Intensive case management for severe mental illness

Abstract

Background

Intensive Case Management (ICM) is a community based package of care, aiming to provide long term care for severely mentally ill people who do not require immediate admission. ICM evolved from two original community models of care, Assertive Community Treatment (ACT) and Case Management (CM), where ICM emphasises the importance of small caseload (less than 20) and high intensity input.

Objectives

To assess the effects of Intensive Case Management (caseload <20) in comparison with non-Intensive Case Management (caseload > 20) and with standard community care in people with severe mental illness. To evaluate whether the effect of ICM on hospitalisation depends on its fidelity to the ACT model and on the setting.

Search methods

For the current update of this review we searched the Cochrane Schizophrenia Group Trials Register (February 2009), which is compiled by systematic searches of major databases, hand searches and conference proceedings.

Selection criteria

All relevant randomised clinical trials focusing on people with severe mental illness, aged 18 to 65 years and treated in the community-care setting, where Intensive Case Management, non-Intensive Case Management or standard care were compared. Outcomes such as service use, adverse effects, global state, social functioning, mental state, behaviour, quality of life, satisfaction and costs were sought.

Data collection and analysis

We extracted data independently. For binary outcomes we calculated relative risk (RR) and its 95% confidence interval (CI), on an intention-to-treat basis. For continuous data we estimated mean difference (MD) between groups and its 95% confidence interval (CI). We employed a random-effects model for analyses.

We performed a random-effects meta-regression analysis to examine the association of the intervention’s fidelity to the ACT model and the rate of hospital use in the setting where the trial was conducted with the treatment effect.

Main results

We included 38 trials (7328 participants) in this review. The trials provided data for two comparisons: 1. ICM versus standard care, 2. ICM versus non-ICM.

1. ICM versus standard care

Twenty-four trials provided data on length of hospitalisation, and results favoured Intensive Case Management (n=3595, 24 RCTs, MD −0.86 CI −1.37 to −0.34). There was a high level of heterogeneity, but this significance still remained when the outlier studies were excluded from the analysis (n=3143, 20 RCTs, MD −0.62 CI −1.00 to −0.23). Nine studies found participants in the ICM group were less likely to be lost to psychiatric services (n=1633, 9 RCTs, RR 0.43 CI 0.30 to 0.61, I²=49%, p=0.05).

One global state scale did show an Improvement in global state for those receiving ICM, the GAF scale (n=818, 5 RCTs, MD 3.41 CI 1.66 to 5.16). Results for mental state as measured through various rating scales, however, were equivocal, with no compelling evidence that ICM was really any better than standard care in improving mental state. No differences in mortality between ICM and standard care groups occurred, either due to ’all causes’ (n=1456, 9 RCTs, RR 0.84 CI 0.48 to 1.47) or to ’suicide’ (n=1456, 9 RCTs, RR 0.68 CI 0.31 to 1.51).

Social functioning results varied, no differences were found in terms of contact with the legal system and with employment status, whereas significant improvement in accommodation status was found, as was the incidence of not living independently, which was lower in the ICM group (n=1185, 4 RCTs, RR 0.65 CI 0.49 to 0.88).

Quality of life data found no significant difference between groups, but data were weak. CSQ scores showed a greater participant satisfaction in the ICM group (n=423, 2 RCTs, MD 3.23 CI 2.31 to 4.14).

2. ICM versus non-ICM

The included studies failed to show a significant advantage of ICM in reducing the average length of hospitalisation (n=2220, 21 RCTs, MD −0.08 CI −0.37 to 0.21). They did find ICM to be more advantageous than non-ICM in reducing rate of lost to follow-up (n= 2195, 9 RCTs, RR 0.72 CI 0.52 to 0.99), although data showed a substantial level of heterogeneity (I²=59%, p=0.01). Overall, no significant differences were found in the effects of ICM compared to non-ICM for broad outcomes such as service use, mortality, social functioning, mental state, behaviour, quality of life, satisfaction and costs.

3. Fidelity to ACT

Within the meta-regression we found that i. the more ICM is adherent to the ACT model, the better it is at decreasing time in hospital (’organisation fidelity’ variable coefficient −0.36 CI −0.66 to −0.07); and ii. the higher the baseline hospital use in the population, the better ICM is at decreasing time in hospital (’baseline hospital use’ variable coefficient −0.20 CI −0.32 to −0.10). Combining both these variables within the model, ’organisation fidelity’ is no longer significant, but ’baseline hospital use’ result is still significantly influencing time in hospital (regression coefficient −0.18 CI −0.29 to −0.07, p=0.0027).

Authors’ conclusions

ICM was found effective in ameliorating many outcomes relevant to people with severe mental illnesses. Compared to standard care ICM was shown to reduce hospitalisation and increase retention in care. It also globally improved social functioning, although ICM’s effect on mental state and quality of life remains unclear. ICM is of value at least to people with severe mental illnesses who are in the sub-group of those with a high level of hospitalisation (about 4 days/month in past 2 years) and the intervention should be performed close to the original model.

It is not clear, however, what gain ICM provides on top of a less formal non-ICM approach.

We do not think that more trials comparing current ICM with standard care or non-ICM are justified, but currently we know of no review comparing non-ICM with standard care and this should be undertaken.

BACKGROUND

Description of the condition

Worldwide, more than 25% of people develop one or more mental or behavioural disorders, during their entire lifetime (WHO 2001). Schizophrenia is one illness heavily contributing to the numbers of people considered severely mentally ill. The lifetime prevalence of schizophrenia alone is 0.58% in the adult population (Warner 1995). Currently this one illness is 26th on the list of diseases ranked according to contribution to overall burden in term of disability-adjusted life years (DALYs). Its ranking is projected to rise to 20th by the year 2020, with more than 17 million DALYs lost (accounting for 1.25% of overall burden, (Murray 1996). Other psychiatric/psychological conditions, however, can also cause very considerable problems in a person’s functioning. Many people with other types of non-organic psychotic illness, or even non-psychotic disorders such as personality disorder, can be considered severely mentally ill.

There has been lack of consensus about the definition of ’severe mental illness’ but the most common dimensions used to identify this group are i. diagnosis, ii. disability, iii. duration and iv. abnormal behaviour. There is, however, little consistency between dimensions and thresholds used in different settings (Slade 1997). The definition of severe mental illness with the widest consensus is that of the National Institute of Mental Health (NIMH) (Schinnar 1990). Their definition is based on three criteria: i. diagnosis of non-organic psychosis or personality disorder; ii. duration characterised as involving ’prolonged illness’ and ’long term treatment’ and operationalised as a two-year or longer history of mental illness or treatment; and iii. disability, which includes dangerous or disturbing social behaviour, moderate impairment in work and non-work activities and mild impairment in basic needs (National Institute of Mental Health 1987).

A survey was conducted in Europe aiming to calculate prevalence rates of severe mental illness according to the NIMH definition. This put the total population-based annual prevalence at approximately two per thousand (Ruggeri 2000).

Description of the intervention

Since the 1960s there has been an almost worldwide trend towards the closure of institutions for the mentally ill. Coupled with these closures, many government policies have focused on reducing the number of hospital beds for people with severe mental illness in favour of providing care in a variety of non-hospital settings - outpatient clinics, day centres or community mental health centres. These changes were consistent with the increasing shift from hospital-based care in favour of a more community-focused approach (Malone 2007).

A major thrust towards community care has been the development of community mental health teams (Bouras 1986; Bennett 1991). Over the last 20 years various types of approaches to working within community mental health teams have evolved. These approaches fall into two main categories: i. services with well-delimited aims, such as crisis resolution and home treatments teams, vocational rehabilitation and early intervention service; and ii. services aimed at meeting a wide range of patient needs, such as Assertive Community Treatment (ACT) and Case Management (CM) (Ruggeri 2008).

Assertive Community Treatment and Case Management are community-based packages of care, developed in the in the early 1970s. These packages of care were initially conceived to co-ordinate the care of severely mentally ill people discharged from closing mental hospitals. They were, however, soon more widely applied as a means of caring for severely mentally ill patients who did not require immediate admission (Thompson 1990).

Assertive Community Treatment and Case Management do share common goals such as maintaining contact, reducing hospitalisation (and hence cost) and improving outcome (Table 1). There are, however, at least in theory and respect to the original models, important structural distinctions between them. Nonetheless, through clinical practice, across time, the two interventions have evolved and tended to converge into a package of care that can be called ’Intensive Case Management’, containing elements from both the two original models (Scott 1995; Burns 2008). Brokerage case management was rapidly abandoned in favour of Clinical Case Management (Holloway 1995) and more sophisticated, but poorly defined models were developed. In these models case managers have clinical training, provide at least some clinical services, and operate with low caseloads (Rubin 1992; Solomon 1992). Assertive Community Treatment is sometimes described as a team-based case management where Case Management involves individual caseloads, but this distinction is not maintained consistently (Burns 2001). Thus, in both clinical trials and clinical practice, what is presently called ’Case Management’ is likely to contain some elements of ACT practice. These models can be called ’Clinical Case Management’, ’Intensive Case Management’ and ’Strengths Case Management’ (Solomon 1992). However, Intensive Case Management is a broader term often used interchangeably with Assertive Community Treatment but distinguished from it on the grounds that it often lacks one or more ACT programme elements (Burns 2001). Intensive Case Management emphasises the importance of small caseload (usually considerably less than 20) and high intensity input, but is otherwise not clearly defined (Marshall 2008). Intensive case managers are usually clinicians who act as therapist in addition to their case management duties (Marshall 2008).

Table 1. Case management and assertive community treatment.

Case management (CM) The key principle of case management is that a single person - the ‘case manager’ - takes primary responsibility for a defined group of patients in the community. The case manager is responsible for (Holloway 1991): Assessing the patient’s needs Developing a care plan Arranging suitable care from community services Keeping contact with the patient Initially, in its simplest form (referred to as ‘brokerage’), case managers were not mental health professionals, did not provide any direct care and worked independently Assertive community treatment (ACT) Assertive community treatment intervention (Stein 1980) should be practiced according to a defined and validated model, based on the consensus of an international panel of ACT experts (McGrew 1994, McGrew 1995). A key aspect of this model is that ACT is a team-based approach, characteristically a multi-disciplinary team, including social workers, nurses and psychiatrists, caring exclusively for a defined group of patients (McGrew 1995, Olfson 1990). Team members share responsibility for their clients, so it is common for several members to work together with the same person. Other characteristics of the model are (Stein 1980): Provide all necessary care themselves, rather than arranging for it to be provided from other services. Provide care at home or in places of work Carry low caseloads (usually 10-15 patients per member) Practice ‘assertive outreach’, meaning that they persist in attempts to engage uncooperative clients Place particular emphasis on medication compliance Provide 24 hour emergency cover

How the intervention might work

The theory behind care in the community is that it enables people to live as independently as possible within their own homes or ’homely settings’ out of hospital, because unnecessary hospital care is wasteful, un-therapeutic and stigmatising. It was hoped that living in the community would increase opportunities for people with severe mental illness to achieve their full potential as autonomous members of society (Department of Health 1990). Community care policies are also aimed at promoting choice and independence for people experiencing mental health difficulties. Intensive Case Management is an intervention at the level of local service organisation. It is a way of organising teams rather than a specific treatment model (Johnson 2008). Intensive Case Management should provide a mental health service which is a reliable, systematic, flexible and a co-ordinated care method, addressed to answer the unique combination of health and social care needs of people with severe mental illness. It represents a long term intensive approach to the patient in the community (Killaspy 2008), providing a comprehensive range of treatment, rehabilitation and support services (Scott 1995). Intensive Case Management aims to helping people to acquire material resources (such as food, shelter, clothing and medical care) and to improve psychosocial functioning; to provide sufficient support to keep the patient involved in community life and to encourage growth toward greater autonomy; to develop coping skills to meet the demands of community life; to insure continuity of care among treatment agencies (Stein 1980). Key purposes of Intensive Case Management are to improve outcome, reduce hospitalisation and prevent loss of contact with services.

Recently it has been suggested that the success of Intensive Case Management depends on its fidelity to the Assertive Community Treatment model (i.e. if a team approach is properly implemented) and on the setting (i.e. it would work better where there is a high baseline level of bed use) (Burns 2007).

Why it is important to do this review

With the evolution of the original intervention models there was a need to update and merge two previous relevant Cochrane reviews (Marshall 2000a; Marshall 2000b) and to take into account the findings of work by the same authoring team (Burns 2007). During the last ten years intervention models have not only been modified, merged and become more difficult to distinguish in practice, but also research has been more widespread with new studies evaluating these approaches outside of the USA.

Inpatient treatment remains the major cost of healthcare (Burns 2001). The effects of the currently implemented packages of care in different settings should be fully understood across a range of outcomes.

OBJECTIVES

1. Primary objective

To assess the effects of Intensive Case Management as a means of caring for severely mentally ill people in the community in comparison with non-Intensive Case Management (caseload >20) and with standard community care. The review does not distinguish between models of Intensive Case Management.

2. Secondary objective

To assess whether the effect of Intensive Case Management on hospitalisation (mean number of days per month in hospital) is influenced by intervention’s fidelity to Assertive Community Treatment model and by the rate of hospital use in the setting where the trial was conducted (baseline level of hospital use).

METHODS

Criteria for considering studies for this review

Types of studies

All relevant randomised controlled trials were considered, and economic evaluations conducted alongside included randomised controlled trials. Quasi-randomised studies, such as those allocating by using alternate days of the week, were excluded. Where trials were described in some way as to suggest or imply that the study was randomised and where the demographic details of each group’s participants were similar, trials were included and sensitivity analysis was undertaken to the presence or absence of these data.

Types of participants

A majority of participants were required to be i. within the age range 18 to 65, ii. suffering from severe mental illness preferably as defined by National Institute of Mental Health 1987 but, in the absence of that, from illness such as schizophrenia, schizophrenia-like disorders, bipolar disorder, depression with psychotic features or/and personality disorder; and iii. not acutely ill. Substance abuse was not considered to be a severe mental disorder in its own right, however studies were eligible if they dealt with people with both diagnoses, that is those with severe mental illness plus substance abuse. Dementia and mental retardation are not considered to be a severe mental disorder, hence studies focusing on these populations were excluded. We considered only participants treated in the community-care setting.

Types of interventions

We considered only interventions and management packages not focused primarily on alternatives to acute hospital admission.

1. Intensive Case Management

   We defined Intensive Case Management as where the majority of people received:

   1. A package of care shaped either on the:

      • Assertive Community Treatment model, being based on the Treatment in Community Living, Assertive Community Treatment (Stein 1980); or
      • Assertive Outreach model (Witheridge 1982, Witheridge 1991) (i.e. multidisciplinary team-based approach, practicing ’assertive outreach’, providing 24 hrs emergency cover, providing care themselves)(McGrew 1995); or the
      • Case Management model (Intagliata 1982) however described as such in the trial report.

   2. With a caseload up to and including 20 people.

2. Non-Intensive Case Management

   We defined non-Intensive Case Management as where the majority of people received:

   1. A package of care shaped either on the:

      • Assertive Community Treatment model, being based on the Treatment in Community Living, Assertive Community Treatment (Stein 1980), or
      • Assertive Outreach model (Witheridge 1982, Witheridge 1991) (i.e. multidisciplinary team-based approach, practicing ’assertive outreach’, providing 24 hrs emergency cover, providing care themselves)(McGrew 1995); or
      • on the Case Management model (Intagliata 1982) however described as such in the trial report.

   2. With a caseload over 20 people.

3. Standard care

   We defined standard care as where the majority of people received:

   1. A community or outpatient model of care not specifically shaped on either the model of Assertive Community Treatment and Case Management, and not working within a specific designated named package or approach to care.

   2. If data were available on the standard care caseload, a final sensitivity analysis was undertaken testing how prone the primary outcomes were to change when trials comparing Intensive Case Management with standard community care only (caseload >20) were included.

Types of outcome measures

All outcomes were grouped by time into short term (up to 6 months), medium term (7 months to 12 months) and long term (over 12 months). Where available, 24 months was the preferred follow-up point for calculating mean days per months in hospital. If more than one follow-up point within the same period were available, we reported the latest one.

Primary outcomes

• 1Service use
• 1.1Hospitalisation: mean number of days per month in hospital
• 1.2Not remaining in contact with psychiatric services

Secondary outcomes

• 1 Adverse effects

• 1.1 Death - all causes and suicide

• 2 Service use

• 2.1 Admitted to hospital

• 2.2 Hospital admission rate

• 2.3 Use of services outside of mental health provision (i.e. emergency services)

• 3 Global state

• 3.1 Leaving the study early (lost to follow-up)

• 3.2 Relapse (as defined in trial)

• 3.3 Not improved to a clinically meaningful extent (as defined in trial)

• 3.4 Not improved

• 3.5 Average endpoint score

• 3.6 Average change score

• 3.7 Compliance with medication

• 3.8 Average endpoint score

• 3.9 Average change score

• 4 Social functioning

• 4.1 Contact with legal system (i.e. police contacts, arrests, imprisonments)

• 4.2 Employment status (number unemployed at end of study)

• 4.3 Accommodation status (number homeless or not living independently during or at the end of the study, mean days homeless and mean days in stable accommodation per month in study)

• 4.4 Alcohol use

• 4.5 Illicit drug use

• 4.6 Average endpoint score

• 4.7 Average change score

• 5 Mental state

• 5.1 General symptoms

• 5.1.1 Not improved to a clinically meaningful extent (as defined in trial)

• 5.1.2 Not improved

• 5.1.3 Average endpoint score

• 5.1.4 Average change score

• 5.2 Specific symptoms

• 5.2.1 Positive symptoms (delusions, hallucinations, disordered thinking)

• 5.2.1.1 Not improved to a clinically meaningful extent (as defined in trial)

• 5.2.1.2 Not improved

• 5.2.1.3 Average endpoint score

• 5.2.1.4 Average change score

• 5.2.2 Negative symptoms (poor volition, poor self-care, blunted affect)

• 5.2.2.1 Not improved to a clinically meaningful extent (as defined in trial)

• 5.2.2.2 Not improved

• 5.2.2.3 Average endpoint score

• 5.2.2.4 Average change score

• 5.2.3 Mood depression

• 5.2.3.1 Not improved to a clinically meaningful extent (as defined in trial)

• 5.2.3.2 Not improved

• 5.2.3.3 Average endpoint score

• 5.2.3.4 Average change score

• 6 Behaviour

• 6.1 General behaviour

• 6.2 Not improved to a clinically meaningful extent (as defined in trial)

• 6.3 Not improved

• 6.4 Average endpoint score

• 6.5 Average change score

• 6.6 Specific behaviours (i.e. self-harm; injury to others or property)

• 7 Quality of life

• 7.1 Not improved to a clinically meaningful extent (as defined in trial)

• 7.2 Not improved

• 7.3 Average endpoint score

• 7.4 Average change score

• 8 Satisfaction

• 8.1 Participant satisfaction

• 8.1.1 Not improved to a clinically meaningful extent (as defined in trial)

• 8.1.2 Not improved

• 8.1.3 Average endpoint score

• 8.1.4 Average change score

• 8.2 Carer satisfaction

• 8.2.1 Not improved to a clinically meaningful extent (as defined in trial)

• 8.2.2 Not improved

• 8.2.3 Average endpoint score

• 8.2.4 Average change score

• 9 Costs Measures of cost were expressed as mean monthly costs per patient in the trial. Three main types of cost were reported:

• 9.1 Direct costs of psychiatric hospital care

• 9.2 Direct health care costs (including all medical and psychiatric care and the costs of case management, but excluding accommodation other than hospital care)

• 9.3 Direct costs of all care (including costs of accommodation and subtracting benefits, such as earnings, where these are known)

Search methods for identification of studies

For search methods of first versions of this review please see Appendix 1 and Appendix 2.

Electronic searches

1. Cochrane Schizophrenia Group Trials Register (February 2009)

The register was searched using the phrase: (*ca?e management* OR *cpa* OR *community treatment* OR *community team* OR *community cent* OR *community care approach* OR *madison model* OR *outreach* OR *hostel* OR *aftercare* OR *residential* OR *housing* OR *transitional* OR *posthospital* OR *partial hospitali?ation* OR *Foster* OR *Guardianship* OR *daily living programme* OR *crisis intervention* OR *early intervention* OR *Ambulatory treatment* OR *Ambulatory care* OR *community living* OR *social support* OR *patient care team* OR *community mental health* OR *patient participation* OR *assertive outreach* OR *drop-in hospital* OR *drop-in care* OR *drop-in treatment* OR *dropin cent* OR *drop-in unit* OR *drop in hospital* OR *drop in care* OR *drop in treatment* OR *drop in cent* OR *drop in unit* OR *day hospital* OR *day care* OR *day treatment* OR *day cent* OR *day unit* OR *Intensive care* OR *Intensive interven* OR *Intensive treat* OR *Intensive therap* OR *Intensive management* OR *Intensive model* OR *Intensive programmem* OR *Intensive team* OR *Intensive service* OR *mobile care* OR *mobile interven* OR *mobile treat* OR *mobile therap* OR *mobile management* OR *mobile model* OR *mobile programmem* OR *mobile team* OR *mobile service* OR *outreach care* OR *outreach interven* OR *outreach treat* OR *outreach therap* OR *outreach management* OR *outreach model* OR *outreach programmem* OR *outreach team* OR *outreach service* OR *community care* OR *community interven* OR *community treat* OR *community therap* OR *community management* OR *community model* OR *community programmem* OR *community team* OR *community service* OR *community base* OR *home care* OR *home interven* OR *home treat* OR *home therap* OR *home management* OR *home model* OR *home programmem* OR *home team* OR *home service* OR *home base* OR *aggressive outreach* OR *broker* OR *programme* OR *care programme approach* OR *care programme* in title, abstract, index terms of REFERENCE and in interventions of STUDY) OR (*Pact* OR *tcl* In title) OR (Pact* OR tcl* in abstract and index terms of REFERENCE and in interventions of STUDY)

This register is compiled by systematic searches of major databases, hand searches and conference proceedings (see Group Module - Specialised Register).

Searching other resources

1. References

Should an included or excluded study suggest that another study was of relevance, the reference was identified and the full text acquired.

2. Personal contact

Where required for additional data we contacted authors of trials for this information. We did not systematically contact all authors for additional papers.

Data collection and analysis

Selection of studies

The principal reviewer (MD) inspected all abstracts of studies identified as above and identified potentially relevant reports. In addition, to ensure reliability, CBI inspected a random sample of these abstracts, comprising 10% of the total. Where disagreement occurred this was resolved by discussion, or where there was still doubt, the full article was acquired for further inspection. The full articles of relevant reports were acquired for reassessment and carefully inspected for a final decision on inclusion (see Criteria for considering studies for this review). Once the full articles were obtained, in turn MD and CBI inspected all full reports and independently decided whether they met inclusion criteria. MD and CBI were not blinded to the names of the authors, institutions or journal of publication. Where difficulties or disputes arose, we asked author MM for help and if it was impossible to decide, these studies were added to those awaiting assessment and the authors of the papers contacted for clarification.

Data extraction and management

1. Extraction

1.1 Data regarding criteria and outcomes

The principal reviewer (MD) extracted data from all included studies. In addition, to ensure reliability, CBI independently extracted data from a random sample of these studies, comprising 10% of the total. Again, any disagreement was discussed, decisions documented and, if necessary, authors of studies were contacted for clarification. With remaining problems MM helped clarify issues and those final decisions were documented. Data presented only in graphs and figures were extracted whenever possible, but were included only if two reviewers independently had the same result. Attempts were made to contact authors through an open-ended request in order to obtain missing information or for clarification whenever necessary. Where possible, we extracted data relevant to each component centre of multi-centre studies separately.

1.2 Additional data

1.2.1 Fidelity

This rating related to the Intensive Case Management intervention. This rated fidelity of the intervention to assertive community treatment on the ’team membership’ and ’team structure and organisation’ sub-scales of the Index of Fidelity to Assertive Community Treatment (IFACT) (McGrew 1994).

This index was derived from a survey of 20 clinical experts in assertive community treatment and validated in a survey of 18 programmes.

1. The ’team membership’ sub-scale comprises four items:

   • -
     ratio of patients to staff
   • -
     total size of team
   • -
     extent of psychiatric input
   • -
     extent of nursing input to the team
2. The ’structure and organisation’ sub-scale comprises seven items, whether the team is:

   • -
     the primary source of care for its patients
   • -
     is situated away from the hospital
   • -
     meets daily
   • -
     shares responsibility for caseloads
   • -
     is available 24 hours a day
   • -
     has a team leader who is also a case manager
   • -
     offers unlimited time for its services

We chose IFACT because the sub-scales are brief and are possible to be completed from published or unpublished text. For each item on the index, a score of one indicates high fidelity to the model. Score ranges from 0 to 11, where the maximum score available on ’team membership’ sub-scale is 4, and on ’structure and organisation’ sub-scale is 7, with higher scores indicating higher fidelity to the model.

We obtained fidelity data from published and unpublished trial reports, direct contact with trialists, and data previously obtained directly from trialists by previous reviews (Burns 2001, Catty 2002, Burns 2007). Two raters (MD and CBI) independently combined these data into a single fidelity score. Multicentre trials of Intensive Case Management often struggle to implement a uniform approach, with centres operating at different degrees of fidelity. Where possible, we rated each component centre separately.

1.2.2 Baseline hospital use

The average number of days per month in hospital for all participants in the two years before the study began.

We obtained this data from published and unpublished trial reports and from direct contact with trialists.

1.2.3 Service use: hospitalisation

We obtained the primary outcome mean number of days per month in hospital for the included studies from published and unpublished trial reports, direct contact with trialists and data previously obtained directly from trialists reported by a previous review (Burns 2007).

2. Management

2.1 Forms

Data were extracted onto standard, simple forms.

2.2 Data from multi-centre trials

Where possible MD and CBI verified independently calculated centre data against original trial reports.

3. Scale-derived data

We included continuous data from rating scales only if:

1. the psychometric properties of the measuring instrument had been described in a peer-reviewed journal (Marshall 2000); and

2. the measuring instrument was not written or modified by one of the trialists for that particular trial; and

3. the measuring instrument is either i. a self-report or ii. completed by an independent rater or relative (not the therapist).

4. Endpoint versus change data

We preferred to use scale endpoint data, which typically cannot have negative values and is easier to interpret from a clinical point of view. Change data are often not ordinal and are very problematic to interpret. If endpoint data were unavailable, we used change data.

5. Skewed data

5.1 General

Continuous data on clinical and social outcomes are often not normally distributed. To avoid the pitfall of applying parametric tests to non-parametric data, we aim to apply the following standards to all data before inclusion: a) standard deviations and means are reported in the paper or obtainable from the authors; b) when a scale starts from the finite number zero, the standard deviation, when multiplied by two, is less than the mean (as otherwise the mean is unlikely to be an appropriate measure of the centre of the distribution, (Altman 1996); c) if a scale started from a positive value (such as PANSS which can have values from 30 to 210) the calculation described above was modified to take the scale starting point into account. In these cases skew is present if 2SD>(S-S min), where S is the mean score and S min is the minimum score. Endpoint scores on scales often have a finite start and end point and these rules can be applied. When continuous data are presented on a scale which includes a possibility of negative values (such as change data), it is difficult to tell whether data are skewed or not. Skewed data from studies of less than 200 participants were entered in additional tables rather than into an analysis. Skewed data pose less of a problem when looking at means if the sample size is large and were entered into syntheses.

5.2 Specific - mean number of days in hospital

We implemented one exception to the above rule (5.1) in order to present more data, recognising that this is a ’post hoc’ decision, but also that the rules as regards management of skewed data and how robust skewed data are within meta-analysis is unclear (Higgins 2008). Where mean number of days in hospital data were skewed, and they were provided by studies of less than 200 participants, we nevertheless entered those data into a sub-group of the overall analysis. We also presented the overall effect from all data pooled.

6. Common measure

To facilitate comparison between trials, we intended to convert variables that can be reported in different metrics, such as days in hospital (mean days per year, per week or per month) to a common metric (e.g. mean days per month).

7. Conversion of continuous to binary

Where possible, efforts were made to convert outcome measures to dichotomous data. This could be done by identifying cut-off points on rating scales and dividing participants accordingly into ’clinically improved’ or ’not clinically improved’. It was generally assumed that if there had been a 50% reduction in a scale-derived score such as the Brief Psychiatric Rating Scale (BPRS, Overall 1962) or the Positive and Negative Syndrome Scale (PANSS, Kay 1986), this could be considered as a clinically significant response (Leucht 2005a, Leucht 2005b). If data based on these thresholds were not available, we used the primary cut-off point presented by the original authors.

8. Direction of graphs

Where possible, we entered data in such a way that the area to the left of the line of no effect indicates a favourable outcome for Intensive Case Management.

9. Summary of findings table

We anticipated including the following long term main outcomes in a summary of findings table.

• Service use

  • 1.1 Hospitalisation: mean number of days per month in hospital

  • 1.2 Hospital admission across time

• Global state

  • 2.1 Relapse (as defined in trial)

  • 2.2 Leaving the study early (lost to follow-up)

• Mental state: general symptoms

  • 3.1. Not improved to a clinically meaningful extent (as defined in trial)

• Adverse effect

  • 4.1Death - suicide
• 5Social functioning
  • 5.1Employment - unemployed at end of study

Assessment of risk of bias in included studies

Again working independently, the principal reviewer (MD) assessed risk of bias of all included studies and the second reviewer (CBI) assessed risk of bias from a random sample of these studies, comprising 10% of the total. MD and CBI assessed risk of bias using the tool described in the Cochrane Collaboration Handbook (Higgins 2008). This tool encourages consideration of how the sequence was generated, how allocation was concealed, the integrity of blinding at outcome, the completeness of outcome data, selective reporting and other biases. We would have excluded studies where allocation was clearly not concealed.

Trials with high risk of bias (defined as at least three out of five domains categorised as ’No’) were not included in the meta-analysis. If the raters disagreed, the final rating was made by consensus with the involvement of another member of the review group. Where inadequate details of randomisation and other characteristics of trials are provided, authors of the studies were contacted in order to obtain further information. Non-concurrence in quality assessment was reported.

Measures of treatment effect

1. Binary data

For binary outcomes we calculated a standard estimation of the random-effects risk ratio (RR) and its 95% confidence interval (CI). It has been shown that RR is more intuitive (Boissel 1999) than odds ratios and that odds ratios tend to be interpreted as RR by clinicians (Deeks 2000). For statistically significant results we had planned to calculate the number needed to treat to provide benefit /to induce harm statistic (NNTB/H), and its 95% confidence interval (CI) using Visual Rx (http://www.nntonline.net/) taking account of the event rate in the control group. This, however, was superseded by Summary of findings for the main comparison and the calculations therein.

2. Continuous data

2.1 Summary statistic

For continuous outcomes we estimated a random-effects mean difference (MD) between groups. We preferred not to calculate effect size measures (standardised mean difference SMD). However, in the case of where scales were of such similarity to allow presuming there was a small difference in measurement, we calculated it and, whenever possible, we transformed the effect back to the units of one or more of the specific instruments.

Unit of analysis issues

1. Cluster trials

Studies increasingly employ ’cluster randomisation’ (such as randomisation by clinician or practice) but analysis and pooling of clustered data poses problems. Firstly, authors often fail to account for intra-class correlation in clustered studies, leading to a ’unit of analysis’ error (Divine 1992) whereby p values are spuriously low, confidence intervals unduly narrow and statistical significance overestimated. This causes type I errors (Bland 1997, Gulliford 1999).

Where clustering is not accounted for in primary studies, we presented data in a table, with a (*) symbol to indicate the presence of a probable unit of analysis error. In subsequent versions of this review we will seek to contact first authors of studies to obtain intra-class correlation coefficients for their clustered data and to adjust for this by using accepted methods (Gulliford 1999). Where clustering had been incorporated into the analysis of primary studies, we present these data as if from a non-cluster randomised study, but adjusted for the clustering effect.

We have sought statistical advice and have been advised that the binary data as presented in a report should be divided by a ’design effect’. This is calculated using the mean number of participants per cluster (m) and the intra-class correlation coefficient (ICC) [Design effect = 1+(m-1)*ICC] (Donner 2002). If the ICC was not reported it was assumed to be 0.1 (Ukoumunne 1999).

If cluster studies has been appropriately analysed taking into account intra-class correlation coefficients and relevant data documented in the report, synthesis with other studies would have been possible using the generic inverse variance technique.

2. Cross-over trials

A major concern of cross-over trials is the carry-over effect. It occurs if an effect (e.g. pharmacological, physiological or psychological) of the treatment in the first phase is carried over to the second phase. As a consequence on entry to the second phase the participants can differ systematically from their initial state despite a wash-out phase. For the same reason cross-over trials are not appropriate if the condition of interest is unstable (Elbourne 2002). As both effects are very likely in severe mental illness, we only used data of the first phase of cross-over studies.

3. Studies with multiple treatment groups

Where a study involved more than two treatment arms, if relevant, the additional treatment arms were presented in comparisons. Where the additional treatment arms were not relevant, these data were not reproduced.

Dealing with missing data

1. Overall loss of credibility

At some degree of loss of follow-up, data must lose credibility (Xia 2007). For any particular outcome should more than 50% of data be unaccounted for, we did not reproduce these data or use them within analyses. If, however, more than 50% of those in one arm of a study were lost, but the total loss was less than 50%, we marked such data with (*) to indicate that such a result may well be prone to bias.

2. Binary

In the case where attrition for a binary outcome is between 0 and 50% and where these data were not clearly described, data were presented on a ’once-randomised-always-analyse’ basis (an intention to treat analysis). Those leaving the study early were all assumed to have the same rates of negative outcome as those who completed, with the exception of the outcome of death. A sensitivity analysis was undertaken testing how prone the primary outcomes were to change when data from only those who completed the study were compared with intention to treat data using the assumption outlined above.

3. Continuous

3.1 Attrition

In the case where attrition for a continuous outcome is between 0 and 50% and data from only those who completed the study are reported, we have reproduced these.

3.2 Standard deviations

3.2.1 General

Where there are missing measures of variance for continuous data, but exact standard errors or confidence intervals for group means, or either ’p’ or ’t’ values for differences in means, we calculated standard deviation value according to method described in Section 7.7.3 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). If standard deviations were not reported and could not be calculated from available data, we asked authors to supply the data. In the absence of data from authors, we used the mean standard deviation from other studies.

3.2.2 Standard deviation mean number of days per month in hospital

For the primary outcome, mean number of days per month in hospital, if standard deviations were not reported and could not be calculated from available data, we asked authors for additional information. In the absence of data from authors, we imputed the missing standard deviations using a regression analysis of SD against mean from those trials that provided both. We documented in Table 2 in what studies we imputed SDs according to the above technique.

Table 2. Average number of days in hospital per month - at about 24 months - entering meta-regression.

INTENSIVE CASE MANAGEMENT vs STANDARD CARE	ICM	ICM	ICM	SC	SC	SC
Study ID	Mean	SD	Total	Mean	SD	Total	Note
Audini-UK	0.95	2.84*	33	0.93	2.03*	33	* SD imputed
Bjorkman-Sweden	0.83	3.13	33	2.15	4.13	44
Bond-Chicago1	3.22	4.55	42	5.3	5.42	40
Bond-Indiana1 (A)	1.28	3.17*	29	7.72	8.99*	32	* SD imputed.
Bond-Indiana1 (B)	2.72	4.54*	34	3.62	5.24*	30	* SD imputed.
Bond-Indiana1 (C)	0.05	1.89*	21	3.38	4.98*	21	* SD imputed.
Chandler-California1 (A)	0.47	2.34*	102	0.78	1.84*	101	* SD imputed.
Chandler-California1 (B)	0.67	2.55*	115	0.96	2.07*	114	* SD imputed.
Curtis-New York	1.77	1.79	146	1.02	1.18	143
Ford-UK	3.07	6.9	39	1.76	3.67	38
Hampton-Illinois (A)	1.75	3.63*	48	4.83	6.49*	47	* SD imputed.
Hampton-Illinois (B)	3.25	5.01*	34	3.42	5.02*	36	* SD imputed.
Holloway-UK	2.4	5.1	34	1.2	3	26
Jerrell-SCarolina1	0.53	2.40*	40	0.8	1.86*	40	* SD imputed.
Lehman-Maryland1	3.04	5.15	77	5.41	7	75
Marshall-UK	1.04	2.18	40	1.56	4.45	40
Muijen-UK2	2.53	5.55	41	2.45	5.83	41
Muller-Clemm-Canada	1.68	3.56*	61	1.63	2.93*	57	* SD imputed
OPUS-Denmark	5.11	7.7	263	6.57	8.73	244
Quinlivan-California	1.09	2.65	30	5.53	8.65	30
Rosenheck-USA-GMS (A)	3.63	3.89	44	3.71	2.76	35
Rosenheck-USA-GMS (B)	6.99	4.85	47	4.23	5.18	47
Rosenheck-USA-NP (C)	18.52	11.16	50	19.16	12.19	43
Rosenheck-USA-GMS (D)	2.8	3.31	49	3.26	3.98	53
Rosenheck-USA-NP (E)	4.13	5.24	34	3.05	4.61	33
Rosenheck-USA-GMS (F)	2.39	3.16	43	2.58	2.45	35
Rosenheck-USA-NP (G)	7.68	7.72	40	12.2	10.65	31
Rosenheck-USA-NP (H)	4.63	8.58	59	11.21	13.38	55
Rosenheck-USA-GMS (I)	5.62	4.67	44	7.8	6.63	44
Sytema-Netherlands	3.4	5.4	58	4.3	7.3	57
Test- Wisconsin	0.42	2.29*	72	2.13	3.54*	41	* SD imputed.
INTENSIVE CASE MANAGEMENT vs NON-INTENSIVE CASE MANAGEMENT	ICM	ICM	ICM	NON-ICM	NON-ICM	NON-ICM
Study ID	Mean	SD	Total	Mean	SD	Total	Note
Bush-Georgia	1.58	3.46*	14	2.39	3.85*	14	* SD imputed.
Drake-NHamp (A)	0.5	0.94	7	2.17	3.21	9
Drake-NHamp (B)	0.85	1.43	16	1.41	2.06	14
Drake-NHamp (C)	2.28	3.2	10	1.67	3.84	12
Drake-NHamp (D)	1.04	2.44	13	0.63	0.91	11
Drake-NHamp (E)	1.08	4.15	30	1.39	2.36	27
Drake-NHamp (F)	1.66	4.49	10	0.84	2.33	13
Drake-NHamp G	2.05	3.06	9	0.87	0.92	8
Essock-Connecticut1	2.87	7.82	130	4.3	9.52	132
Essock-Connecticut2	0.64	1.9	99	0.72	1.3	99
Harrison-Read-UK	2.94	5.74	97	3.76	5.83	96
Johnston-Australia	4.0	5.75	35	3.08	4.3	33
McDonel-Indiana (A)	3.15	7.1	61	1.43	2.91	64
McDonel-Indiana (B)	1.22	3.66	14	0.58	1.29	17
Quinlivan-California	1.09	2.65	30	2.8	4.74	30
REACT-UK	9.0	8.9	124	8.0	7.8	119
Salkever-SCarolina	1.12	3.01*	91	1.3	2.51*	53	* SD imputed.
UK700-UK (A)	3.08	5.77	94	2.64	3.49	95
UK700-UK (B)	3.2	4.79	77	3.16	4.97	73
UK700-UK (C)	3.29	5.41	76	2.48	4.71	75
UK700-UK (D)	2.74	4.69	91	3.79	5.22	98

ICM: Intensive Case Management.

NON-ICM: Non-Intensive Case Management.

SC: Standard Care.

SD: Standard Deviation.

Study ID: Study identification name.

3.3 Last observation carried forward

We anticipated that in some studies the method of last observation carried forward (LOCF) would be employed within the study report. As with all methods of imputation to deal with missing data, LOCF introduces uncertainty about the reliability of the results. Therefore, where LOCF data has been used in the trial, if less than 50% of the data had been assumed, we reproduced these data and indicated that they are the product of LOCF assumptions.

3.4 Incomplete data for meta-regression

In some cases we anticipated that IFACT score variables would not all be available. If IFACT score could not be calculated from available data, we imputed it by multiple imputation using the mi library in R (R 2008). As explained above, we only made these assumptions if we were able to directly rate over 50% of the data. We documented in Table 3 in what studies we calculated IFACT score according to the above technique.

Table 3. Covariates entering meta-regression.

INTENSIVE CASE MANAGEMENT vs STANDARD CARE	Baseline hospital use	Baseline hospital use	IFACT	IFACT	IFACT
Study ID	Mean	Total	Total score	Organisation sub-scale score	Staff sub-scale score	Note
Audini-UK	1.08	66	6.7	3.5	3.2
Bjorkman-Sweden	5.63	77	7	4.5	2.5
Bond-Chicago1	7.83	88	6	4	2
Bond-Indiana1 (A)	14.17	61	9.2	7	2.2
Bond-Indiana1 (B)	4.95	64	2.2	1	1.2
Bond-Indiana1 (C)	10.86	42	7.4	5	2.4
Chandler-California1 (A)	0.5	203	8.5	5	3.5
Chandler-California1 (B)	1.14	229	6.6	5	1.6
Curtis-New York	0.95*	289	5.8	3.5	2.3	* Mean Imputed.
Ford-UK	2.61	77	4.8	2	2.8
Hampton-Illinois (A)	5.6	95	6	4	2
Hampton-Illinois (B)	5.2	70	5	3	2
Holloway-UK	7.37	70	9.3	6	3.3
Jerrell-SCarolina1	2.85	80	8.8	5.5	3.3
Lehman-Maryland1	4,94*	152	11	7	4	* Mean Imputed.
Marshall-UK	3.31*	80	4.9	4	0.9	* Mean Imputed.
Muijen-UK2	8.43*	82	5.4	3	2.4	* Mean Imputed.
Muller-Clemm-Canada	4.07	123	6.2	4	2.2
OPUS-Denmark	NA	547	8	4	4	* Baseline hospital use: not applicable as first episode.
Quinlivan-California	4.50*	60	6.4	4	2.4	* Mean Imputed.
Rosenheck-USA-GMS (A)	3.96	79	6	2	4
Rosenheck-USA-GMS (B)	5.83	94	3.8	2	1.8
Rosenheck-USA-NP (C)	19.8	93	7.7	5	2.7
Rosenheck-USA-GMS (D)	4.19	102	7	3	4
Rosenheck-USA-NP (E)	5.33	67	6.4	3.5	2.9
Rosenheck-USA-GMS (F)	3.22	78	6.6	3	3.6
Rosenheck-USA-NP (G)	11.42	71	8.4	5	3.4
Rosenheck-USA-NP (H)	11.4	114	6.4	4	2.4
Rosenheck-USA-GMS (I)	8.28	88	5.8	2	3.8
Sytema-Netherlands	12.17*	118	7.6*	5.1*	2.5*	* Mean and IFACT score imputed.
Test-Wisconsin	2.33	122	8,5	5.5	3
INTENSIVE CASE MANAGEMENT vs NON-INTENSIVE CASE MANAGEMENT	Baseline hospital use	Baseline hospital use	IFACT	IFACT	IFACT
Study ID	Mean	Total	Total score	Organisation subscale score	Staff subscale score	Note
Bush-Georgia	3.99	28	3.1	2	1.1
Drake-NHamp (A)	2.88	19	8	5	3
Drake-NHamp (B)	1.72	33	3.8	3	0.8
Drake-NHamp (C)	3.02	25	8.8	5.5	3.3
Drake-NHamp (D)	1.78	26	7.8	4.5	3.3
Drake-NHamp (E)	2.76	66	8.5	4.5	4
Drake-NHamp (F)	2.34	22	3.5	3	0.5
Drake-NHamp (G)	4.1	19	5	2	3
Essock-Connecticut1	2.81*	262	8.5	4.5	4	* Mean Imputed.
Essock-Connecticut2	1.08*	198	10*	7*	3*	* Mean, and IFACT score imputed.
Harrison-Read-UK	4.11	193	7.6	4	3.6
Johnston-Australia	3.66	71	7.3	3.5	3.8
McDonel-Indiana (A)	4.2	152	4.2	3	1.2
McDonel-Indiana (B)	1.16	39	4.4	3	1.4
Quinlivan-California	2.96*	60	6.4	4	2,4	* Mean Imputed.
REACT-UK	7.3	251	10.3	6.5	3.8
Salkever-SCarolina	3.06	144	7	5	2
UK700-UK (A)	4.55	196	8.8	5	3.8
UK700-UK (B)	4.66	153	4.5	3	1.5
UK700-UK (C)	4.33	158	4.2	2	2.2
UK700-UK (D)	4.59	200	8.5	5	3.5

Baseline hospital use: average number of days per month in hospital for all participants in the two years before the study began.

IFACT: Index of Fidelity to Assertive Community Treatment.

Study ID: Study identification name.

In some cases we anticipated that baseline hospital use data would not all be available. Missing data was imputed as for the IFACT scores. As explained above, we only make these assumptions if we were able to directly rate over 50% of the data. We documented for which studies we calculated baseline hospital use data according to the above technique (Table 3).

A sensitivity analysis was undertaken testing how prone the results from meta-regression were to change when data from only those who completed the studies were compared with the imputed data using the assumption outlined above.

Assessment of heterogeneity

1. Clinical heterogeneity

We considered all included studies initially, without seeing comparison data, to judge clinical heterogeneity. We simply inspected all studies for clearly outlying situations or people, which we had not predicted would arise. When such situations or participant groups arose, these were fully discussed.

In addition two potential sources of heterogeneity were specified a priori (fidelity and baseline level of hospital use (Data extraction and management). These data were extracted as described above.

2. Methodological heterogeneity

We considered all included studies initially, without seeing comparison data, to judge methodological heterogeneity. We simply inspected all studies for clearly outlying methods which we had not predicted would arise. Should such methodological outliers arise these will be fully discussed.

3. Statistical heterogeneity

3.1 Visual inspection

We visually inspected graphs to investigate the possibility of statistical heterogeneity.

3.2 Employing the I²statistic

Heterogeneity between studies was investigated by considering the I² method alongside the Chi² ’p’ value. The I² provides an estimate of the percentage of inconsistency thought to be due to chance (Higgins 2003). The importance of the observed value of I² depends on i. magnitude and direction of effects and ii. strength of evidence for heterogeneity (e.g. ’p’ value from Chi² test, or a confidence interval for I²). I² estimate greater than or equal to 50% accompanied by a statistically significant Chi² statistic, was interpreted as evidence of substantial levels of heterogeneity (Section 9.5.2 - Higgins 2008). When substantial levels of heterogeneity were found in the primary outcome, we explored reasons for heterogeneity (Subgroup analysis and investigation of heterogeneity).

Assessment of reporting biases

Reporting biases arise when the dissemination of research findings is influenced by the nature and direction of results (Egger 1997). These are described in Section 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). We are aware that funnel plots may be useful in investigating reporting biases but are of limited power to detect small-study effects. We did not use funnel plots for outcomes where there were ten or fewer studies, or where all studies were of similar sizes. In other cases, where funnel plots were possible, we sought statistical advice in their interpretation.

Data synthesis

Where possible we employed a random-effects model for analyses. We understand that there is no closed argument for preference for use of fixed or random-effects models. The random-effects method incorporates an assumption that different studies are estimating different, yet related, intervention effects. According to our hypothesis of an existing variation across studies, to be explored further in the meta-regression analysis despite being cautious that random-effects methods does put added weight onto the smaller of the studies - we favoured using random-effects model.

Subgroup analysis and investigation of heterogeneity

1. Subgroup analyses

We anticipate two sub-group analyses. For the first version of the protocol for this review we did not anticipate any sub-group analyses. On further consideration we now realise that analysis at separate time periods could be thought of as sub-groups. The second sub-group is within the primary outcome and relates to skewed and non-skewed data. This has been introduced late into this protocol and could be considered post hoc. However, we are also conscious that our original rule for management of these data could be considered overly cautious and result in some important data not being presented (Higgins 2008).

2. Investigation of heterogeneity

2.1 Anticipated heterogeneity - outcome of mean days per month in hospital

Investigation of heterogeneity formed part of the secondary objectives of the review. We hypothesised that the effect of Intensive Case Management on one of our primary outcomes (mean number of days per month in hospital) differs according to fidelity of intervention to the assertive community treatment model and the baseline level of hospital use.

The association of the IFACT score and the baseline number of days in hospital with the treatment effect was examined by performing random-effects meta-regression analysis in R (R 2008). The script we used to perform meta-regression analyses is reported in Appendix 5. Meta-regression was also carried out using both variables within the same model. The relationship between the treatment effect and the two variables was also examined using a thin plate spline. If possible data from multi-centre studies were to be entered in the meta-regression disaggregated into the component centre with outcome and fidelity data for each.

Meta-regression was performed:

1. only if at least ten studies per comparison are available (Higgins 2008)

2. all included studies were entered into the meta-regression. Comparison type was also tested as an additional regressor in the model.

2.2 Unanticipated heterogeneity - other outcomes

2.2.1 For outcomes other than the second primary outcome (not remaining in contact with psychiatric services)

If inconsistency was high this was reported and no exploration undertaken.

2.2.2 For outcome ’not remaining in contact with psychiatric services’

If inconsistency was high this was reported. First we investigated whether data had been entered correctly. Second, if data had been correct, the graph was visually inspected and studies outside of the company of the rest were successively removed to see if heterogeneity was restored. Should this occur with no more than 10% of the data being excluded, data were presented. If not, data were not pooled.

Should unanticipated clinical or methodological heterogeneity be obvious we simply stated hypotheses regarding these for future reviews or versions of this review. We did not anticipate undertaking analyses relating to these.

Sensitivity analysis

1. Implication of randomisation

We aimed to include trials in a sensitivity analysis if they are described in some way as to imply randomisation. For the primary outcomes we included these studies and if there was no substantive difference when the implied randomised studies were added to those with better description of randomisation, then all data were employed from these studies.

2. Standard care caseload

If data were available, a sensitivity analysis was undertaken testing how prone the primary outcomes were to change when trials comparing Intensive Case Management to standard community care caseload ≦ 20 were compared with trials comparing Intensive Case Management to standard community care caseload >20. If there was a substantial difference, we reported results and discussed them but continued to pool the data.

3. Assumptions for lost binary data

Where assumptions had to be made regarding people lost to follow-up (see Dealing with missing data) we compared the findings of the primary outcomes when we used our assumption compared with completer data only. If there was a substantial difference, we reported results and discussed them but continued to employ our assumption.

4. Assumptions for incomplete data for meta-regression

Where assumptions had to be made regarding missing SDs data in studies entering meta-regression (see Dealing with missing data), we compared the findings of the meta-regression on our primary outcome when we used our assumption compared with data taken from only those who completed the studies. A sensitivity analysis was undertaken testing how prone results from meta-regression were to change when data from those who completed were compared with imputed data using the assumption outlined above. If there was a substantial difference, we reported results and discussed them but continued to employ our assumption.

RESULTS

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification; Characteristics of ongoing studies.

Results of the search

1. Selection of studies (see Selection of studies)

When CBI inspected a random sample of study abstracts identified as above (see Search methods for identification of studies) comprising 10% of total abstracts the principal reviewer (MD) had inspected, disagreement did occur. Full articles were therefore acquired for further inspection. At this next stage MD and CBI had full agreement on the total sample of reports selected for further inspection, with the exception of Huang-China which now awaits assessment.

When MD and CBI in turn inspected all full articles of relevant reports and independently decided whether they met inclusion criteria (see Criteria for considering studies for this review), they fully agreed and no difficulties or disputes arose on any report. It was not necessary contact third reviewer (MM) for clarifying issues.

2. Data extraction (see Data extraction and management)

When CBI independently extracted data from a random sample of included studies (10% of total) disagreements were discussed. MD and CBI reached full agreement on final decisions.

3. Original Cochrane reviews (Marshall 2000a, Marshall 2000b)

3.1 ACT (Marshall 2000b)

Fourteen trials had met inclusion for the ACT versus standard care comparison. The trials included were Aberg-Wistedt-Sweden, Audini-UK, Bond-Chicago1, Bond-Indiana1, Chandler-California1, Hampton-Illinois, Herinckx-Oregon, Jerrell-SCarolina1, Lehman-Maryland1, Morse-Missouri1, Quinlivan-California, Rosenheck-USA, Solomon-Pennsylvania and Test-Wisconsin. All trials were included in the current review in the comparison ICM versus standard care.

For the ACT versus hospital-based rehabilitation comparison, three trials had been eligible for inclusion (De Cangas-Canada, Lafave-Canada, Marx-Wisconsin). All three have been excluded from the current update as they did not met inclusion criteria.

Six trials were included in the ACT versus case management comparison (Bush-Georgia, Essock-Connecticut1, Jerrell-SCarolina1, Morse-Missouri2, Quinlivan-California, Solomon-Pennsylvania). Of the original six studies described above, two were included in the ICM versus standard care comparison as both are three arm studies comparing ICM (in two arms) versus standard care (Jerrell-SCarolina1, Solomon-Pennsylvania). Three trials were included in the ICM versus non-ICM comparison (Bush-Georgia, Essock-Connecticut1, Quinlivan-California). Quinlivan-California was included in both comparisons ICM versus standard care and ICM versus non-ICM, as it had three arms comparing three differentiated interventions (ICM, non-ICM and standard care). The sixth trial, Morse-Missouri2, although previously included, was excluded from the current update because it contained no usable data, as number for treatment groups were not presented.

3.2 Case Management (Marshall 2000a)

Ten randomised controlled trials were included in the comparison of case management versus standard care (Curtis-New York, Ford-UK, Franklin-Texas, Jerrell-SCarolina1, Macias-Utah, Marshall-UK, Muijen-UK2, Quinlivan-California, Solomon-Pennsylvania, Tyrer-UK). Of these eight are now included (Curtis-New York, Ford-UK, Jerrell-SCarolina1, Macias-Utah, Marshall-UK, Muijen-UK2, Quinlivan-California, Solomon-Pennsylvania).Franklin-Texas and Tyrer-UK had to be excluded as they did not meet inclusion criteria on type of intervention.

4. 2009 search

The February 2009 update search of Cochrane Schizophrenia Group’s Register of trials yielded 2565 references. We selected 55 for further inspection. Of these 14 trials met the inclusion criteria and were included (Bjorkman-Sweden, Drake-NHamp, Essock-Connecticut2, Harrison-Read-UK, Johnston-Australia, Morse-Missouri3, Muller-Clemm-Canada, Okpaku-Tennessee, OPUS-Denmark, Pique-California, REACT-UK, Sytema-Netherlands, Salkever-SCarolina, UK700-UK). Thirty one trials were excluded. We added ten English language trials and five Chinese trials to those awaiting assessment and sought further information. Three trials which had been previously been awaiting assessment were able to be included as more reports had become available (Fekete 1998 now included as McDonel-Indiana, Holloway-UK, Shern-USA1).

4.1 Problematic trials

There are two problematic trials that are worth special mention. Jerrell-SCarolina1 is a three arm trial, with two of the arms qualifying as Intensive Case Management (Programme Assertive Community Treatment and Intensive Broker Case Management) and one a control (standard care). As results are reported separately for each arm, it was not possible to present continuous data from two ICM arms pooled together. One option was to treat each arm as a separate ’site’, effectively treating the study as two trials, but with the same control group. A second option was to include only one of the experimental arms. Although conscious of excluding potentially useful data on an arbitrary basis, we decided to include only one of the arms compared to standard care, according to the second option. The main reason for this was avoiding a unit of analysis error occurring in the first option. A sensitivity analysis was undertaken testing how prone results were to change when this trial was not included in meta-analysis.

Curtis-New York is a trial comparing ICM with standard care and presents two main difficulties to the reviewer. The first is regarding the ICM caseload size. Study reported caseload ratio as 1:35 (therefore above the 1:20 ratio defining an ICM intervention). As we derived estimation of caseload size by dividing the number of intervention participants by the number of whole time equivalent clinical staff in the team (not just those who are formally classified as ’case managers’), we found that the actual staff:participant ratio is about 1:17. We therefore stated this trial as eligible. The second issue is regarding the peculiar way this trial provides the ICM intervention. Both experimental and control interventions are community-oriented and they fully fit into the review’s inclusion criteria, but the ICM team is located in hospital. The fact that the team office is based in hospital is undesirable but not unusual. In any case, the case management was given to people in the community. We, therefore, confirmed its inclusion, aiming not to penalise it because it had reported details that are not available for all trials. We found, however, a discrepancy in data Curtis-New York provided on service use outcomes (average number of days in hospital per month, admitted to hospital). This study was an outlier, being the only one clearly favouring standard care over ICM. A sensitivity analysis was undertaken testing how prone results were to change when this trial was not included in the meta-analysis. Neither for the primary outcome ’average number of days in hospital’ nor ’admitted to hospital’ results changed significantly when Curtis-New York was dropped, but it did significantly affect the level of heterogeneity. We could just advance the hypothesis that reason for heterogeneity could be the unusual way the intervention was provided in this trial (Table 4).

Table 4. Interventions in Curtis-New York.

ICM: ‘Intensive outreach case management’ from a multi-disciplinary team at HHC. This implemented a discharge treatment plan, and monitored clinical and social problems. The team did not ‘assume direct responsibility for care but [..] help[ed] the patient enrol in a day hospital programme, adult mental health clinic, rehabilitation programme, or alcohol treatment programme’. Caseload: 1:17. N=147 Standard Care: routine aftercare, within the discharge treatment plan prescribed for each patient from Harlem Hospital Centre (HHC); “most received at least initial treatment form various divisions of the departments of psychiatry within the Health and Hospitals Corporation”. N=145

Included studies

The current update of previous two reviews includes 176 reports describing 38 studies. This review now includes data on 7328 randomised people from within these 38 separate trials. Twenty four of these studies had already been included in two original reviews (as included or awaiting assessment), with 14 more derived from the 2008 search. Twenty six trials provide data for the ICM versus standard care comparison, 11 for the ICM versus non-ICM comparison and only one for both comparisons. Please note that it was possible to report data for several separate centres of seven multi-centre trials (Bond-Indiana1 (3), Chandler-California1 (2) Drake-NHamp (7) Hampton-Illinois (2) McDonel-Indiana (2) Rosenheck-USA (9) and UK700-UK (4)). Several of these centres are reported separately in the table of Characteristics of included studies.

1. Study length

Only one study fell into the short term category, with a maximum length of six months (Bond-Indiana1); nine studies reported medium term data, between them only one reported data by seven months (Okpaku-Tennessee) as all the remaining eight reported data by 12 months (Bond-Chicago1; Bush-Georgia, Hampton-Illinois, Johnston-Australia, Lehman-Maryland1, Morse-Missouri1, Solomon-Pennsylvania, Sytema-Netherlands). The remaining 28 all fell into the long term category with a maximum length of four years (Test-Wisconsin) and an average length of 23.5 months.

2. Design

All included studies presented a parallel longitudinal design. Twelve were multi-centre trials, but only seven provided data for single centres (see above).

3. Participants

A total of 7328 participants are included from 38 trials. Most trials included from previous reviews were conducted in USA (16 trials, including 3474 participants) and only five were European (including 345 participants). In the current revision we added 1964 participants from seven trials conducted in Europe, and 1545 from ten studies conducted in Australia, Canada and the USA.

Nineteen studies included participants with severe mental illness. None provided operationalised definitions addressing dimensions of diagnosis, duration and disability. Thirteen, however, provided criteria for either the diagnosis, impairment or level of service use. Six studies did not provide any criteria for defining serious mental illnesses at all. Diagnoses of severe mental illness varied across studies from schizophrenic disorder alone, to wider diagnostic groups including schizophrenic disorder, affective and personality disorder. Of the remaining 19 studies, 17 involved participants with various diagnoses, but the great majority had some psychotic disorder and most trials reported criteria for service use and/or impairment. Two studies included participants with a high level of impairment or service use due to psychiatric illness, but not any diagnostic criteria for inclusion (Harrison-Read-UK, Okpaku-Tennessee). Most trials (22) involved participants that had been diagnosed using operationalised criteria (DSM, ICD, OPCRIT, RDC, SADS, see Characteristics of included studies footnotes) whilst 16 (ten in the group including participants with serious mental illnesses and five in the group of studies including people with various diagnoses) did not report using any diagnostic tool, but only stated type of illness or level of impairment. Only OPUS-Denmark included participants with a first episode of psychotic illness.

Four studies included a total of 742 dually diagnosed participants (Drake-NHamp, Essock-Connecticut2, Morse-Missouri3, Muller-Clemm-Canada) and eight 1337 homeless participants.

Data on mean age were available from 31 trials (6339 participants). The average age was about 38 years old. Only Macias-Utah did not report information on participant age.

4. Settings

As stated in the inclusion criteria, all the included studies took place in a community setting, provided both by private and public mental health services. No study was carried out in a low-income country, as all the included studies were from Europe or USA.

5. Interventions

Twenty seven trials were included in the comparison ICM versus standard care and 12 in ICM versus non-ICM comparison. Quinlivan-California was a three arm trial (ICM, non-ICM and standard care) and provided data for both comparisons. REACT-UK was considered in the ICM versus non-ICM comparison due to our assumption that standard care could be considered as Care Programme Approach, even if not clearly reported by trial authors. The Care Programme Approach was introduced in England in the mid-1990s and become standard care thereafter: it is a combination of non-Intensive Case Management and care from a Community Mental Health Team (CMHT), hence to be considered as non-ICM, according definitions in this review.

5.1 ICM

On average the ICM included in this review was well defined. The majority of experimental interventions were explicitly modelled on the assertive community treatment model, being based on the Treatment in Community Living model of Stein 1980. Only a few studies based ICM on the case management model. The experimental intervention was provided either by an already existing team or Intensive Case Management services newly established for the trial.

5.2 Non-ICM

There were no discernable differences in the practice of non-ICM and ICM except for the intensity of contact. Staff’s degree of training and skill were similar in the ICM teams and non-ICM teams. In some studies non-ICM was itself an experimental intervention, but mostly it represented the average standard care, as what in this review we call ’standard care’ increasingly shifted towards non-ICM across decades. Mental health systems increasingly included elements from ICM, melting them with community mental health service. English mental health policy is one example, where the Care Programme Approach was introduced in United Kingdom in the mid-1990s and become standard care thereafter. It is a combination of non-Intensive Case Management and care from a community mental health team (CMHT), hence to be considered as non-ICM, according definitions in this review. Therefore the REACT-UK study was considered in the ICM versus non-ICM comparison due to our assumption that standard care could be considered as Care Programme Approach, even if not clearly reported by trial authors.

5.3 Standard care

On average the standard care definition was blurred as this intervention was modelled on a generalist model. Its core was being provided by a community mental health service, but its features were variable across trials run in different countries at different time periods. Presence of further specialised services, such as rehabilitation or psychotherapist services, were variable within standard care services. In a few studies both ICM and standard care were incorporating services for substance abuse treatment and homelessness care.

6. Outcomes

Many trials used different scales in assessing treatment effects in various outcomes (global state, mental state - general and specific symptoms, social functioning, behaviour, quality of life and satisfaction). As most of the scales were used by only one study in each comparison group, it was not possible to enter these data in a unique analysis. Even where studies used the same scale, they often applied different rating scores. Therefore, again, data could not be entered together in the analysis. Some studies failed to clearly report the rating score they used for a pre-stated scale. This is noted in the Risk of Bias tables (see ’Outcomes’ in Characteristics of included studies table). No studies assessed improvement measuring it on scales. We did not calculate effect size measures (standardised mean difference SMD, see Measures of treatment effect).

6.1 Outcome scales: only details of the scale that provided usable data are shown below. Reasons for exclusions of data are given under ’Outcomes’ in Characteristics of included studies table

6.1.1 Global state

 

6.1.1.1 Global Assessment Scale (GAS) (Endicott 1976): in Audini-UK, Muijen-UK2, Rosenheck-USA

This is an observer-rated scale for evaluating the overall functioning of a person during a specified time period on a continuum from psychological or psychiatric sickness to health. Score ranges from 0 to 100, where higher score indicates a better outcome.

A modified version of the GAS was included in DSM-III as the Global Assessment of Functioning Scale (GAF) (APA 1987): in Bjorkman-Sweden. Outcomes from the two scales are reported together as GAF, as these two scales are very similar and they report results on the same score range.

6.1.1.2 Health of the Nation Outcome Scale (HoNOS) (Wing 1998; Stein 1999): in Harrison-Read-UK, REACT-UK

Provides a systematic summary of behaviours and functioning, measuring mental health and social/behavioural functioning. It consists of four areas (behaviour, impairment, symptoms and social), each assessed through 12 items on a five-point scale (0 to 4). Ratings are from 0 to 48; high score means severe dysfunction.

6.1.2 Global state: compliance with medication

 

6.1.2.1 Rating of Medication Influence (ROMI) (Weiden 1994): in Harrison-Read-UK, REACT-UK

This is a 20 item scale, measuring the influence of factor on medication adherence. Each item is rated according to the degree of influence on medication-taking behaviour: none (1), mild (2) and strong (3). It has 2 sub-scales - patient reported compliance (items 1-7) and patient-reported non compliance (items 8-20). A high score on the compliance sub-scale indicates high compliance, high score on the non-compliance sub-scale indicates high non-compliance. Results from the two studies are presented on different range score (in Harrison-Read-UK range score of 1 to 3; in REACT-UK range score is not clearly reported).

6.1.3 Social functioning

 

6.1.3.1 Disability Assessment Schedule (DAS) (WHO 2001a): in Holloway-UK

The WHO’s Psychiatric Disability Assessment Schedule (DAS) is a measure of functioning and disability. It contains 36 items with six domains of functioning including: understanding and communicating, getting around, self care, getting along with others, household and work activities and participation in society. Higher scores indicate a worse outcome.

6.1.3.2 Interview Schedule for Social Interaction - abbreviated version (ISSI) (Henderson 1980; Unden 1989): in Bjorkman-Sweden

The ISSI scale is a self-report scale, it consists of 30 items and measures social integration and attachment. The maximum score is 30 points and higher scores indicate a better social integration and attachment.

6.1.3.3 Life Skills Profile (LSP) (Rosen 1989; Parker 1992): in REACT-UK

The Life Skills Profile is a clinician-rated questionnaire that was developed in Australia primarily for use with people with psychotic illnesses. There are 39-item and five sub-scales that assess the general domain of disability over the last 3 months. The five sub-scales measure self-care, non-turbulence, social contact, communication and responsibility. For each of the 39 items on a scale ranging from ’not at all disabled’ to ’extremely disabled’.

6.1.3.4 REHAB Scale (REHAB) (Baker 1983): in Marshall-UK

An observer-rated measure of social functioning, covering social activity, self-care, speech disturbance and community skills. It rates the frequency of items of embarrassing or disruptive behaviour, such as violence, self harm, shouting and swearing, and sexual offensiveness (deviant behaviour - REHAB DB); and lack of general skills (general behaviour - REHAB GB) .The scale ranges from 0-144 with higher scores indicating poorer functioning.

6.1.3.5 Role Functioning Scale (RFS) (Green 1987): in Jerrell-SCarolina2

Self-report scale whereby the total of four sub-scales measures global role functioning. Higher scores indicate better functioning.

6.1.3.6 Social Adjustment Scale (SAS) (Weissman 1971, Weissman 1974): in Audini-UK, Muijen-UK2

Measures social functioning in a number of life domains (work, social, extended family, marital, parental, family unit, and economic adequacy). Score ranges from 1 to 7, with high score indicating poor outcome.

6.1.3.7 Social Adjustment Scale-II (SAS-II) (Schooler 1979): in Jerrell-SCarolina1

Revised version of the Social Adjustment Scale (see above), used to assess social adjustment. Self-reported scale, similar to SAS, but adapted for schizophrenia; it is composed of 24 items covering seven areas including social, family and work functioning. The scoring system appears to differ between the two versions, perhaps because this was an adapted version. High score indicates better outcome.

6.1.3.8 Social Functioning Questionnaire (SFQ) (Tyrer 1990, Tyrer 2005): in Harrison-Read-UK

An eight-item self-report scale (score range is 0 to 24). It provides a quick assessment of perceived social function. Higher score indicates poorer social functioning.

6.1.3.9 Strauss-Carpenter scale (Strauss 1972; Strauss 1974): in Bjorkman-Sweden

Strauss-Carpenter Outcome Scale assesses a total of 21 items exploring frequency of social contacts, employment duration, symptomatology, and duration of rehospitalisation. The scaling of each item extends from 0 (maximal negative) to 4 (maximal positive). The scoring range of the scale extends from 0 (maximal negative) to 84 (maximal positive).

6.1.4 Social functioning: substance abuse

 

6.1.4.1 Alcohol Use Scale (AUS) (Mueser 1995; Drake 1996): in Drake-NHamp

A five point scale based on clinicians’ ratings of severity of disorder, ranging from one (abstinence) to five (severe dependence).

6.1.4.2 Dartmouth Assessment of Lifestyle Interview (DALI) (Rosenberg 1998): in Sytema-Netherlands

An 18-item, interviewer-administered scale, addressing the detection of substance use disorder in people with severe mental illness. DALI focuses on alcohol, cannabis, and cocaine use disorders. DALI-alcohol: scores range from −4 to +6, higher scores indicate higher risk of alcohol abuse. DALI-drugs: scores range from −4 to +4, higher scores indicate higher risk of drugs abuse. As scale ranges from negative to positive value, skew is difficult to detect. Therefore data from this scale were entered in additional tables rather than into an analysis.

6.1.4.3 Substance Abuse Treatment Scale (SATS) (McHugo 1995; Drake 1996): in Drake-NHamp

An eight-point scale indicating progression toward recovery ranging from one (early stages of engagement) to eight (relapse prevention). Higher scores indicate greater progression.

6.1.4.4 Time-Line-Follow-Back (TLFB) (Sobell 1980): in Drake-NHamp

Scale administered by an interviewer to assess days of alcohol and drug use over the previous six months. Outcome reported as binary data.

6.1.5 Mental state: general symptoms

 

6.1.5.1 Brief Psychiatric Rating Scale (Overall 1962): in Audini-UK, Drake-NHamp, Muijen-UK2, REACT-UK, Sytema-Netherlands (BPRS 24-items - Velligan 2005); in Ford-UK, Rosenheck-USA (BPRS 18-item)

The BPRS is a scale measuring positive symptoms, general psychopathology and affective symptoms. The original scale has 16 items, but a revised 18-item scale is commonly used. There are several ways symptoms are reported (i.e. on a scale of 0 to 6 or a scale of 1 to 7), but the most common is that each item is rated on a seven-point scale (1 = not present to 7 = extremely severe). The 18-item scale might range from 18 to 126, or from 0 to 108 (as in Ford-UK and Rosenheck-USA). There is a further version, a 24-item scale, it ranges from 24-168. For all the scales high scores indicate more severe symptoms.

6.1.5.2 Brief Symptom Inventory (BSI) (Derogatis 1983): in Rosenheck-USA

A brief rating scale used by an independent rater to assess the severity of psychiatric symptoms. Scores range from 0-4 with higher scores indicating more symptoms.

6.1.5.3 Comprehensive Psychopathological Rating Scale (CPRS) (Asberg 1978): in Holloway-UK, UK700-UK

This is an interview rating scale covering a wide range of psychiatric symptoms, and can be used in total or as sub-scales. It consists of 65 items that covers the range of psychopathology over the preceding week (40 symptom items are rated by the participant and 25 observed items are rated by the rater during the interview). Each item is rated on a 0-3 scale varying from ’not present’ to ’extremely severe’, with high scores indicating more severe symptoms. High scores indicate a worse outcome.

6.1.5.4 Colorado Symptom Index (CSI) (Shern 1994): in Lehman-Maryland1, Shern-USA1

A brief rating scale used by an independent rater to assess the severity of a range of psychiatric symptoms. A lower score indicates more symptoms.

6.1.5.5 Krawiecka Scale (KS) alias Manchester Scale (Krawiecka 1977): in Harrison-Read-UK

This scale rates severity of psychiatric symptoms. It consists of eight categories of symptoms assessed on a five-point scale. These are depression, anxiety, hallucinations, delusions, flattened and incongruous effect, psychomotor retardation, incoherence and irrelevance of speech, and poverty of speech. A score of zero or one denotes an absence of pathology while ratings of two, three or four denote the presence of the target symptoms in increasing severity. Rating is from 0 to 36. Higher scores indicate a worse outcome.

6.1.5.6 Present State Examination (PSE) (Wing 1974): in Audini-UK, Muijen-UK2

This is a clinician-rated scale measuring mental status. 140 symptom items are rated and combined to give various syndrome and sub-syndrome scores. A short version, covering the first 40 “neurotic” symptoms, has been used in several population surveys. Score range from 1 to 120. Higher scores indicate greater clinical impairment.

6.1.5.7 SCL-90 (Hopkins symptoms check list) (Derogatis 1974): in Bjorkman-Sweden

This is a self-report clinical rating scale of psychiatric symptomatology containing 90 symptom related questions. It consists of 90 items, with 83 items representing nine sub-scales: somatisation (12 items), obsessive-compulsive (10 items), interpersonal sensitivity (9 items), depression (3 items), anxiety (10 items), anger-hostility (6 items), phobic anxiety (7 items), paranoid ideation (6 items) and psychoticism (10 items). Seven additional items include disturbances in appetite and sleep. The SCL-90 also utilises three global distress indices: Global Severity Index (GSI), Positive Symptom Distress Index (PSDI), Positive Symptom Total (PST). The participant assesses the degree of severity of each symptom. Items are rated on a 5-point Likert scale, ranging from “not at all distressing” (0) to “extremely distressing” (4), with higher scores indicating greater symptomatology.

6.1.6 Mental state: specific symptoms

 

6.1.6.1 Beck Depression Inventory (BDI) (Beck 1979): in Holloway-UK

A 21-item self-rating scale for depression. Each item comprises 4 statements (rated from 0 to 4) describing increasing severity of the abnormality concerned. Person completing the scale is required to read each group of statements and identify the one that best describes the way they have felt over the preceding week. Score ranges from 0 to 84, with higher score indicating more severe symptoms.

6.1.6.2 Hospital Anxiety and Depression Scale (HAD-S) (Zigmond 1983): Harrison-Read-UK

This scale is a questionnaire composed of statements relevant to either generalised anxiety or depression referring to the past week. The number of items in the questionnaire are seven reflecting anxiety, and seven reflecting depression. Each item is answered by the participant on a four point (0-3) response category, so the possible scores range from 0 to 21 for anxiety and 0 to 21 for depression. High score indicates a worse outcome.

6.1.6.3 Scale for the Assessment of Negative Symptoms (SANS) (Andreasen 1982, Andreasen 1989): in Holloway-UK, UK700-UK

This assesses five symptom complexes to obtain clinical ratings of negative symptoms in people with schizophrenia over the preceding week. They are: affective blunting, alogia (impoverished thinking), avolition/apathy, anhedonia/asociality and disturbance of attention. The final symptom complexes seem to have less obvious relevance to negative symptoms than the other four complexes. Assessments are conducted on a six-point scale (from 0 indicating ’not at all’ to 5 indicating ’severe’). High scores indicate a worse outcome.

6.1.7 Behaviour

 

6.1.7.1 Social Behaviour Schedule (SBS) (Wykes 1986): in Holloway-UK

The SBS is a 21-item scale designed to assess a range of areas of functioning in people with long term mental illness. The scale covers areas such as social behaviour and communication, self-care and inappropriate behaviour. The respondent’s behaviour on each item during the previous month is scored by someone familiar with him or her. Each item is rated on a five-point Likert scale (from 0 to 4), with higher scores indicating greater deficits.

6.1.8 Quality of life

 

6.1.8.1 Lancashire Quality of Life Profile (LQoLP) (Oliver 1996; Oliver 1997): in Bjorkman-Sweden, Holloway-UK, UK700-UK

A structured self-report interview with 105 items, combining objective and subjective measures in the following nine life domains (range of values 1-7): living situation, social relationships, work and education, legal status and safety, religion, family relations, leisure activities, finances and health. The LQoLP also measures the following additional areas: positive and negative affect (with the Bradburn Affect-Balance Scale), self-esteem, global well-being (Cantril’s Ladder and Happiness Scale), perceived quality of life and the quality of life of the patient independently of the patient’s own opinion (with the Quality of Life Uniscale). The measures from LQoP used in Bjorkman-Sweden were: overall quality of life (which is the mean of subjective quality of life in nine life domains), and global well-being. Higher score indicates better subjective quality of life/satisfaction.

6.1.8.2 Manchester Short Assessment of Quality of Life (MANSA) (Priebe 1999): REACT-UK, Sytema-Netherlands

A 16-item scale composed of 4 objective and 12 subjective questions. 12 subjective items are rated on a 7 point scale (from ’couldn’t be worse’ to ’couldn’t be better’, scored from 1 to 7, range 12-84) assessing satisfaction with life ’in general’, and in a range of domains such as vocational, financial, friendships, leisure, personal safety, physical health and mental health. Four objective items, answered yes or no, assess the existence of a close friend, contacts with friends per week, accusation of a crime and victimisation of physical violence. Higher score indicates better quality of life. In REACT-UK and Sytema-Netherlands score is reported as a mean ranging from 1 to 7.

6.1.8.3 Lehman’s Quality of Life Interview (QOLI) (Lehman 1988): in Drake-NHamp, Lehman-Maryland1, Shern-USA1; (Lehman 1993) in Ford-UK; (Lehman 1983) in Marshall-UK

The QOLI contains 153 items that measure global life satisfaction as well as objective and subjective quality of life, in eight life domains (living situations, daily activities and functioning, family relations, social relations and finances, work and school, legal and safety issues and health). Rated on a 7-point scale with higher scores indicating better quality of life. Subjective assessment of general life satisfaction ranges from 1 to 7 (terrible to delighted). Ford-UK reported objective quality of life and Marshall-UK reported subjective quality of life.

6.1.9 Client satisfaction

 

6.1.9.1 Camberwell Assessment of Need interview (CAN) (Phelan 1993): in Bjorkman-Sweden, Harrison-Read-UK, UK700-UK

The Camberwell Assessment of Need assesses the health and social needs of people with mental health problems. This scale assesses 22 areas to yield numbers of met and unmet needs as rated by the participant. Possible scores range from 0 to 22, with a higher score indicating poorer level of met needs.

6.1.9.2 Abbreviated Camberwell Assessment of Need Interview(CANSAS)(Slade 1999, Phelan 1995): in REACT-UK, Sytema-Netherlands

This is an abbreviated form of the above CAN.

6.1.9.3 Client Satisfaction Questionnaire (CSQ) (Larsen 1979): in Audini-UK; in Muijen-UK2, OPUS-Denmark, Sytema-Netherlands

The CQS is a self-report instrument designed to measure global patient’s satisfaction with services. Items are concerned with quality of services received, how well services met the client’s needs and general satisfaction. The CSQ is substantially correlated with treatment attrition, number of therapy sessions attended, and with change in client-reported symptoms. It consists of 8 items which are scored on 4-point Likert scales (1 to 4). Total score ranges from 8 to 32. Higher scores indicate greater satisfaction.

6.1.9.4 Client Satisfaction Questionnaire (CSQ) - modified version (Larsen 1979, Gerber 1999): in REACT-UK

The survey has 35 questions covering the location of services, services clients expect, delays in obtaining services, clients’ input into treatment, information received about drug treatment, satisfaction with treatment, access to clinical files, satisfaction with the therapist, family involvement in treatment, the treatment process, and overall satisfaction. Possible responses to most items range from 1 (most negative) to 7 (most positive). A rating of zero on certain items enables the respondent to indicate that the question was not relevant to his or her situation. Six items within the questionnaire, also on a 7-point scale, form a general satisfaction questionnaire. Higher score indicates greater satisfaction.

6.1.9.5 Patients’ satisfaction with health services (Tyrer 1979): in UK700-UK

A self-reporting questionnaire that rates nine components of satisfaction with services each on a four-point scale (1-4). Scores may range from 9 to 36, higher values indicating less satisfaction.

6.2 Missing outcomes

No trial reported or rated relapse, positive symptoms and carer satisfaction.

Excluded studies

We excluded 81 studies from the review. Half had been excluded by authors of the original Assertive Community Treatment and Case Management reviews and another 31 had to be excluded after the 2009 search.

Four trials included in the original reviews have been excluded in this update as they now do not match the new inclusion criteria (De Cangas-Canada, Franklin-Texas, Lafave-Canada, Marx-Wisconsin). Morse-Missouri2 had to be excluded in this update as it did not report the number of people randomised to each treatment group.

One further trial (Tyrer-UK), originally included in the CM review, was now excluded because of a methodological issue.

Tyrer-UK is a trial comparing ICM to standard care. The first issue was to clarify the ICM caseload. Thanks to further information provided from author, it was clarified there were 25 key workers in the service looking after 400 patients on the register. Therefore each key worker had a caseload of 1 to 16 or ’high intensity’ case management by this review’s definition. The second issue arose for the case managers in the treatment group being also workers in the control group. The problem was therefore one of contamination, as requiring someone to carry out close monitoring of one participant in the treatment group, that could affect their care of a similar participant in the control group in an unpredictable way. We decided to treat this as an excluded study on the grounds that we cannot be sure that high intensity case management was really being compared with standard care.

Three trials, which had been awaiting assessment in the old Cochrane reviews, were also excluded for different reasons (Jerrell-California, Godley-Illinois, Mulder-Missouri, more details are given in Characteristics of excluded studies).

Jerrell-California had been a partially published trial and had been previously classified as ’awaiting further assessment’ because information was required on number of people excluded after randomisation (participants were excluded if they refused to participate after randomisation or if they had not been discharged from hospital within 6 months of entering the study). As these data did not become available, the trial has now been excluded.

Godley-Illinois was an unpublished two-centre trial and it was initially classified as ’awaiting further assessment’ because it was not possible to determine if the intervention was ACT or CM. During the current update the intervention was classified as ICM versus standard care. It had to be excluded because it contained no usable data due to incomplete data reporting, and no further information became available (i.e. there was an apparent error in the reporting of numbers admitted to hospital; in one table admission rates are reported as: 31/52 experimental group and 33/45 control group; in another table admission rates are reported as: 31/52 experimental and 25/45 control).

Mulder-Missouri was a report of data from randomised and non-randomised participants. These data were not reported separately and, in addition, the intervention also did not fit inclusion criteria (ICM is compared to acute hospital admission).

Overall, 27 studies were excluded because they were not randomised or because randomisation was compromised (Jerrell-California). Five studies had to be excluded as participants were requiring immediate hospital admission (Fenton-Canada, Hoult-Australia, Muijen-UK1, Mulder-Missouri, Stein-Wisconsin), one (Martin-UK) because participants were dually diagnosed with intellectual disability and mental illness, and one more because the majority of participants were simply homeless and not clearly ill (Toro-New York). Most trials had to be excluded because of the intervention: 26 because the experimental intervention was not ICM. Modcrin-Kansas had to be excluded as caseload was not reported in either experimental or control group. Ten trials were excluded as the comparison intervention was not standard care nor non-ICM. Ten trials were excluded as the intervention administered to the experimental group was not only ICM (Chandler-California2, COAST-UK, Cosden-California, Gold-SCarolina, Grawe-Norway, Lehman-Maryland2, LEO-UK, McHugo-Washington DC, Shern-USA2, Shern-USA3). Finally two trials, Godley-Illinois and Morse-Missouri2, were excluded because no usable data could be extracted from the study report (as previously explained).

1. Awaiting classification

Fifteen trials, of which five are in the Chinese language are awaiting classification until further information is obtained (Agius-Croatia, Dick-UK, Johnson-UK, Kane-Virginia, Klotz-California, Linszen-Netherlands, O’Donnell-Australia, Rivera-New York, Sells-Connecticut, Verhaegh-Netherlands, Guo-China, Huang-China, Li-China, NCT00781079, Tan-China).

2. Ongoing studies

One of the two trials described as ’ongoing’ in previous reviews has now been included (UK700-UK), while the second trial (Hu-California) has not been included as a trial itself, but it has been moved to reference both Chandler-California1 and Jerrell-SCarolina1. We are aware of only one study currently ongoing (Walsh-Connecticut).

Risk of bias in included studies

For multi-centre trials provided data for individual single centres, we did not assess the risk of bias for each centre. Our judgments regarding the overall risk of bias in individual studies is illustrated in Figure 1.

Figure 1.

Methodological quality summary: review authors’ judgements about each methodological quality item for each included study.

Allocation

All 38 studies were stated to be randomised, but only ten provided descriptions of the methods used to generate the sequence. These studies were therefore classified as of high quality with a low risk of selection bias. The most common method was random allocation according to a sequence of random numbers generated by a computer programme (Bjorkman-Sweden, Essock-Connecticut2, Ford-UK, Harrison-Read-UK, OPUS-Denmark in one of two sites). Three trials used permuted block (Marshall-UK, REACT-UK, Sytema-Netherlands), one used table of random permutation (Pique-California) and one coin tossing (Rosenheck-USA). In one of the two sites of OPUS-Denmark trial (Aahrus site) allocation was performed by drawing lots - from among five red and five white lots from a black box. Overall, however, most studies are classified as of unclear quality with a moderate risk of selection bias and an overestimate of positive effect, as no description of the methods used to generate the sequence had been provided.

Regarding the allocation concealment, only four studies were rated as high quality as they provided descriptions of the methods used to conceal random allocation (OPUS-Denmark, REACT-UK, Sytema-Netherlands, UK700-UK). In all four cases it was a centralised allocation carried on by telephone, fax or mail. All remaining thirty-three studies are classified as of unclear quality with a moderate risk of selection bias and overestimate of positive effect.

Blinding

We classified blinding in respect only to primary outcomes. Due to intervention characteristics, i.e. being a model of service organisation, we assumed participants and clinicians being implicitly not blind to treatment assignment. We also assumed that primary outcomes were likely to be influenced by participant and clinician lack of blinding, as the knowledge of treatment allocation could determine both performance and attrition bias at a level which is difficult to predict/quantify. Whereas we did not consider the primary outcomes as interviewer mediated, hence we assume that lack of interviewer blinding would produce less detection bias. Therefore all studies providing primary outcome data are classified as of unclear quality with a moderate risk of performance and attrition bias. This gathers further potential for overestimate of positive effects and underestimate of negative ones.

We report blinding to secondary outcomes in the risk of bias table but we do not account for it in the global rating of the study blinding risk of bias. Again, if the secondary outcome was clinician/participant mediated, we rated it as unclear. If it was interviewer rated, we assessed it according to information provided in the study. Only Shern-USA1 was rated at high risk of bias, as it provided only secondary outcome data and was only interviewer mediated. It was therefore possible to assess its’ risk of bias with higher confidence.

Incomplete outcome data

Where information were available, we assessed incomplete outcome data separately for primary and secondary outcome and presented both assessments in Figure 1. We, however, only rated the risk of bias in respect to primary outcome. Only three trials provided information for incomplete primary and secondary outcome data apart, and the risk of bias could be assessed separately (Holloway-UK, Johnston-Australia, REACT-UK). Nine trials were judged as adequately addressing incomplete outcome data and rated as low risk of attrition bias. Four of these were so rated because there were no missing outcome data (Bush-Georgia, Holloway-UK, REACT-UK, Sytema-Netherlands). Three more did make the number and reason for missing data explicit as well as balanced across groups (Audini-UK, Essock-Connecticut1, Johnston-Australia). The last two because they undertook intention to treat analysis (OPUS-Denmark, UK700-UK).

Bond-Chicago1 and Ford-UK were judged as high risk of attrition bias because, although clearly reporting number and reason for missing data, reasons for missing outcome data were likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention group. However, our protocol somewhat compensated for this (see Dealing with missing data) and despite the high rating, data from these studies remains included.

The remaining twenty-seven trials were rated as ’unclear’ with a moderate risk of attrition bias. They either did not address this issue or presented insufficient information of attrition/exclusions to permit judgment (i.e. no reasons for missing data provided or number of randomised not stated. Jerrell-SCarolina1 reported only number of randomised participants completing the study period).

Some specific examples may serve to illustrate the difficulty in rating this issue.

Essock-Connecticut2 was not an intention to treat analysis - seven participants were excluded from the study straight after randomisation because they were lost to follow up, but authors failed to provide information on what intervention those participants had been allocated and reason for leaving the study early. As this study was providing only continuous outcome data, we reproduced completer-only data which were reported. No action was undertaken to deal with other missing data.

Macias-Utah had three problems. First, the study was not an intention to treat analysis - seven participants were excluded from the study because they were lost to follow up. The authors of this study broke with usual practice by failing to provide any data on participants lost to follow up, in particular, data on admissions to hospital. Second, one participant, (presumably randomised to the treatment group) was excluded after randomisation, having refused to participate (again, it was not clear whether this person had been admitted to hospital). Third, five further participants were added (randomly) to the treatment and control groups part way through the study, some as late as ’late 1990’ (final assessments took place in February 1991). No further information has become available since first review publication, which potentially could substantially affect findings.

In Morse-Missouri1 28 further participants were added (randomly) to the initial randomised sample, to replace participants leaving the study within the first month after entering. As replacement was carried out through randomised assignment, reviewers did not raise any question on the replacement issue and the study was included. We presented data from the final sample, obtained after randomised replacement occurred.

In Muller-Clemm-Canada the number randomised was not clearly reported, as authors declared that “Clients who withdrew from the study within the first 6 months were replaced by other clients”. Finally Sytema-Netherlands randomised 119 participants, but one was excluded because moved to another area directly after randomisation (We performed intention to treat analysis on remaining 118 participants).

Selective reporting

Most of the trials (23) were rated as of low quality with a high risk of reporting bias as they presented data in reports in a way we could not consider as free of suggestion of selective outcome reporting (i.e. pre-specified outcomes were not reported or reported incompletely so that they could not be entered in the analysis, or reported outcomes not pre-specified). Thirteen studies were rated as of high quality in reporting outcome with a low risk of reporting bias. For two studies the risk of bias was assessed as unclear with a moderate risk of reporting bias.

Other potential sources of bias

The presence of other potential threats to validity was unclear in only Hampton-Illinois. In this trial it was unclear whether the study was interrupted early in one of the two centres. Therefore the risk of other potential sources of bias was rated as moderate in this trial.

All remaining trials were rated as low risk of other potential sources of bias, as no evidence of other bias occurred. Most were publicly funded. No declaration of interest was made by authors, and we assume there was none to be made. However, many study authors were active pioneers of developing and implementation of the experimental intervention model across the scientific community and clinical world. This raises the issue on how researcher beliefs could affect the entire process of evaluating an intervention in a randomised clinical trial. Although conscious of this issue, we decided not to make any attempt in rating it as it is very difficult to judge, and erroneous quantification could drive bias into our conclusions.

Effects of interventions

See: Summary of findings for the main comparison INTENSIVE CASE MANAGEMENT compared with STANDARD CARE for severe mental illness; Summary of findings 2 INTENSIVE CASE MANAGEMENT compared to NON-INTENSIVE CASE MANAGEMENT for severe mental illness

The nine main indices of outcome were: i. service use; ii. adverse effects; iii. global state; iv. social functioning; v. mental state; vi. behaviour vii. quality of life; viii. satisfaction; and ix. direct costs. Each index will be considered in turn for each of the two comparisons.

1. COMPARISON 1: INTENSIVE CASE MANAGEMENT versus STANDARD CARE (Summary of findings for the main comparison)

Primary outcomes

1.1 Service use

1.1.1 Service use: Average number of days in hospital per month - at about 24 months

Data were available from five studies presenting skewed data from a sample size greater than or equal to 200 participants and from nineteen trials reporting skewed data from sample sizes less than 200. We entered these data in separate subgroups.

In the first sub-group analysis (i.e. skewed data from studies with sample size greater than or equal to 200 participants) we found no significant difference in length of hospitalisation per month (n=1812, 5 RCTs, MD −0.46 CI −0.95 to 0.03), although data suggested a trend favouring ICM (p=0.06). In the second sub-group analysis (i.e. skewed data from study sample size less than 200 participants) there was a significant difference between groups, favouring the ICM group in reducing length of hospitalisation (n= 1783, 19 RCTs, MD −1.01 CI −1.74 to −0.28), but these data were heterogeneous (I²=77%, p<0.00001).

When synthesising data from the two sub-groups, we found that the length of hospitalisation was significantly reduced in the ICM group (n=3595, 24 RCTs, MD −0.86 CI −1.37 to −0.34), but the level of heterogeneity was high (I²=74%, p< 0.00001, Figure 2). We investigated the heterogeneity by checking again for correctness of data and removing one outlier study from the analysis (Curtis-New York), as it was the only one favouring standard care. By excluding Curtis-New York the level of heterogeneity was still high (I²=59%, p< 0.0002). Therefore we removed the second most outlier study from the analysis (one of three centres from a multi-centre study, Bond-Indiana1 (A)) as this was the most extreme result (favouring ICM). By excluding Bond-Indiana1 (A) data remained significant, favouring ICM (n=3245, 22 RCTs, MD −0.79 CI −1.22 to −0.36). The heterogeneity was reduced to just within our cut off point (I²=49%, p=0.005). Removing two further outliers (Bond-Indiana1 (C), Quinlivan-California) reduced heterogeneity still further (I²=36%, p=0.05) as well as the overall estimate, but ICM still seemed to significantly decrease time in hospital (n=3143, 20 RCTs, MD −0.62 CI −1.00 to −0.23), Figure 3).

Figure 2.

Forest plot of comparison: 1 INTENSIVE CASE MANAGEMENT vs STANDARD CARE, outcome: 1.1 Service use: 1. Average number of days in hospital per month - at about 24 months.

Figure 3.

Service use: 1. Average number of days in hospital per month - at about 24 months - restoring homogeneity - 4 studies removed from analysis.

No substantial reporting biases were highlighted when investigated through visual inspection of funnel plot (Figure 4). Two studies - Bond-Indiana1 (A) and Quinlivan-California - seemed most heterogeneous (see above).

Figure 4.

Funnel plot of comparison: 1 INTENSIVE CASE MANAGEMENT vs STANDARD CARE, outcome: 1.1 Service use: 1. Average number of days in hospital per month - by about 24 months.

Meta-regression was run on 52 trials providing data for primary outcome ’Average number of days in hospital per month - at about 24 months’ (combining data from all ICM studies within Comparison 1 and 2). Within the meta-regression we found that i. the more ICM is adherent to the organisation model, the better it is at decreasing time in hospital (’organisation fidelity’ variable coefficient −0.36 CI −0.66 to −0.07, Figure 5); and ii. the higher the baseline hospital use in the population, the better ICM is at decreasing time in hospital (’baseline hospital use’ variable coefficient −0.20 CI −0.32 to −0.10, Figure 6). Combining both these variables within the model, ’organisation fidelity’ is no longer significant (regression coefficient −0.24 CI −0.52 to 0.04, p=0.089), but ’baseline hospital use’ result is still significantly influencing time in hospital, although it seems to lose some of its potency (regression coefficient −0.18 CI −0.29 to −0.07, p=0.0027) (Figure 7). Figure 7 shows the interaction of the two variables on study outcome graphically through the use of thin plate spline modelling. The plot provides a locally weighted two dimensional representation of the collinearity between the variables used in the regression.

Figure 5.

Meta-regression: Scatter plot of IFACT organisation sub-score v mean days per month in hospital Intensive case

Figure 6.

Meta-regression: Scatter-plot of mean baseline days in hospital v mean days per month in hospital

Figure 7.

Weighted thin plate spline regression showing combined effect of baseline days in hospital and Organizational Fidelity Score on Treatment effect

1.1.2 Service use: Not remaining in contact with psychiatric services

We included only one short term study and found no significant difference between treatment group (n=95, 1 RCT, RR 0.54 CI 0.28 to 1.05).

Medium-term data were available from three studies showing a significant difference between treatment groups, favouring the ICM group, where participants had a lower risk of not remaining in contact with psychiatric services compared with participants in the SD group (n=1063, 3 RCTs, RR 0.51 CI 0.36 to 0.71).

Six long term studies data confirmed this trend favouring ICM, showing a significant advantage for the ICM group (n=653, 6 RCTs, RR 0.35 CI 0.18 to 0.68), but data were heterogeneous (I²=63%, p=0.02). Herinckx-Oregon seemed to be the sole cause of this (n=475, 5 RCTs, RR 0.27 CI 0.11 to 0.66, I²=44%, p= 0.13) and on further consideration, post hoc, we think we were in error to include the outcome from this study because it was defined so differently from the other trials. Herinckx-Oregon did not include refusing to re-interview, moving out, and death - as all the other studies had done - and it was impossible to amend this at this stage. We therefore feel justified in removing this study altogether from this part of the review outcomes.

Overall, when pooling together studies from different time sub-groups, we found a significant advantage in the ICM group, where people were less likely to be lost to psychiatric services than people in the standard care group (n=1633, 9 RCTs, RR 0.43 CI 0.30 to 0.61, I²=49%, p=0.05, Figure 8).

Figure 8.

Forest plot of comparison: 1 INTENSIVE CASE MANAGEMENT vs STANDARD CARE, outcome: 1.2 Service use: 2. Not remaining in contact with psychiatric services by short, medium, long term and overall.

We did not use a funnel plot for this outcome as there were less than ten studies (see Assessment of reporting biases).

Secondary outcomes

1.1.3 Service use: Admitted to hospital across time

Data were available from two short term studies showing no significant differences between treatment groups (n=244, 2 RCTS, RR 0.61, CI 0.22 to 1.69), but these data were heterogeneous (I²= 81%, p=0.02).

We found medium term data in five studies and that these favoured the ICM group which had less admission to hospital across time compared with standard care (n=1303, 5 RCTs, RR 0.85 CI 0.77 to 0.93).

Long-term data were provided by eleven studies. As in the short term data, they showed no significant differences between treatment groups (n=1516, 11 RCTs, RR 0.96 CI 0.74 to 1.23), but these data were heterogeneous (I²=70%, p=0.0003). Only one study reported data by long term, but referring to the number of admissions across the previous year. This, therefore, could not be entered in the long term data sub-group analysis. It showed a significant effect favouring the ICM group (n=547, 1 RCT, RR 0.67 CI 0.52 to 0.86), therefore not consistent with long term data showed above. As these findings were based on data from one study only, they are considered less robust than those available from the eleven long term studies.

1.1.4 Service use: Hospital admission across time

Data were available from one medium term and three long term studies, describing the average number of admissions per month. All these data were skewed and did not enter the analysis. Data by the medium term study suggested a trend favouring the ICM group. Data from long term studies did not show any trend favouring one specific group over the other. Audini-UK and Muller-Clemm-Canada did not report variance measurements. We assumed consistency between studies and used the fully reported variance for Sytema-Netherlands and employed these data for Audini-UK and Muller-Clemm-Canada as well.

1.1.5 Service use: Use of Emergency Room

Data were available from three studies: one providing binary data as ’number admitted to ER’ (Emergency Room) and two providing continuous data as ’average number of admission to ER’. The first study describing ’number admitted to ER’, showed a non significant difference between two group (n=178, 1 RCT, RR 1.13 CI 0.72 to 1.76). The two studies describing the average number of admission to ER reported skewed data. Skewed data were not consistent as one did not show any trend in the direction of effect and the other showed a trend favouring the ICM group. As in one study (Jerrell-SCarolina1) the variance measurement was not reported in the study, the SD was carried over from the other available study (Lehman-Maryland1).

1.2 Adverse events

1.2.1 Adverse event: Death all causes

By the short term, two deaths occurred in the 81 people treated with ICM compared with two in the 80 people treated with standard care (n= 161, 2 RCTs, RR 1.04 CI 0.16 to 6.91).

We found the same result by medium and long term. By medium term five deaths occurred in 453 people treated with ICM compared with six deaths in 448 people treated with standard care (n= 901, 6 RCTs, RR 0.78 CI 0.23 to 2.62). By long term twenty-four deaths occurred in 741 people treated with ICM compared with twenty-seven deaths in 715 people treated with standard care (n=1456, 9 RCTs, RR 0.84 CI 0.48 to 1.47).

1.2.2 Adverse event: Suicide

We found similar results in mortality due to suicide as we found in mortality due to all causes. Data by short term were available from two studies, where no suicides occurred in 62 people treated with ICM compared with two suicides in 65 people treated with standard care (n=127, 2 RCTs, RR 0.35 CI 0.04 to 3.27). Data by medium term were available from four studies where two suicides occurred in 412 people treated with ICM, compared with two suicides in 407 people treated with standard care (n=819, 4 RCTs, RR 0.98 CI 0.17 to 5.60). Finally data by long term were available from nine studies, where ten suicides occurred in 741 people treated with ICM, compared with fourteen suicides in 715 people treated with standard care (n=1456, 9 RCTs, RR 0.68 CI 0.31 to 1.51).

1.3 Global state

1.3.1 Global state: Leaving the study early

We included five short term studies and found no significant differences between treatment groups for number of participants leaving the study early (n=598, 5 RCTs, RR 0.74 CI 0.41 to 1.33), but data were heterogeneous (I²=86%, p=0.0001).

We included eight medium term studies and found the risk of leaving the study early was lower for participants in the ICM group (n=1699, 8 RCTs, RR 0.60 CI 0.51 to 0.70).

Data from thirteen long term studies confirmed data by medium term studies, showing a significant advantage for ICM (n=1798, 13 RCTs, RR 0.68 CI 0.58 to 0.79).

1.3.2 Global state: GAF average endpoint score

We found GAF score favoured ICM during short term assessment in four studies (n=797, 4 RCTs, MD 2.07 CI 0.28 to 3.86) and during long term assessment in 5 studies (n=818, 5 RCTs, MD 3.41 CI 1.66 to 5.16). Medium-term GAF data from 3 studies (n=722, 3 RCTs, MD 0.09 CI −3.11 to 3.28) were equivocal, not showing any significant difference between groups.

1.3.3 Global state: Not compliant with medication

We found data only from one long term study, favouring the ICM group (n=71, 1 RCT, RR 0.35 CI 0.15 to 0.81).

1.4 Social functioning

1.4.1 Social functioning: Contact with legal system

For short term outcomes we found data only from one study, describing the outcome ’contact with the police’. This study did not reveal any significant difference in the rate of contact with the police between treatment groups (n=61, 1 RCT, RR 2.57 CI 0.73 to 9.04).

For medium term outcomes we found three studies, describing the outcome ’number of arrested’. They failed to show a significant difference between the two intervention groups (n=604, 3 RCTs, RR 1.08 CI 0.61 to 1.90). Only one medium term study was available providing data on ’contact with the police’. These data favoured the ICM group in reducing the number of contacts with the police (n=88, 1 RCT, RR 0.23 CI 0.09 to 0.55). We found four medium term studies describing ’number of imprisoned’ and we found no significant advantage for ICM group (n=361, 4 RCTs, RR 0.80 CI 0.39 to 1.64).

For long term outcomes we found data from one study, describing the outcome ’number of arrested’ and it showed no significant advantage for ICM (n=178, 1 RCT, RR 0.66 CI 0.32 to 1.37). We found also four long term studies reporting data on ’number of imprisoned’, again not showing any significant advantage for ICM in reducing the number of participants imprisoned by long term (n=361, 4 RCTs, RR 0.72 CI 0.31 to 1.67).

1.4.2 Social functioning: Employment status

We found medium term data for various outcomes. One study reported data on ’not competitively employed at the end of trial’ and these data did not show a significant advantage for ICM in improving the number of competitively employed people (n=88, 1 RCT, RR 1.00 CI 0.91 to 1.10). We found four medium term trials reporting the same outcome also failing to show a significant difference, although data did suggest a trend favouring ICM (n= 1136, 4 RCTs, RR 0.89 CI 0.79 to 1.00), but heterogeneity was present (I²=75%, p=0.008).

The same outcome ’number of not employed at the end of the trial’ was described in four long term studies. As in the medium term comparison, data again failed to show a significant difference,although they suggested a trend favouring the ICM group (n= 1129, 4 RCTs, RR 0.70 CI 0.49 to 1.00), but, again, there is considerable heterogeneity (I²=94%, p<0.00001).

1.4.3 Social functioning: Accommodation status

We found data on outcome ’homelessness’ from one short term study. This small study revealed a significant reduction in the rate of homelessness in the ICM group (n=95, 1 RCT, RR 0.04 CI 0.00 to 0.70).

Medium-term data on ’homelessness’ outcome were available from one small study. Data did not reveal any significant difference between groups in the rate of homelessness by the medium term (n= 88, 1 RCT, RR 0.32 CI 0.03 to 2.95). Medium-term data from five trials were also available on ’not living independently’ outcome. These data showed a significant advantage for ICM in reducing the number of not living independently (n=1303, 5 RCTs, RR 0.80 CI 0.66 to 0.97).

Data on the accommodation status were available by long term, describing different outcomes. We found ’homelessness’ data in three long term studies, not revealing any significant difference between intervention groups (n=418, 3 RCTs, RR 0.78 CI 0.34 to 1.82).

We found four studies providing data on ’not living independently’. These data favoured the ICM group where the incidence of not living independently was lower compared with the standard care group (n=1185, 4 RCTs, RR 0.65 CI 0.49 to 0.88).

The outcome ’not living in stable accommodation’ was available only from one study. We found that data favoured the ICM group in reducing the number of not living in stable accommodation (n=168, 1 RCT, RR 0.80 CI 0.65 to 0.98).

Skewed data were available on ’average days in stable accommodation’ and ’average days per month in sheltered homes’. Medium-term data on ’average days in stable accommodation’ showed a trend favouring the ICM group, therefore consistent to results previously described on ’not living in stable accommodation’ by long term.

Long-term data on ’average days per month in sheltered homes’ were provided by two studies. These data were equivocal as data from one study favoured ICM whilst data from the second study favoured the standard care group.

1.4.4 Social functioning: Substance abuse

Long-term data were available from one study, reporting alcohol and illicit drug abuse. Data failed to show a significant difference both in alcohol (n=547, 1 RCT, RR 0.55 CI 0.26 to 1.17) and illicit drug abuse (n=547, 1 RCT, RR 0.96 CI 0.63 to 1.47) between groups.

We were unable to enter two types of data into the analysis: medium term data assessing alcohol and drug abuse on DALI scale, and other skewed data. As the DAS scale averages values from positive to negative, skew is very difficult to detect, we did not enter these data in the analysis. These data tended to favour the standard care group both on alcohol and drug abuse outcome. Medium and long term data from the outcome ’days of substance use per month’ were equivocal, but skewed.

1.4.5 Social functioning: Average endpoint score (various scales)

Three studies provided data from five different scales (DAS, ISSI, RFS, SAS-adapted version, Strauss-Carpenter scale) assessing social functioning by short, medium and long term. As no more then one study per time period used the same scale, we did not enter more than one study per sub-group. Data from each available time-period failed to show any significant difference between treatment groups.

Data by short term were available on the RFS (n=80, 1 RCT, MD 0.62 CI −0.99 to 2.23) and SAS-adapted scale scores (n=80, 1 RCT, MD 3.34 CI −0.87 to 7.55).

Data by medium term were available on the DAS (DAS scale: n= 55, 1 RCT, MD 0.10 CI −0.40 to 0.60), RFS (n=80, 1 RCT, MD 0.86 CI −1.00 to 2.72), and SAS-adapted scale scores (n=80, 1 RCT, MD 3.30 CI −1.23 to 7.83).

Data by long term were available on the DAS (n=58, 1 RCT, MD −0.20 CI −0.67 to 0.27), ISSI (n=62, 1 RCT, MD −3.20 CI −6.29 to −0.11), RFS (n=80, 1 RCT, MD 2.35 CI 0.65 to 4.05), and SAS-adapted (n=80, 1 RCT, MD 2.75 CI −1.63 to 7.13), and Strauss-Carpenter scale scores (n=60, 1 RCT, MD −0.10 CI −1.37 to 1.17). Skewed data on SAS scale score were available by short, medium and long term from two studies. These data were equivocal and not consistent between two studies. Long-term data on REHAB scale score were provided by one study, these data tended to favour the ICM group but were also skewed.

1.5.1 Mental state: General symptoms

Two sets of data were available: i. non skewed data or skewed data from a sample size greater than or equal to 200 participants per study: entering analysis together; ii. skewed data: not entering analysis.

1. Short-term data were available from different scales. Short-term data mental state scores assessed with BPRS scale were not significantly different between two groups (n=668, 2 RCTs, MD −1.56 CI −6.85 to 3.73), but data were heterogeneous (I²=92%, p=0.0004). Short-term mental state scores assessed with BSI scale were available from the same two studies providing BPRS short term data. Again data were not significantly different (n=668, 2 RCTs, MD −0.06 CI −0.19 to 0.06), but, different from the BPRS results, these data were homogeneous. One further trial was available providing short term data on mental state assessed with CSI. These data show a significant difference favouring the ICM group (n=125, 1 RCT, MD 0.56 CI 0.28 to 0.84).

   Medium-term data were available from different scales. Mental state scores assessed with BPRS and BSI scales were both not significantly different between the two groups (BPRS scale: n=662, 2 RCTs, MD −0.96 CI −2.42 to 0.51; BSI scale: n=662, 2 RCTs, MD −0.02 CI −0.15 to 0.10), but medium term data assessed with CSI favour the ICM group (n=125, 1 RCT, MD 0.35 CI 0.05 to 0.65).

   We found significant results in mental state by long term: data were available from two studies assessing mental state scores with BPRS and BSI scale. We found in both sub-groups data favouring the ICM group, where people reached a better mental state score by long term (BPRS scale: n=647, 2 RCTs, MD −2.65 CI −4.11 to −1.20; BSI scale: n=647, 2 RCTs, MD −0.18 CI −0.31 to −0.06). Finally one study assessed the mean change from baseline on CSI scale, data showed no difference between groups (n=168, 1 RCT, MC −0.32 CI −0.53 to −0.11).

2. Skewed data were available from six studies assessing mental state by different time periods and with different scales (BPRS-18 items, BPRS-24 items, CPRS, PSE, SCL-90). Data failed to show any significant trend favouring one group over the others, this report being consistent across different studies and different rating scales. These data were considered not robust as they were skewed, but were in accord with short and medium term data, but not long term data.

1.5.2 Mental state: Specific symptoms

We found data on specific symptoms from only one study. This study provided data on depression incidence and there was no significant difference between groups by medium and long term (by medium term: n=547, 1 RCT, RR 0.77 CI 0.56 to 1.04; by long term: n=547, 1 RCT, RR 0.83 CI 0.57 to 1.21).

We found a second small study (n=70) providing skewed data on depression symptoms assessed with BDI scale, and negative symptoms assessed with SANS scale. Neither set of data is entered into the analysis. Skewed depression scores favoured the ICM group at medium and long term, whilst skewed negative symptoms scores by medium and long term were equivocal.

1.6 Behaviour: Specific - self harm

Medium and long term data showed there were no significant differences in number of participants who committed self-harm between groups: by medium term 30 events occurred in 312 people treated with ICM compared with 30 events in 308 people treated with standard care (n=620, 2 RCTs, RR 0.99 CI 0.61 to 1.59); by long term 11 events occurred in 190 people treated with ICM compared with 13 events in 184 people treated with standard care (n=374, 2 RCTs, RR 0.82 CI 0.38 to 1.78). These data were confirmed from long term data on self-harm during the previous twelve months available from a single study (n=547, 1 RCT, RR 0.81 CI 0.47 to 1.38).

Skewed data were available from one small study (n=70) assessing behaviour with SBS scale by medium and long term. These data tended to favour the ICM group.

1.7 Quality of life: Average endpoint score (various scales)

We found six studies assessing quality of life with various scales by different time periods.

The only significant result we found was by short term: data were available from a single study and showed a significant higher quality of life in the ICM group as assessed on the QOLI general-well-being sub-scale (n=125, 1 RCT, MD 0.53 CI 0.09 to 0.97). Medium-term data assessing quality of life with LQoLP (one study) and MANSA (one study) did not show a significant difference between groups (LQoLP scale n=52, 1 RCT, MD 0.09 CI −0.60 to 0.78; MANSA scale n=81, 1 RCT, MD 0.20 CI −0.29 to 0.69).

Long-term data on quality of life were assessed with LQoLP (two studies) and QOLI (two studies). As with the medium term data, results did not suggest significant differences between groups (LQoLP scale: n=113, 2 RCTs, MD −0.23 CI −0.55 to 0.08; QOLI scale: n=132, 2 RCTs, MD 0.09 CI −0.24 to 0.42).

We found one further study providing data by long term for this comparison, but it was not entered in the analysis as data were skewed, measuring mean change from baseline on QOLI scale. These data tended to favour the ICM group.

1.8 Participant satisfaction/need

We found that participant satisfaction assessed with CSQ scale was significantly greater in the ICM group compared with standard care in all three time period assessments. Short-term data were available from only one small study and showed a significant difference between groups favouring the ICM intervention (n=61, 1 RCT, MD 6.20 CI 2.60 to 9.80); medium term data from two studies confirmed these results (n=500, 2 RCTs, MD 1.93 CI 0.86 to 3.01) and long term data was also favourable for the ICM group (n=423, 2 RCTs, MD 3.23 CI 2.31 to 4.14). One further small trial provided short term data but it was skewed, the standard care arm’s attrition was higher than 50%. Participant satisfaction was assessed with CSQ scale and it tended to favour the ICM group. We found more skewed data from three studies assessing participant need on two other scales (CAN, CANSAS). Medium-term data from one study failed to show any difference between groups. Long-term data were assessed with two different scales in two different studies: one study, using the CANSAS, did not show any trend in the direction of effect, data from the second study, using the CAN, tended to favour the ICM group.

1.9 Costs

Studies reported cost data assessing ’direct costs of psychiatric hospital care, ’health care costs’ and ’costs of all care’ at different periods of time.

Direct medium term costs of psychiatric hospital care were available from two studies (Chandler-California1 (A), Chandler-California1 (B),) reporting skewed data but with a sample size greater than 200. Data favoured ICM (n=426, 2 RCTs, MD −US$143.74 CI −272.40 to −15.08). Four more studies did describe ’direct costs of psychiatric hospital care’, but data were markedly skewed. Some of these data showed a trend favouring ICM, some standard care. Therefore we could not highlight any trend confirming the findings from meta-analysis.

Long-term data for direct health care cost were also available from two studies: again studies reporting skewed data but with a sample size greater than 200 so could be entered into a meta analysis. These data were inconclusive (n=873, 2 RCTs, MD US$-529.24 CI −2143.59 to 1085.10), as they were highly heterogeneous (I²= 94%, p<0.0001) with inconsistency in direction of effect.

Other skewed, medium term data from a study with a sample size of less than 200 could not be entered into the meta analysis, this data did not show any significant difference between interventions.

2. COMPARISON 2: INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT (Summary of findings 2)

Primary outcomes

2.1 Service use

2.1.1 Service use: Average number of days in hospital per month - at about 24 months

We found three studies reporting skewed data but with a sample size greater than or equal to 200. There was no significant difference between groups for reducing length of hospitalisation (n=694, 3 RCTs, MD −0.58 CI −1.93 to 0.76). These findings were in accord with the second sub-group analysis - skewed data from 18 studies with a sample size less than 200 participants. These data again did not show a significant difference between groups (n=1526, 18 RCTs, MD −0.03 CI −0.33 to 0.28). Overall, combining the two pools of studies, found no clear difference between ICM and non-ICM (n=2220, RCTs 21, MD −0.08 CI −0.37 to 0.21, Figure 9).

Figure 9.

Forest plot of comparison: 2 INTENSIVE CASE MANAGEMENT vs NON-INTENSIVE CASE MANAGEMENT, outcome: 2.1 Service use: 1. Average number of days in hospital per month - at about 24 months.

A funnel plot did not show any significant reporting bias (Figure 10).

Figure 10.

Funnel plot of comparison: 2 INTENSIVE CASE MANAGEMENT vs NON-INTENSIVE CASE MANAGEMENT, outcome: 2.1 Service use: 1. Average number of days in hospital per month - by about 24 months.

2.1.2 Service use: Not remaining in contact with psychiatric service

Short-term data were not available.

Medium-term data were available from one small study showing a significant difference favouring ICM group (n=73, 1 RCT, RR 0.27 CI 0.08 to 0.87).

Long-term data were available from three studies. Pooled data were not statistically significant (n=1182, 3 RCTs, RR 0.82 CI 0.34 to 1.98) but data were heterogeneous (I²=82%, p=0.004) with inconsistency in direction of effect.

When we pooled studies from medium and long term, data did not show significant differences between interventions but data heterogeneity was substantial (n=1255, 4 RCTs, RR 0.63 CI 0.27 to 1.49, I²=81%, p=0.001). We addressed this finding checking again for correctness of data and explored heterogeneity by dropping each study out of the analysis. Only by removing both Drake-NHamp and UK700-UK is homogeneity restored. We could not ascertain any clear reason for the heterogeneity. We have therefore chosen not to pool these data as it could be misleading to quote an average value for the intervention effect - particularly as this is the case, when there is inconsistency in direction of effect (Figure 11). We did not use a funnel plot for this outcome as there were less than ten studies (see Assessment of reporting biases).

Figure 11.

Forest plot of comparison: 2 INTENSIVE CASE MANAGEMENT vs NON-INTENSIVE CASE MANAGEMENT, outcome: 2.2 Service use: 2. Not remaining in contact with psychiatric services; data not pooled.

Secondary outcomes

2.1.3 Service use: Use of hospital

Binary data describing this outcome were available from three long term studies, reporting ’number of admitted to hospital’. These data showed a not significant difference in number of admitted to hospital between groups (n=1132, 3 RCTs, RR 0.91 CI 0.75 to 1.12).

Only one long term study provided continuous data on this outcome, reporting ’average number of admission’. Data were skewed, but as the trial sample size was greater than or equal to 200, we entered these data in the analysis. They failed to show any significant differences between groups (n=678, 1 RCT, MD −0.18 CI −0.41 to 0.05).

A trial that included less than 200 participants presented skewed data that we could not meta analyse. Data was reported for the outcome ’average number of admissions’ by medium term. As with previous findings, they failed to show any trend in effect between groups.

2.2 Adverse events: Death

2.2.1 Adverse events: Death all causes

Short-term data on mortality were available from one study (n=193), not reporting any deaths and therefore a measure of effect was not estimable.

Medium-term data were available from three studies, where one death occurred in 148 people treated with ICM, compared with no deaths in 146 people treated with non-ICM. There were no significant differences in mortality between groups (n=294, 3 RCTs, RR 2.92 CI 0.12, 69.43).

We found long term data in five studies, reporting 16 deaths occurring in 816 people treated with ICM, compared with 18 deaths in 821 people treated with non-ICM. These data confirmed medium term findings not showing differences in mortality between groups (n=1634, 5 RCTs, RR 0.90 CI 0.46, 1.75).

2.2.2 Adverse events: Suicide

Short-term data on suicide mortality were available from one study (n=193), again not reporting any deaths and therefore a measure of effect was not estimable.

Medium-term data were available from six studies, where five suicides occurred in 464 people treated with ICM, compared with three suicide in 465 people treated with non-ICM. There were no significant differences in the suicide rate between groups (n=929, 6 RCTs, RR 1.61 CI 0.26 to 9.85).

Long-term data were available from three studies, reporting six suicides occurring in 577 people treated with ICM, compared with seven suicides in 575 people treated with non-ICM. These data confirmed medium term data overall mortality and on suicide, with no significant differences in the suicide rate between groups (n=1152, 3 RCTs, RR 0.88 CI 0.27 to 2.84).

2.3 Global state

2.3.1 Global state: Leaving the study early

We found studies providing medium and long term data, but not short term.

Medium-term data showed no treatment effect in reducing number of lost to follow up (n=225, 2 RCTs, RR 0.64 CI 0.13 to 3.07), but these data presented a substantial level of heterogeneity (I²= 84%, p=0.01). In addition, there was inconsistency in the direction of effect between the two studies.

Long-term data were available from seven studies and we found a significant advance for ICM in reducing the number of lost to follow up (n=1970, 7 RCTs, RR 0.70 CI 0.52 to 0.95). Different from the medium term data, long term sub-group analyses did not show substantial heterogeneity. However, pooling data from two time sub-groups, we found data still significant, showing an advantage for ICM in reducing the number of lost to follow up (n= 2195, 9 RCTs, RR 0.72 CI 0.52 to 0.99). Although heterogeneity was reduced in comparison to the medium term sub-group data, these pooled data still showed a substantial level of heterogeneity (I²=59%, p=0.01).

2.3.2 Global state: HoNOS scale

One study reported long term data on global state assessed with HoNOS scale. These data were skewed, but entered the analysis as the study sample size was greater than or equal to 200 participants. Data failed to show a significant difference between interventions (n=239, 1 RCT, MD −0.40 CI −1.77 to 0.97).

We found skewed data describing global state with HoNOS scaleby medium and long term. These data trended to favour the standard care group.

2.3.3 Global state: Compliance with medication

One study reported medium term data for the binary outcome ’number of participants not compliant with medication’. There was no significant difference between groups (n=73, 1 RCT, RR 1.14 CI 0.42 to 3.05).

Long-term compliance scores assessed with the ROMI compliance and non-compliance sub-scale were not significantly different (compliance sub-scale: n=239, 1 RCT, MD 0.60 CI −0.05 to 1.25), (non-compliance sub-scale, n=239, 1 RCT, MD −0.60 CI −1.63 to 0.43), although both sub-scale scores tended to favour ICM.

Medium and long term skewed data were available from one study assessing compliance scores with the ROMI compliance and non-compliance sub-scale. These data tended to favour standard care, where participants had a higher level of compliance.

2.4 Social functioning

2.4.1 Social functioning: Contact with legal system

Medium-term data were available from one study reporting ’contact with the police’. We found no significant difference betwee groups (n=73, 1 RCT, RR 0.32 CI 0.04 to 2.97).

Long-term data were available both on binary outcome ’imprisoned’ and ’arrested’. We found two studies reporting the outcome ’imprisoned’ and data failed to show any difference in the number of people imprisoned (n=959, 2 RCTs, RR 1.15 CI 0.64 to 2.08). Only one study provided data on the outcome ’arrested’. Again there was no clear difference between groups (n=251, 1 RCT, RR 0.87 CI 0.53 to 1.42).

2.4.2 Social functioning: Employment status

Data were available from one medium term study, reporting ’participant who spent >1 day employed’ and ’participants on paid employment’. Both these outcomes did not show significant difference between groups (’spent >1 day employed’: n=73, 1 RCT, RR 1.46 CI 0.45 to 4.74; ’on paid employment’: n=73, 1 RCT, RR 0.97 CI 0.14 to 6.54).

2.4.3 Social functioning: Accommodation status

The outcome ’living in supported accommodation’ was only available for the medium term, as reported from one study. Data failed to show a significant difference between groups (n=73, 1 RCT, RR 2.59 CI 0.75 to 9.01).

The outcome ’homelessness’ was only available for the long term (one study). There were no significant differences between treatment groups in the number of people being homeless (n=251, 1 RCT, RR 0.69 CI 0.34 to 1.38).

The continuous outcome ’average days per month in stable accommodation’ was available for short, medium and long term (2 RCTs). Data did not find significant differences between groups at any time period (by short term: n=203, 1 RCT, MD −0.20 CI −2.48 to 2.08; by medium term: n=203, 1 RCT, MD 0.10 CI −2.15 to 2.35; by long term: n=901, 2 RCTs, MD −0.19 CI −1.37 to 1.00).

2.4.4 Social functioning: Substance abuse

Data describing this outcome were available both as a binary and continuous measure. Long-term binary data on ’substance abuse’ differentiated between ’alcohol abuse’ and ’illicit drug abuse’. There was no difference in the number of people abusing alcohol (n=251, 1 RCT, RR 1.10 CI 0.67 to 1.83) nor in those abusing illicit drug (n=251, 1 RCT, RR 1.08 CI 0.69 to 1.71). Binary data on ’remission from alcohol use disorder’ (defined AUS score < 3) also did not show any significant difference between groups by long term (n=223, 1 RCT, RR 0.86 CI 0.65 to 1.14).

Short, medium and long term continuous data were available, assessing the substance abuse with SATS scale. These data failed to show significant difference between groups at any time period assessment (short term: n=203, 1 RCT, MD 0.07 CI −0.28 to 0.42; medium term: n=203, 1 RCT, MD −0.11 CI −0.55, 0.33), (long term: n=203, 1 RCT, MD 0.11 CI-0.41 to 0.63).

Skewed data were available from one study assessing ’days using alcohol’ and ’AUS scale score’ by short, medium and long term. Data on ’days using alcohol’ showed a trend towards a higher alcohol consumption in the ICM group in each assessed time period. This trend was confirmed by the short and medium term findings on the AUS scale, where the ICM group rating had a worse outcome than the non-ICM group, however, long term data showed a worse outcome in the non-ICM group.

2.4.5 Social functioning: Scale data

We found LSP social functioning score did not favour any group over the other by long term (n=239, 1 RCT, MD 4.00 CI −0.61 to 8.61).

Skewed data on SFQ social functioning scores showed equivocal results by medium term, and a worse outcome in the ICM group by long term.

2.5 Mental state

2.5.1 Mental state: General symptoms

We found BPRS mental state scores did not favour any group during the short, medium and long term analysis (by short term, n=203, 1 RCT, MD −0.65 CI −3.99 to 2.69; by medium term, n= 203, 1 RCT, MD −1.62 CI −4.76 to 1.52; by long term, n=203, 1 RCT, MD −0.22 CI −3.32 to 2.88). These data were confirmed by long term findings on the CPRS, where there was no significant difference between groups (n=595, 1 RCT, MD 0.40 CI −1.83 to 2.63).

Medium and long term skewed data from the Krawiecka Scale could not be entered into meta analysis, along with long term skewed data from the BPRS scale. These data consistently suggested a better mental state outcome in the non-ICM group.

2.5.2 Mental state: Specific symptoms

Long-term continuous data on negative symptoms from one long term trial did not favour either group (n=593, 1 RCT, MD 0.20 CI −2.32 to 2.72).

Skewed data were available on anxiety and depression symptoms assessed with HADS anxiety and depression sub-scale by medium and long term. These data, provided by the same study, are notconsistent between medium and long term, as they showed a better outcome in the ICM group by medium term and a worse outcome in the ICM by long term, compared with the non-ICM group.

2.6 Behaviour (various measures)

Medium-term data reporting ’harm to self or others’ did not find significant difference between groups (n=73, 1 RCT, RR 0.88 CI 0.40 to 1.90).

Long-term data were available both on ’patient committing self-harm’ and ’patient committing injury to others’. There were no significant differences in these two data sets (’patient committing self-harm’: 708, 1 RCT, RR 1.06 CI 0.70, 1.61; ’patient committing injury to others’: n=959, 2 RCTs, RR 1.09 CI 0.85 to 1.40).

2.7 Quality of life

We found three studies assessing quality of life with three different scales (LQoL, MANSA, QOLI) at different time period. There were no significant differences in any of these measures - at any time period. Data on QOLI scores were available by short and medium term from one study (short term QOLI: n=203, 1 RCT, MD −0.02 CI −0.43 to 0.39; medium term QOLI: n=203, 1 RCT, MD −0.04 CI −0.43 to 0.35). Long-term data were available from three different studies measuring quality of life with three difference scales, therefore not entering the analysis together. None of them showed any significant difference between groups (LQoL: n=526, 1 RCT, MD 0.03 CI −0.10 to 0.16; MANSA: n=166, 1 RCT, MD 0.10 CI −0.19 to 0.39; QOLI: n=203, 1 RCT, MD 0.10 CI −0.25 to 0.45).

2.8 Participant satisfaction/need

Long-term data were available from one study, assessing participant need with CAN, and participant satisfaction with health service with the Homonym scale. There were no significant differences between groups as assessed with the two scales (CAN: n= 585, 1 RCT, MD −0.29 CI −0.69 to 0.11; Satisfaction with Health Service: n=490, 1 RCT, MD −0.40 CI −1.25 to 0.45).

Skewed data from a different study suggested a long term better outcome on satisfaction with treatment for ICM group (data assessed with CSQ scale).

One study provided data suggesting a difference between intervention in participant need as assessed with CAN scores, but these data were skewed. They showed a worse outcome in the ICM group in the medium and long term.

2.9 Costs

Costs were assessed measuring ’direct costs of psychiatric hospital care’ and ’direct cost of all care’. No data were available on ’direct health care costs’.

Skewed data on direct costs of psychiatric hospital care were available from two studies - by medium and long term. Medium and long term data from one study did not find a significant difference between groups, whilst the second long term study data found costs significantly lower for the ICM group. The latter study was a small one (Quinlivan-California, n=60), where the first one was larger (Harrison-Read-UK, n=193).

Direct cost of all care were available from one study (UK700-UK) reporting skewed data but with a sample size greater than 200. There was no significant difference between groups for reducing direct cost of all care (n=667, 1 RCT, MD 77.00 CI −66.63 to 220.63), but findings showed a trend suggesting greater cost in the ICM group. This trend suggested a cost increasing of £77 per person per month by long term, referred to fiscal year 1997/98. One study (Johnston-Australia, n=58) reported skewed data on the same outcome (direct cost of all care). This set of data did not show any significant differences between groups by the medium term, substantially confirming data by long term.

3. SENSITIVITY ANALYSES

As anticipated in the Methods section (see Methods, Sensitivity analysis), we performed the following sensitivity analysis.

3.1 Implication of randomisation

3.1.1 COMPARISON 1: INTENSIVE CASE MANAGEMENT versus STANDARD CARE

All the included studies were described as randomised and none were described in a way as to imply randomisation, therefore we did not include any trials in the anticipated sensitivity analysis.

3.1.2 COMPARISON 2: INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

All the included studies were described as randomised and none were described in a way as to imply randomisation, therefore we did not include any trials in a sensitivity analysis

3.2 Standard care caseload is over 20

3.2.1 COMPARISON 1: INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Between studies providing primary outcomes, only six studies reported the ratio staff to patients, and for each study it was >1: 20 (Bond-Chicago1, Bond-Indiana1, Herinckx-Oregon, Jerrell-SCarolina1, OPUS-Denmark, Sytema-Netherlands).

When we entered these studies (where standard care group caseload was greater than 20) in the analysis, the first primary outcome ’average days per month in hospital’ showed a significant favourable effect in ICM group in reducing the length of hospitalisation (n=951, 7 RCTs/study centres, MD −2.01 CI −3.36, −0.67), but heterogeneity was substantial (I²=70%, p=0.003). These findings confirm those obtained when all studies were entered in the analysis, regardless of caseload in the standard care comparison group. The second primary outcome ’not remaining in contact with psychiatric services’ showed that participants were significantly less likely to lose contact with psychiatric services than participants in the ICM group compared to the standard care group by medium, long term and overall (n=931, 4 RCTs, MD 0.38 CI 0.23, 0.63), but heterogeneity was substantial (I²=67%, p=0.03). These findings confirm those obtained when all studies were entered in the analysis, regardless of caseload in the standard care comparison group. But data in the sensitivity analysis showed a substantial level of heterogeneity not present when all studies were entered in the analysis.

3.3 Assumptions for lost binary data

3.3.1 COMPARISON 1: INTENSIVE CASE MANAGEMENT versus STANDARD CARE

In this comparison, no assumptions had to be made regarding people lost to follow-up at the primary binary outcome ’not remaining in contact with psychiatric services’, therefore we did not include any trials in the anticipated sensitivity analysis.

3.3.2 COMPARISON 2: INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

In the second comparison, assumptions had to be made regarding people lost to follow-up for the primary binary outcome ’not remaining in contact with psychiatric services’ for one trial (REACT-UK). When we compared the findings of the primary outcomes when we used our assumption compared when we used completer data only, results did not show any substantial difference.

3.4 Assumptions for lost continuous data (for meta-regression)

Sensitivity analysis was undertaken, removing 13 out of 52 trials originally entering meta-regression. These were trials where primary outcome SD was imputed (see Table 2). Therefore meta-regression was run on the remaining pool of 39 non-imputed studies. Results did no longer reach significance for both variables ’organisation fidelity sub-score’ and ’baseline hospital use’ as p is at trend ~ 0.07 (’organisation fidelity sub-score’: regression coefficient −0.2 CI −0.48 to 0.02, p=0.067; ’baseline use of hospital’: regression coefficient −0.18 CI −0.39 to 0.02, p=0.077). Removing a high influence outlier (Rosenheck-USA-NP (C)), significance returns for the effect of ’baseline hospital use’ (−0.26 CI −0.51 to −0.01, p=0.046), but not for ’organisation fidelity’. When combining both two variables within the model (’organisation fidelity sub-score’ and ’baseline hospital use’), the pattern of results did not change for the non-imputed studies.

DISCUSSION

Summary of main results

1. COMPARISON 1: INTENSIVE CASE MANAGEMENT versus STANDARD CARE (Summary of findings for the main comparison)

Primary outcomes

1.1 Service use

1.1.1 Service use: Average number of days in hospital per month - at about 24 months

ICM does seem to reduce length of hospitalisation when compared with standard care (by 0.86 day/month over 24 months). These data are from a heterogeneous analysis - when certain studies are removed from the analysis, the now homogeneous result suggests that the saving is in the region of 0.6 days/month. We are not quite sure which of the two figures is really the most reliable, but in any event trial findings suggest a saving of time in hospital per person of between about 7 and 10 days in hospital per year.

What is important, however, is that there was a high duration of hospital stay for these people in the two previous years (averaging 6 days/month). This is higher than for those in the ICM vs non-ICM comparison (averaging 3.4 days/month) - which found no difference between groups (see Figure 9). The impressive saving of time in hospital for ICM was greater than that of standard care, but the saving in standard care was also considerable (absolute decline for ICM is about 3 days, standard care is 1.8 days). Well organised standard care, for people with what would seem to be a history of considerable periods spent in hospital over the last 24 months, does seem to reduce the time in hospital for the ensuing two years. ICM, however, adds more than an extra day to that gain.

The meta-regression is a tool of limited power, employed on weak data. The results are, however, in keeping with the results of one other review (Burns 2007) which suggests that the gain of ICM was not so much linked with the total fidelity score, the staff sub-scale, or the comparison group. It did suggest a link with the organisational sub-score of the IFACT scale (0.36 day/month out of hospital gained by each point increase of this sub-score). It also suggested that ICM effect is linked to length of hospitalisation during the previous two years (0.2 day/month out of hospital gained by each day increase in the day in hospital/per month during previous two years).

In other words this means that:

• if the ICM team ratio staff:client is less than 1:20 (calculated dividing the number of active clients on the caseload by the number of full time equivalents of direct service staff on the team), then it makes no difference how much the caseload is lower than 1:20;

• the ICM team size does not matter (calculated as the number of full-time clinical staff equivalents, as defined earlier);

• the availability of a psychiatrist on the team is not pivotal; and/or

• the availability of a nurse on the team is not pivotal

It also suggested that gains in IFACT scores can reduce average hospital stay. McGrew 1994 and McGrew 1995 suggest that gain can be mediated by:

• ensuring that the ICM team performed the role of primary therapist for the client (the primary therapist role designates the person, within the local mental health system, with primary clinical and record keeping (e.g., treatment plans) responsibility for the client);

• the ICM team’s offices being located in a separate building from the parent agency’s main offices (i.e. usually away from the hospital site);

• the ICM team sharing caseloads (rated as the degree to which all staff members on the team had contact with all clients on a regular basis, e.g., through rotation), in contrast to individual caseloads in which specific staff workers are responsible for specific clients;

• the ICM team meeting as a group each weekday to discuss their entire caseload;

• the ICM team’s supervisor devoting at least halftime to client contacts, either co-jointly during supervision of team members, or individually as part of his or her duties as a member of the team;

• the ICM team providing 24-hr direct access to the team (If access to the ICM team was triaged through the community mental health centre (CMHC) emergency 24-hr on-call service, intermediate score of 0.5 is obtained, therefore the advantage on decreasing days in hospital per month would be halved (−0,2 days in hospital per month)); and/or

• the ICM team serving clients without any expectation of transferring them to another programme

1.1.2 Service use: Not remaining in contact with psychiatric services

Overall, we found ICM to be better than standard care for retaining people in psychiatric services (n=1811, 10 RCTs, RR 0.45 CI 0.34 to 0.61). This effect, however, was not seen for the short term analysis (n=95, 1 RCT, RR 0.54 CI 0.28 to 1.05), although confidence intervals are compatible with an effect favouring ICM (p=0.07). Medium-term findings did suggest an overall effect of ICM being more effective in reducing loss to follow up to psychiatric service contact (n=1063, 3 RCTs, RR 0.51 CI 0.36 to 0.71). In the longer term this same effect was heterogeneous, due to an outlying finding of a single study (Herinckx-Oregon). This trial had a peculiar definition for this outcome which was different to the other studies. For Herinckx-Oregon ’not remaining in contact with psychiatric services’ did not include refusing re-interview, moving out and death, whereas for the other trials it did. If Herinckx-Oregon is excluded and only five long term trials are retained in the analysis, then ICM appears to be effective in preventing loss to follow (n=475, 5 RCTs, RR 0.27 CI 0.11 to 0.66). The result of a better retention in care for ICM group strengthensthe relevance of the first primary outcome (of ICM decreasing time in hospital). ICM decreases days of hospitalisation in a severe mentally ill population where patients are kept in close contact with services. Therefore, the ICM effect on reducing days in hospital might not be dissipated by a higher rate of loss to follow up.

Secondary outcomes

1.1.3 Service use: Use of hospital

We found ICM to reduce the number of people admitted to hospital more than standard care - at least in the medium term (n=1303, 5 RCTs, RR 0.85 CI 0.77 to 0.93). Skewed data for ’average number of admissions in the medium term’ are in accord with these first findings. Synthesis of short and-long term studies in this outcome produce heterogeneous findings. The short term data are from only two studies and when the most positive one (Bond-Indiana1) is removed the finding becomes clearly null. The long term data is from eleven trials, removing the three clearly outlying studies (Curtis-New York, Macias-Utah, Test-Wisconsin) also restores homogeneity and moves the finding squarely towards the null.

Overall the effect of ICM on admission to hospital is not strong. Whereas the time spent in hospital does seem to be less if allocated ICM at least for those whose baseline use of hospital was high, the number of admissions is not greatly changed. Admissions are shorter but not by much, if any less frequent.

1.1.4 Service use: Use of services outside of mental health provision

The only outcome available was ’use of Emergency Room’ as well as ’rate of use of Emergency Room’. Studies did not report a convincing difference in use. We do not, however, consider these data very robust as they were based on data from one study only or on skewed data that are very difficult to interpret.

1.2 Adverse event: Death

This review did not find any difference in mortality either due to ’all causes’ or to ’suicide’ in the short, medium and long term. Although death is a rare adverse event, the duration of the longer studies (~24 months) means that some deaths would have been expected to occur and did (3.2% ICM vs 3.8% standard care). This difference is not statistically significant but is homogeneous and it suggests a trend in term of risk reduction (RR 0.84 CI 0.48 to 1.47). The same trend is confirmed in the ICM effect on mortality due to ’suicide’ by long term (suicide rate1.3% ICM vs 2% standard care). Again this difference is not statistically significant but is homogeneous and stronger compared to ’death all causes’. If the suicide risk is really so much reduced (RR 0.68 CI 0.31 to 1.51) this would be a really important finding, also considering that the quality of evidence is moderate (Summary of findings for the main comparison).

Few things have changed the outcome of death for people with schizophrenia. If, for this group of people, set in the context of a standard care with a high baseline risk of admission, ICM does really decrease death, and by so much, this really is a strong argument in favour of the ICM. These data are the findings of eight studies with only a total number of participants of 1164. One larger trial might be able to confirm this important suggestion.

1.3 Global state

1.3.1 Global state: Leaving the study early

For the short term ICM data regarding preventing number of lost to follow up were heterogeneous. However we found ICM to be more advantageous than standard care in reducing rate of lost to follow up both in the medium term (n=1699, 8 RCTs, RR 0.60 CI 0.51 to 0.70) and long term period (n=1798, 13 RCTs, RR 0.68 CI 0.58 to 0.79). There remains the impression that ICM, and again with the proviso that these findings may apply most specifically to a group with high baseline admission, holds on to people more tightly across time. This may not lower admission rate much but loss to follow up and length of admission may decrease. Overall these data are not of high quality (Summary of findings for the main comparison) but there is an impression that ICM has more advantage over standard care for the higher risk groups. For example, in groups with only 10% loss to follow up across the long term only three more people are not lost for every 100 given ICM. However, with more realistic figures of 50% loss to follow up, this figure rises to 15 more people out of every 100 who are kept in care compared with those allocated standard care.

1.3.2 Global state: GAF

Short-term studies indicated a better improvement in GAF end-point score for participants on ICM compared with those allocated standard care (short term: n=797, 4 RCTs, MD 2.07 CI 0.28 to 3.86). This effect was confirmed by the long term data (n=818, 5 RCTs, MD 3.41 CI 1.66 to 5.16). Whilst this is favourable for those allocated ICM we are unsure of the clinical meaning of these data as changes of two or three points on a scale that runs to 100 does not seem much. We have found no reference to the clinical meaning of such small changes.

1.3.3 Global state: Not compliant with medication

Only one long term study provided data on compliance with medication and these indicated a higher compliance level in those allocated to ICM intervention compared with people in the standard care group (n=71, 1 RCT, RR 0.35 CI 0.15 to 0.81). Again this is an important finding and should be replicated. We are surprised that such easily recorded data are not reported in more studies.

1.4 Social functioning

1.4.1 Social functioning: Contact with legal system

Studies measuring contact with legal system used different definitions over a variety of periods of time. This makes interpretation difficult. There is no real suggestion that ICM increases or decreases the measures used to record these issues. The ’arrested’ and ’imprisoned’ outcomes were the only ones with some consistency. They did not show any significant difference in the intervention effect between groups. The ’contact with the police’ outcome findings were from only one study and were not significant in the short term, but became significantly different in the medium term (favouring the ICM group). Overall the legal outcomes are not convincing but there is more need for consistency in approach tothis area of research.

1.4.2 Social functioning: Employment status

This review did not reveal any significant difference between ICM and standard care in employment status, as measured by different outcomes. Medium term findings on ’number of people not competitively employed’ were based on only one trial. Both medium and long term data reporting ’not employed’ tended to favour ICM, although data were heterogeneous and overall quality of evidence was very low (Summary of findings for the main comparison). Again, this is an important outcome for which more data are needed before anyone can draw firm conclusions.

1.4.3 Social functioning: Accommodation status

Data on outcome ’homelessness’ were not convincing either by short, medium and long term, but had usually been derived from few trials.

The outcome ’not living in stable accommodation’ was scarcely reported, as it was available just from one long term study.

Regarding the ’not living independently’ outcome, we found people allocated to ICM were more likely to live independently compared with those in the standard care group - in the medium term and, even more so in the long term. This has to be another important finding of this review. If the risk of not living independently really is substantially reduced by this ICM package (18% ICM vs 26% standard care, long term) at least for people with high baseline risk of admission and, if this is a desired outcome for this particular client group then, combined with the other moderate but cumulative advantages, this further highlights ICM’s advantage over standard care.

1.4.4 Social functioning: Substance abuse

Only one study (n=547) reported usable binary data for alcohol and drug abuse. We found a long term single study failed to show any advantage for participants treated with ICM compared to standard care. Skewed data was supplied by Sytema-Netherlands (n= 81) with DALI scores tending to favour the ICM group for both drug use and alcohol consumption. We, however, are not clear of the clinical meaning of these scores. Morse-Missouri3 (n=103) reported skewed data on days substances were used per month. There is no indication of any difference between groups. Currently, there is not compelling evidence that ICM effects abuse of substances or alcohol.

1.4.5 Social functioning: Various scales

Few studies reported usable data for social functioning scale and outcome data were complicated by the use of different scales by single studies making meta-analysis impossible. In this confusion of evidence, we see no advantage for participants treated with ICM compared to those in standard care in terms of social functioning measures. This does not seem to concur with other findings re independent living. It could be that the fine grain measures of functioning are picking up subtle parameters of social function not effected by the package. It could also be that they are not sensitive enough measures to broad and important issues of social function.

1.5 Mental state

Mental state rating in these trials illustrate the confusion of how such symptoms are recorded in randomised studies. Timings of use of the scales differ and the findings are so problematic to interpret from the clinical perspective that were are left making safe but bland conclusions.

There does not seem any compelling evidence that, in this group of people, set where the baseline risk of admission is high, ICM really substantially effects a person’s mental state.

1.6 Behaviour: self-harm

Self harm was not convincingly reduced by use of the ICM model and these are based on findings from the larger of the studies. The mortality finding discussed above, however, does seem to suggest that ICM reduced the risk of death. These findings are a little at odds with each other, although not entirely, but give all the more reason to continue to research into this area for these, the simplest of outcomes.

1.7 Quality of life

Few studies report relevant outcomes. Short and medium term outcome data were complicated by the use of different scales by single studies making meta-analysis impossible. We found one short term study (n=125) showed a better quality of life in the ICM group on the QOLI scale but more medium term and long term data failed to show any advantage for participants treated in ICM group compared with standard care. The few skewed data seems to concur with the impression that for the quality of life measures used in trials, ICM confers no advantage over standard care.

1.8 Participant satisfaction/need

ICM participants were more satisfied with their treatment compared with the standard care in these trials. These findings are based on quite strong data (short term n=61, medium term n=500, long term n=423). More satisfaction with care could enhance medication compliance, the will to keep in services, housing status and a host of other variables. We are left with the doubt about the size and meaning of the overall finding. We are not sure of how encouraged we should be that these packages of care deliver an average of a two to three point improvement in the Client Satisfaction Questionnaire.

Several of the smaller trials did measure need and unmet need. These skewed data are difficult to interpret but do not convincingly seem to favour either of the groups.

1.9 Costs

With respect to cost of in-patient psychiatric care ICM was consistently superior to standard care ’direct costs of psychiatric hospital care’, findings suggesting a saving of money per person of about US$144 per month (fiscal year 1990). Those data were provided from two studies, taking part in the same multi-centre trial. Skewed data did not show any trend confirming or disputing this data, as skewed data were contradictory.

No difference was found respect to direct health care costs or direct costs of all care.

2. COMPARISON 2: INTENSIVE CASE MANAGMENT versus NON-INTENSIVE CASE MANAGMENT (Summary of findings 2)

Primary outcome

2.1 Service use

2.1.1 Service use: Average number of days in hospital per month - at about 24 months

This review failed to show a significant advantage in reducing the average length of hospitalisation when ICM is compared with non-ICM. This could be an important finding and we see no clear reason not to trust this result. The implications from this could be that if services are already providing non-ICM there is no point in investing in further intensiveness. Currently we know of no review comparing non-ICM with standard care and reporting relevant outcomes. This should be undertaken. It is possible that there are other features of ICM that may improve outcome but we have not stipulated that we should specifically investigate for these. What was different between the two sets of trials is the baseline risk of admission in the previous two years (about 6 days/month Comparison 1 vs about 3.4 days/month Comparison 2). This was highlighted to us by the meta-regression process. This generates further hypotheses. Baseline hospital risk is linked to service provision, service culture, severity of illness, and other issues. We do not have the sophistication of data to investigate these. What we are left with is the possibility that in a situation where people with severe mental illness have a duration in hospital of <4 days/month in the two years preceding the ICM package of care, the increased intensity of approach may not be justified.

2.1.2 Service use: Not remaining in contact with psychiatric service

We found ICM to be more effective in increasing the number of people retained in contact with psychiatric service in the medium term, but these findings were not considered robust as they were based only on one small trial (n=73). No difference was found between interventions in the long term data, but data were heterogeneous. Overall, when pooling medium and long term data, we did not find any advantage for participants treated in the ICM compared to non-ICM regarding better retention in psychiatric service but, again, these data were heterogeneous and we found no obvious explanation for the heterogeneity (n=1255, 4 RCTs).

Secondary outcomes

2.1.3 Service use: Use of hospital

This review did not reveal any difference in the risk of being admitted to hospital in the long term (n=1132, 3 RCTs, RR 0.91 CI 0.75 to 1.12). These findings were confirmed by data from one long term study (n=678) on the average number of admission, where no advantage was shown between treatments in reducing number of admissions in the long term.

Therefore data on frequency of admission and those data on length of hospitalisation are consistently showing no effect of ICM on both outcomes.

2.2 Adverse events: Death all causes and suicide

This review did not find provide strong evidence on mortality rate either due to all causes or to suicide in the short, medium and long-term. These data are quite informative, especially those from long term studies where the study length might balance the rarity of the event in detecting any difference between intervention effects. In the long term studies some deaths occurred (2.0% ICM vs 2.2% non-ICM), (n=1634, 5 RCTs, RR 0.90 CI 0.46, 1.75). This impression was confirmed for studies reporting suicide, as shown in Summary of findings 2, where low quality evidence showed no difference between groups.

2.3 Global state

2.3.1 Global state: leaving the study early

For the short term outcome, no studies were available.

By the medium term data showed no difference between interventions, but those data were not strong as they came from a small sample of two studies (n=225) and were heterogeneous with inconsistency of effect.

By the long term we found ICM to be more advantageous than non-ICM in reducing rate of lost to follow-up.

Overall pooling studies from two time sub-groups, we found heterogeneous data, but substantially confirming those homogeneous ones obtained by long term. They confirmed ICM to be more advantageous than non-ICM in reducing rate of lost to follow-up (n=2195, 9 RCTs, RR 0.72 CI 0.52 to 0.99). If we consider this outcome as proxy of a better retention in care, it might overcome the inconsistency of data on “number of people remaining in contact with psychiatric service” (see Discussion - Primary outcomes - 2.1.2 Service use: Not remaining in contact with psychiatric service). Therefore ICM seems to positively reduce number of lost to follow-up, but not affecting length and frequency of admission. These data are not of high quality (Summary of findings 2) but there is an impression that ICM has more advantage over non-ICM for the higher risk groups. For example, in groups with only 10% loss to follow up across the long term only three more people are not lost for every 100 given ICM. However, with more realistic figures of 50% loss to follow up this figure rises to 14 more people out of every 100 are kept in care compared with those allocated standard care.

2.3.2 Global state: HoNOS scale

Not enough studies were available to run a meta-analysis on data derived from the HoNOS scale or from any other scales assessing global state. Considering how comprehensive the scale assessment of other outcomes is, it appears that global state as an outcome is under-recorded by trialists despite being informative and relevant from a clinical point of view.

2.3.3 Global state: Compliance with medication

As for the previous outcome, not enough studies were available to run a meta-analysis. Again, a very meaningful outcome from a clinical point of view is neglected.

2.4 Social functioning

2.4.1 Social functioning: Contact with legal system

As it was for studies included in the first comparison (see Results 1. COMPARISON 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE), studies measuring contact with legal system used different definitions over a variety of time periods. This makes interpretation difficult. Besides only few studies addressed this outcome (3 trials overall). There is no real suggestion that ICM increases or decreases the measures used to record these issues. The ’imprisoned’ outcome was the only one with some consistency. It did not show any significant difference in the intervention effect between groups. The ’contact with the police’ and ’arrested’ outcome findings were from only one study each and were not significant in the short and long term respectively. Overall the legal outcomes are not convincing but there is more need for consistency in approach to this area of research.

2.4.2 Social functioning: Employment status

Data available on this outcome are substantially inconclusive, being provided by only one small trail (n=73). It measured employment status according to two different definitions ’spent > 1 day employed’ and ’on paid employment’. Both findings were not significant in the medium term and overall quality of evidence was low (Summary of findings 2). Again, this is an important outcome underestimated in the current studies and at this stage more data are needed before anyone can draw firm conclusions.

2.4.3 Social functioning: Accommodation status

Surprisingly scarce data were available on this outcome, as it was available just from two studies measuring ’living in supported accommodation’ and ’homelessness’ respectively. Both medium term data on ’living in supported accommodation’ and long term data on ’homelessness’ were based on only one trial and they were not significant. Again, this is an inconclusive finding due to the scarceness of available data, despite being easily recorded and very relevant from a perspective of a community based service.

2.4.4 Social functioning: Substance abuse

Studies measuring substance abuse were only two and they described this outcome as a binary and continuous measures, and not consistently across studies. This makes interpretation difficult and findings not convincing, as a single study entered each measurement, therefore we could not carry out any meta-analysis. As far as we can assess there is no indication of any long term difference between groups in ’number of people abusing alcohol’, in ’number of people abusing illicit drug’ nor in ’remission from alcohol use disorder’ (defined as AUS score <3). These findings were confirmed by those continuous ones, assessing the substance abuse with SATS scale. These data failed to show a significant difference between groups at any time period assessment. As for the first comparison, currently there is no compelling evidence that ICM effects abuse of substances or alcohol.

2.4.5 Social functioning: Scale data

Findings were equivocal on scale data measuring social functioning as provided by only one study. We do not think that future studies should address this outcome by use of scale measurement. Scales are not sensitive measures for social functioning. More effort should be placed on consistent and wide measurements of the main issues of social functioning (such as accommodation status, employment status, contact with legal system, rate of permanent social benefits).

2.5 Mental state: General symptoms and Specific symptoms

Again, outcomes measured on scales are substantially inconclusive as the data are spread across single studies on different scales and at different time periods, making meta-analysis impossible. According to the low quality of data available, there does not seem any compelling evidence that ICM really substantially effects mental state.

2.6 Behaviour: self-harm and injury to others

This review did not reveal any long term significant difference between ICM and non-ICM in the risk of committing self-harm or injury to others. These data are based on findings from the larger of the studies (UK700-UK). They are consistent to the mortality findings discussed above, where no significant difference was shown in death rate between groups, either for suicide and for all causes. Although these data are suggestive of no difference of effects between interventions, they are still quite weak due to limited sample size. More trials should address this outcome, one of the simplest ones to collect.

2.7 Quality of life

Quality of life rating in these trials illustrate the confusion of how such symptoms are recorded in randomised studies. The scales differ and timings of use of the scales also differ. There was such an inconsistency in approach to this area of research to make meta-analysis impossible. None of the available findings showed any significant difference between interventions. There does not seem any compelling evidence that ICM really substantially effects a person’s quality of life.

2.8 Participant satisfaction/need

Findings tended to favour ICM participants in gaining a better satisfaction with health services and in reducing their need. But this difference was not significant and both findings were based on data from the same trial, the largest one (UK700-UK, n=585). Therefore we can not draw any conclusions, but to highlight a possible favourable effect in the ICM group, that needs to be confirmed.

2.9 Costs

Studies assessed ’direct costs of psychiatric hospital care’ and ’direct cost of all care’. Regarding first outcome, findings are based on skewed data, provided from one small trial (Quinlivan-California, n=60) and a larger one (Harrison-Read-UK, n=193). There does not seem any compelling evidence that ICM really substantially effects ’direct costs of psychiatric hospital care’ either by medium and long-term. Again, findings on long term ’direct cost of all care’ did not show any difference between interventions.

Overall completeness and applicability of evidence

1. Completeness

1.1 Duration of follow up

The majority of studies presented long term data, i.e. over one year of follow up. This is a reasonable time length to assess sensibly any difference in the intervention effects.

1.2 Coverage of outcomes

As the experimental intervention is a service organisation model, its realisation involves health policy and research should account for efficacy and cost evaluation. The outcomes reported were mainly service use and social functioning oriented. No studies reported data on relapse (see Summary of findings for the main comparison, Summary of findings 2), positive symptoms, carer satisfaction and family burden. Participant satisfaction was scarcely reported and in a fragmented way, therefore available data are only partially informative on the effects of these approaches. Few studies provided cost data.

2. Applicability

2.1 Origin

The origin of the data has changed in the last decade since the two original reviews (Marshall 2000a, Marshall 2000b). There are now more included trials from Europe whereas in the past the data source was largely North America with a few from Australia. Now 30% of the total sample included in the review is composed of randomised people from Europe. For the primary outcome these studies add power to the result, narrowing the confidence intervals, but not otherwise substantially changing the findings. Furthermore, as all the included studies were from Europe or North America and Australia, this review findings are still lacking applicability to low-income countries and, more generally, to countries where mental health systems are not community-based.

2.2 People

Studies included people presenting a variability that we feel is likely to reflect the heterogeneous population a clinician faces in daily practice when treating people affected by severe mental illness. This variability was in terms of diagnosis (where participants were affected by a wide diagnostic group including schizophrenic disorder, affective and personality disorder); comorbidity (where four studies (Drake-NHamp, Essock-Connecticut2, Morse-Missouri3, Muller-Clemm-Canada) included dually diagnosed participants); and social characteristics (where eight trials included homeless participants). On average, studies included people with a long history of illness, as just OPUS-Denmark included participants with a first episode of psychotic illness. This fits with the concept of severe mental illness, where this label includes certain criteria relating to length of illness.

2.3 Interventions

Some studies showed a greater applicability, because the experimental intervention was provided by pre-existing team, therefore closer to the real world and less contaminated by the experimental setting.

The majority of new included trials compare ICM with non-ICM (8 trials out of 14). This is confirming the trend of psychiatric services to increasingly include some elements of the original model, but also to dilute and contaminate them with the current organisation. Therefore what we call ’standard care’ is converging toward non-ICM. For those of us who practice in Europe it is the second of the two comparisons in this review that may well be more applicable to everyday care. This comparison, it is important to recall, substantially did not illustrate much difference between ICM and non-ICM.

Quality of the evidence

As illustrated in Figure 1 there is the impression of a moderate overall risk in these trials of bias. This would mean, therefore, a moderate risk of overestimate of positive effect. Also, making the difficult judgements about quality has been helped greatly by a discernable improvement in reporting of methodology.

Potential biases in the review process

There are several potential biases. We have worked mainly with published reports, and only in few cases with unpublished material. By doing this we may be perpetuating a reporting and publishing bias. It would be better to have much more original individual patient data. This review follows on from two past Cochrane reviews (Marshall 2000a, Marshall 2000b) as well as much work already published in paper format (Burns 2007). The conduct of these reports has influenced this document and it is possible that we have failed to identify systematic biases in the way we have done the reviews across time.

The authors of this review do include an active pioneer of developing and implementation the experimental intervention model across the scientific community and clinical world (MM) and one included study is his (Marshall-UK). As a team we have tried to ensure that decisions are made by rational consensus and not to have an expert in the team would have been an inadvisable omission.

In some cases protocol rules were not clear enough and need for subsequent clarification arose and post hoc decisions had to be taken (see Differences between protocol and review). In various cases this could affect the review process. We think, probably, this has lowered the quality of data included in the review but not to include so much, for example, skewed data would have omitted much information. Also by breaking down studies into their centres, many fall below the 200 participant cut off point. We have included these data in the <200 category whereas they would have been in the >200 in previous versions. The impression of the overall effect of the changes in protocol are that we have a more inclusive review, with data that are more heterogeneous and also more favourable for the experimental interventions than otherwise would have been the case should we have used are more limited data set. Nevertheless, we do feel it reasonable to present all these data for the reader to consider.

We pre-specified what characteristics of studies could be associated to heterogeneity and therefore we stated in the protocol what variables were to be explored in the meta-regression before inspecting studies’ results. Despite this pre-specification, we were not blind to what variables were more probably related to heterogeneity, as we were familiar to some study results previously published. Therefore the undertaken exploration of heterogeneity might at best lead to generation of hypothesis, but it can not provide reliable conclusions.

Agreements and disagreements with other studies or reviews

This review merges two older Cochrane reviews fully bringing up to date how these data should be considered. It is not that this version disagrees with the older reviews - it just replaces them with a more current viewpoint of the data. A major improvement of this version are data on duration of admission - previously lacking from past reviews. Paper versions of this area do not provide a full summary of available evidence on ICM effects across various outcomes (Burns 2007). This Cochrane review does not disagree with the paper version - it is simply much more comprehensive. Where it comes to the meta-regression, this reviews substantially confirms the hypothesis stated elsewhere (Burns 2007), that the baseline hospital use and fidelity to the model effects outcome.

AUTHORS’ CONCLUSIONS

Implications for practice

1. For people with severe mental illnesses

ICM was found effective in ameliorating many outcomes relevant to people with severe mental illnesses. In fact ICM was shown to reduce hospitalisation, in term of less frequent and shorter admission to hospital; increasing retention in care; probably reducing the risk of death and suicide; globally improving social functioning in terms of a better accomodation status, employment status, and showing a trend reducing contact with legal system. Although it remains not clear the effect on mental state and quality of life, ICM effects seem to significantly help global state compared with standard care. It is not clear, however, what gain ICM provides on top of a less formal non-ICM approach. The latter may suit some people with severe mental illnesses better than the more intensive full ICM model. Data on satisfaction with care for the ICM versus non-ICM models are very few and difficult to interpret.

2. For clinicians

ICM formalises a holistic approach to care of people with severe mental illnesses in the era of limiting hospital admissions and subsequent hospital closure. What this review suggests is that this formalising is helpful across several outcomes over and above a standard care - the latter largely based in outpatients departments - and that this seems acceptable to people with severe mental illnesses. However, when the fully formal holistic approach (ICM) is compared with the less formal but also holistic non-ICM the differences are not so clear. This could be seen as encouraging, as many clinicians are unlikely to rigidly apply full ICM for various reasons. This does not abrogate the need to know and apply key components of the model of care within ICM.

3. For policy makers

What we know at this juncture is that ICM is of value at least to people with severe mental illnesses who are in the sub-group of those with a high level of hospitalisation (about 4 days/month in past 2 years). The intervention should be performed close to the original model, therefore training should be planned for relevant mental health workers. Data on costs are still scarce and conclusions cannot be drawn on cost-effectiveness. Where ICM features are already available in the community psychiatric service (the non-ICM intervention), it is not clear if additional full development to the rigid model of ICM is of value. Results of this review could guide policies on introduction of such an ICM service in those countries where a community psychiatric service is already set up but ICM is not in use, and in countries where a shift from hospital-based care in favour of a more community-focused approach has still to be developed. Particular consideration should be placed on the setting where ICM is going to be set up, as its value was shown where the level of hospitalisation is high. It is not clear whether the introduction of just some of the ICM features (the non-ICM intervention) is of value, compared with the community-based standard care, as more research is needed to clarify the non-ICM effect compared with standard care.

Implications for research

1. General

First, as has already been said, reporting of research does seem to have improved across time and as a result it is much easier to understand details of the practice of modern studies.

However, this review illustrates how scale measurements are much more widespread than simple clinical questions for assessing clinical outcomes. Binary data, we suggest, are less ambiguous than continuous. Further complicating matters is that there are many scales for the same outcome: we often found many studies assessing the same outcome on different scales and did not feel justified to run a meta-analysis. For example, regarding social functioning outcomes, there was need for consistency in approach to this area of research. Heterogeneous measurements were used to describe the same outcome. This is not very informative - but this review illustrates how opportunities have been lost by researchers. Now, as it is possible that the time for more studies has past, we, as a result of not having consistency, will always be left in doubt aboutimportant effects of care. Finally, we presume that use of scales may discourage any worker committed to patient care to take part in an experimental study.

More attention should be placed on patient and family prospectives, in terms of detecting patient and carer satisfaction, quality of life, and family burden.

2. Specific

2.1 More reviews

Currently we know of no review comparing non-ICM with standard care and reporting relevant outcomes. This should be undertaken. In addition several good studies are excluded from this review as they are evaluating mixed models of care, or models plus other interventions such as Cognitive Behavioural Therapy. These studies do merit further attention from reviewers and could help clarify further ways by which either the ICM model develops or new and yet more holistic approaches evolve.

2.2 Developing this review

A full data set with all individual patient data would help avoid some biases and allow reanalysis using unified definitions of outcome. Any relevant studies in this area should make provision for prospectively providing data compatible with this review (Walsh-Connecticut).

2.3 More trials

We do not think that more trials comparing current ICM with standard care or non-ICM are justified. What should be researched are the features of ICM that may improve outcome, as it might be that the model of intervention is effective just because of some of its features. This work may involve more observational studies in order to evolve the ICM model to new and better packages of care.

PLAIN LANGUAGE SUMMARY.

Intensive case management for people with severe mental illness

Severe mental illnesses are defined by diagnosis, degree of disability and the presence of some abnormal behaviour. They include schizophrenia and psychosis, severe mood problems and personality disorder, and can cause considerable inconvenience over a long period of time both for the people are affected by them, and for their families and friends.

Until the 1970s it was common for those suffering from these disorders to stay in an institution for most of their lives, but now in most of the countries of the world, they are managed in the community with one of several different styles of intervention. Intensive Case Management (ICM) is one such intervention. It consists of management of the mental health problem and the rehabilitation and social support needs of the person concerned, over an indefinite period of time, by a team of people who have a fairly small group of clients (less than 20). It also offers 24 hour help and sees clients in a non-clinical setting.

This review compares ICM with non-Intensive Case Management (non-ICM; where people receive the same package of care but the professionals have caseloads of more than 20 people) and standard care (where people are seen as outpatients but their support needs are less clearly defined). Thirty-eight trials were found in the United States, Canada, Europe or Australia involving 7328 people in total.

When ICM was compared to standard care, those in the ICM group were significantly more likely to stay with the service, have improved general functioning, get a job, not be homeless and have shorter stays in hospital (especially when they had had very long stays in hospital previously). There was also a suggestion that it reduced the risk of death and suicide. If ICM was compared to non-ICM, the only clear difference was that those in the ICM group were more likely to be kept in care. There are no trials comparing non-ICM with standard care.

One of the drawbacks of this review is that the healthcare and social support systems of these countries are quite different, so it was quite difficult to make valid overall conclusions. In addition, much of the data on quality of life, and patient and carer satisfaction were not able to be used because the trials used many different scales of measuring these things, some of which were not validated. The development of such an overall scale and its validation would be very beneficial in producing services that people liked.

(Plain language summary prepared for this review by Janey Antoniou of RETHINK, UK www.rethink.org)

SUMMARY OF FINDINGS FOR THE MAIN COMPARISON

INTENSIVE CASE MANAGEMENT compared to STANDARD CARE for severe mental illness.

Patient or population: patients with severe mental illness

Settings: community

Intervention: INTENSIVE CASE MANAGEMENT

Comparison: STANDARD CARE

Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	STANDARD CARE	INTENSIVE CASE MANAGEMENT
Service use: Average number of days in hospital per month - by about 24 months (any sample size) Follow-up: 24 months		The mean Service use: Average number of days in hospital per month -by about 24 months (any sample size) in the intervention groups was 0.86 lower (1.37 to 0.34 lower)		3595 (24 studies)	⊕⊕○○ low 1,2
Service use: Average number of admissions	Study population		RR 0 (0 to 0)	0 (O3)	See comment
	See comment	See comment
	Medium risk population
Global state: Relapse	See comment	See comment	Not estimable	0 (04)	See comment
Global state: Leaving the study early - by long term Follow-up: > 12 months	Low risk population		RR 0.71 (0.59 to 0.85)	2766 (21 studies)	⊕⊕○○ low 1,5
	100 per 1000	71 per 1000 (59 to 85)
	Medium risk population
	300 per 1000	213 per 1000 (177 to 255)
	High risk population
	500 per 1000	355 per 1000 (295 to 425)
Mental state: Not improved to an important extent	See comment	See comment	Not estimable	0 (04)	See comment
Adverse event: Death - suicide - by long term Follow-up: > 12 months	18 per 1000	13 per 1000 (6 to 27)	RR 0.71 (0.34 to 1.51)	1723 (12 studies)	⊕○○○ moderate 6
Social functioning: Unemployed at the end of the trial - by long term	Low risk population		RR 0.7(0.49 to 1)	1129 (4 studies)	⊕○○○ very low1,7
	400 per 1000	280 per 1000 (196 to 400)
	Medium risk population
	800 per 1000	560 per 1000 (392 to 800)
	High risk population
	1000 per 1000	700 per 1000 (490 to 1000)

^*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence

High quality: Further research is very unlikely to change our confidence in the estimate of effect.

Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

Very low quality: We are very uncertain about the estimate.

¹Randomisation not well described, problematic to blind.

²I-square=74%; p<0.00001.

³Only skewed data available, not entering analysis.

⁴No studies reported this outcome.

⁵I-square=51%, p=0.005.

⁶Randomisation not well described.

⁷I-square=94%; p<0.00001.

ADDITIONAL SUMMARY OF FINDINGS

INTENSIVE CASE MANAGEMENT compared to NON-INTENSIVE CASE MANAGEMENT for severe mental illness.

Patient or population: patients with severe mental illness

Settings: COMMUNITY

Intervention: INTENSIVE CASE MANAGEMENT

Comparison: NON-INTENSIVE CASE MANAGEMENT

Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	NON-INTENSIVE CASE MANAGEMENT	INTENSIVE CASE MANAGEMENT
Service use: Average number of days in hospital per month - by about 24 months (any sample size) Follow-up: 24 months		The mean Service use: Average number of days in hospital per month -by about 24 months (any sample size) in the intervention groups was 0.08 lower (0.37 lower to 0.21 higher)		2220 (21 studies)	⊕⊕⊕○ moderate 1,2
Service use: Average number of admissions (skewed data - sample size ⧠200) - by long term Follow-up: 12 months		The mean Service use: Average number of admissions (skewed data -sample size ⧠200) - by long term in the intervention groups was 0.18 lower (0.41 lower to 0.05 higher)		678 (1 study)	⊕⊕⊕○ moderate 1,3
Global state: Relapse	See comment	See comment	Not estimable	0 (0)	See comment	Outcome pre-stated to be of interest - no data available from trials
Global state: Leaving the study early - by long term Follow-up: 12 months	Low risk population		RR 0.72 (0.52 to 0.99)	2195 (9 studies)	⊕⊕○○ low 1,4
	100 per 1000	72 per 1000 (52 to 99)
	Medium risk population
	200 per 1000	144 per 1000 (104 to 198)
	High risk population
	500 per 1000	360 per 1000 (260 to 495)
Mental state: Not improved to an important extent	See comment	See comment	Not estimable	0 (0)	See comment	Outcome pre-stated to be of interest - no data available from trials
Adverse event: Death - suicide - by long term	12 per 1000	11 per 1000 (3 to 34)	RR 0.88 (0.27 to 2.84)	1180 (4 studies)	⊕⊕○○ low 5,6,7
Social functioning: Employment status - spent less than 1 day employed - by medium term Follow-up: 12 months	Low risk population		RR 1.46 (0.45 to 4.74)	73 (1 study)	⊕⊕○○ low 1,3,8
	400 per 1000	584 per 1000 (180 to 1000)
	Medium risk population
	800 per 1000	1000 per 1000 (360 to 1000)
	High risk population
	1000 per 1000	1000 per 1000 (450 to 1000)

^*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence

High quality: Further research is very unlikely to change our confidence in the estimate of effect.

Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.

Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.

Very low quality: We are very uncertain about the estimate.

¹Randomisation not well described, problematic to blind.

²I-square=0%, p=0.49.

³Heterogeneity: not applicable, as only one study entering analysis.

⁴I-square 59%, p=0.01.

⁵Randomisation not well described.

⁶I-square=4%, p=0.31.

⁷Confidence interval (95%) 0.27 to 2.84.

⁸Confidence interval (95%) 0.45 to 4.74.

CHARACTERISTICS OF STUDIES

Characteristics of included studies [ordered by study ID]

Aberg-Wistedt-Sweden

Methods	Allocation: randomised. Design: single centre. Duration: 24 months. Country: Stockholm, Sweden.
Participants	Diagnosis: schizophrenic disorder (DSM-III-R). N=40. Setting: community setting. Age: 25-55 years, mean ~ 38 years. Sex: 65% M (26M, 14F). Ethnicity: no reported. History: recently admitted to ward or currently in outpatient department
Interventions	ICM: interdisciplinary team based. Caseload: 1:2.5. N=20. Standard care: multi-disciplinary psychiatric outpatient team, specialised in schizophrenic patients. Caseload: ~ 1:10/15. N=20
Outcomes	Global state: leaving the study early. Unable to use - Service use: average number of days in hospital per month and number of emergency visits (no mean, no SD). Quality of life: (no mean, no SD). Social network size: (measure validated on children only, no mean, no SD). Burden of care: (no mean, no SD).
Notes
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: not available. Secondary outcomes: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	The study did not address this outcome.
Free of selective reporting?	High risk	More outcomes of interest are reported incompletely.
Free of other bias?	Low risk	No details. No evident other bias are occurring.

Audini-UK

Methods	Allocation: randomised. Design: single centre. Duration: 15 months. Country: London, UK.
Participants	Diagnosis: serious mental illness (30% schizophrenia according DSM-III R). N=66. Setting: psychiatric community services. Age: 18-64 years, median ~ 37 years. Sex: 45% M (30M, 36F). Ethnicity: 26% Afro-Caribbean. History: mean ~ 0.2 admissions in last year; completed at least 18 months in ACT programme*. Excluded: primary addiction, primary organic brain damage.
Interventions	ICM: ACT (Stein&Test model). Caseload: 1:12. N=33.** Standard care: routine care from psychiatric services, as out-patients and/or in-patients as necessary, with community support services. N=33
Outcomes	Service use: average number of days in hospital per month***; admitted to hospital, average number hospital admissions***. Death: suicide. Global state: leaving the study early; GAF. Social functioning: SAS, employment status. Mental state: BPRS 24-items, PSE. Participant satisfaction: CSQ. Costs: costs of all care. Unable to use - Carer satisfaction: Relative’s Satisfaction (scale not peer-reviewed, attrition >50%)
Notes	Loss to follow-up: 12.1% * People in this study are recruited from the treatment arm of a trial on ACT 20-30 months long. **Authors report that “the team became depleted and demoralised” in the course of this trial. *** Variance not reported - data from another study used.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated - rating - Unclear. No details provided
Incomplete outcome data addressed? All outcomes	Low risk	Number and reason for attrition reported. Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups
Free of selective reporting?	Low risk	All listed outcomes are reported completely.
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Bjorkman-Sweden

Methods	Allocation: randomised. Design: single centre. Duration: 36 months. Country: Lund, Sweden.
Participants	Diagnosis*: serious mental illness, according to DSM-III-R. N=77. Setting: community psychiatric service. Age*: 19-55 years, mean ~ 37 years. Sex: 36M, 41F. Ethnicity: not reported. History: i. serious mental illness for >2yrs, ii. impairment due to illness (social-relationship, housing or work situation) for more than 2 yrs, iii. no primary diagnosis of substance or alcohol related disorders, iv. informed consent given
Interventions	ICM: Case Management service based on the Strength Model. Caseload: ~ 1:9. N=33. Standard care: psychiatric service comprehensive with a joint management for outpatient, in-patient and day care facilities. N=44
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services, admitted to hospital. Death: suicide. Global state: leaving the study early, GAF. Social functioning: Strauss-Carpenter Scale, social network (ISSI). Mental state: general symptoms, SCL-90. Quality of life: LQoLP. Participant satisfaction: CAN. Unable to use - Client satisfaction: questionnaire by the Swedish Institute for Health Services Development (modified version, not peer reviewed)
Notes	*51.9% schizophrenia-like disorder. **ICM group significantly older than standard care group (on average, 5 yrs older)
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised: used a computer random number generator.
Allocation concealment?	Unclear risk	Random selection performed by one of the trialist. No further details
Blinding? All outcomes	Unclear risk	Primary outcomes: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated -rating - Unclear. Interviewers formally blind to patient group allocation. Not tested
Incomplete outcome data addressed? All outcomes	Unclear risk	Analysis performed on an ITT basis, but “clients who were not available for or refused contact at follow-up were excluded from the respective analysis on an individual basis”
Free of selective reporting?	Low risk	All listed outcomes of interest reported.
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Bond-Chicago1

Methods	Allocation: randomised. Design: single centre. Duration: 12 months. Country: Chicago, Illinois, USA.
Participants	Diagnosis*: serious mental illness (schizophrenia, schizoaffective disorder, affective disorder, personality disorder) according to RDC and SADS. N=88. Setting: private psychiatric rehabilitation agency. Age: >18 years, mean ~ 34 years. Sex: 56.8% M (50M, 38F). Ethnicity: non-white 36%. History**: i. 3 admissions in the last 2 years either total of 5 admissions in life, ii. had no prior contact of more than 1 month’s duration with either study programme, iii. informed consent given
Interventions	ICM***: a large city adaptation of Assertive Community Treatment model according to Stein&Test. Caseload 1:10. N=45. Standard care***: provided from drop-in centre (day treatment, central meeting place, no requirement for frequent contacts). Caseload:1: = 20. N=43
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services, admitted to hospital, average number of admissions. Death: all causes, suicide. Global state: leaving the study early. Social functioning: police contact, arrested, imprisoned, accomodation status Unable to use - Global functioning: GAS (no SD, no sample size providing data). Social functioning: Areas of Difficulty Checklist (ADC) (scale not peer-reviewed, no SD). Quality of life: Life Satisfaction Checklist (LSC) (scale not peer-reviewed, no SD). Satisfaction with care: Satisfaction with Services (SWS), author modified version of CSQ (not peer-reviewed, no SD). Cost: no SD, no sample size.
Notes	*Schizophrenia ~ 37%; primary or secondary diagnosis of substance abuse = 26% . **At the time of study admission, 62% in ICM group and 54% in SC group were in a psychiatric hospital. ***Author reporting: “ACT had been in existence for 6 yrs, SC service for 2 yrs. In both programmes staff were enthusiastically committed to the respective programme philosophy”
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	Random assignment achieved by individual sealed envelops (not specified if opaque) , the assignment is carried out by a person unconnected to research programme
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: most are clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	High risk	Missing outcome data not balanced in numbers across intervention groups, but similar reasons for missing data across groups
Free of selective reporting?	High risk	Some outcomes of interest are reported incompletely and are not usable (due to missing SD or sample size providing data)
Free of other bias?	Low risk	Public funded (by the State Department of Mental Health). No further details. No evidence other bias are occurring

Bond-Indiana1

Methods	Allocation: randomised. Design: multi-centre, 3 centres (3 CMHCs)*. Duration: 6 months. Country: Indiana, USA.
Participants	Diagnosis**: psychotic disorder (DSM-III). N=67. Setting: Community Mental Health Centres (CMHCs). Age: >17 years, mean ~ 35 years. Sex: 61.6% M (103M, 64F). Ethnicity: black 34%. History***: high risk of hospitalisation, described as: i. discharged within last year and either rehospitalised ≥ 3 times in previous 2 years or determined by mental health services staff to be at high risk for readmission, ii. committed or awaiting commitment to hospitals, iii. presenting for admission at CMHC inpatient unit and having ≥ 4 psychiatric hospitalisations in last 2 years
Interventions	ICM: Assertive Case Management, according to the ACT model (Stein&Test), in addition to all other available mental health programme. Caseload: 1:8-12. N=84. Standard care: as provided at CMHCs (including case management services with large caseload). N=83
Outcomes	Service use: average number of days in hospital per month****, (data available for single centre, see below); admitted to hospital. Death: all causes (data from centre A only). Global state: leaving the study early. Social functioning: contact with the police (data from centre A only) Unable to use - Quality of Life: self-report instrument (no data, scale not peer reviewed, compiled by the therapist). Global state: compliance with medication (data not reported). Costs: available only for Centre A (no SD).
Notes	*Among client groups at the 3 centres, significant differences were found in: age, gender, race, education, employment, diagnosis and history at the baseline. **Schizophrenia-like disorder: 61%; substance abuse: 39%. ***Average 8.8 life-time hospitalisation and 1.5 hospitalisation in the previous year. **** Variance not reported - data from another study used.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised for single centre. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: most are clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Number of lost to follow up reported, but no reasons for missing data provided
Free of selective reporting?	High risk	Pre-specified outcomes not reported (medication compliance) or reported incompletely (days in hospital: SD missing)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Bond-Indiana1 (A)

Methods	Methods: see above Bond-Indiana1.
Participants	Participants: see above Bond-Indiana1. N=61.
Interventions	Interventions: see above Bond-Indiana1. ICM: N=29. Standard care: N=32.
Outcomes	Service use: average number of days in hospital per month*. Death:suicide. Social functioning: contact with the police. Unable to use Costs: total inpatient cost, total treatment cost (no SD).
Notes	* Variance not reported - data from another study used.

Bond-Indiana1 (B)

Methods	Methods: see above Bond-Indiana1.
Participants	Participants: see above Bond-Indiana1. N=64.
Interventions	Interventions: see above Bond-Indiana1. ICM: N=34. Standard care: N=30.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Variance not reported - data from another study used.

Bond-Indiana1 (C)

Methods	Methods: see above Bond-Indiana1.
Participants	Participants: see above Bond-Indiana1. N=42.
Interventions	Interventions: see above Bond-Indiana1. ICM: N=21. Standard care: N=21.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Variance not reported - data from another study used.

Bush-Georgia

Methods	Allocation: randomised. Design: single site. Duration: 12 months. Country: Atlanta, USA.
Participants	Diagnosis*: severe mental illness. N=28. Age: 25-56 years. Sex: 57% M (16M, 12F). Ethnic: 50% African-American. History: high rates of hospital readmissions (2-18 previous admissions), difficulty in community living
Interventions	ICM: clinical case management, providing intensive support and outreach (according to the Stein&Test model TCL). N=14. non-ICM: standard care providing case-management at a lower level of intensity and rehabilitation services. N=14
Outcomes	Service use: average number of days in hospital per month**. Death: all causes, suicide. Unable to use - Service use: no of hospital admissions (no individual group data); emergency room visits (no individual group data). Global state: compliance: adherence to service and medication plan (incompletely reported data). Social functioning: appropriate living status (incompletely reported data). Costs: no individual group data.
Notes	* 86% schizophrenia. ** Variance not reported - data from another study used.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcome: leaving the study early - clinician/participant mediated - rating -Unclear
Incomplete outcome data addressed? All outcomes	Low risk	No missing outcome data.
Free of selective reporting?	High risk	Outcomes not pre-stated. Most of the reported outcomes are reported incompletely (not usable data)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Chandler-California1

Methods	Allocation: randomised. Design*: multi-centre, 2 sites** (site A: Long Beach, urban site; site B: Stanislaus County, rural site). Duration: 36 months. Country: California, USA.
Participants	Diagnosis: serious & persistent mental illness (DSM-III R). N=516. Setting: 1 urban, 1 rural but integrated service agencies. Age: 30% > 45 years. (N=439). Sex: 52% M. (N=439). Ethnicity: 26% non-white. (N=439). History**: i. functional impairment due to mental disorder, ii. eligibility for public assistance, iii. not a primary diagnosis of substance abuse, iv. informed consent given
Interventions	ICM***: Assertive Community Treatment provided by integrated service agencies, according to Training in Community Living Programme (Stein&Test). Caseload: 1:10. N=252. Standard care: usual mental health service (i.e. outpatients, day treatment, case management, inpatient, minimal rehabilitation services). N=264
Outcomes	Service use: average number of days in hospital per month (available for single centre, see below). Following outcomes available for single centre. Service use: not remaining in contact with psychiatric services, admitted to hospital. Global state: leaving the study early. Social functioning: employment; arrested, imprisoned; accomodation status. Costs: cost of psychiatric hospital care. Unable to use - Mental state: general symptoms: Colorado Symptoms Index (no data reported). Self-esteem: New York Self-Esteem Scale (no data reported). Quality of life: Lehman’s Quality of Life Instrument (no data reported). Social functioning: level of social activities (not peer-reviewed scale). Family burden: sub-scales adapted from Tessler’s Family Burden Interview (not peer reviewed, attrition > 50%). Participant satisfaction: scale for overall satisfaction with mental health services (not peer-reviewed scale). Costs: direct costs of health care and of all care (no SD), all mental health care (not listed as review outcome of interest)
Notes	*61% schizophrenia, N=439. **28% admitted in previous year, N=439. ***Intervention programme in 2 sites slightly different: Site A puts more emphasis on employment services, social and therapeutic activities. Site B emphasis, avoiding hospitalisation through use of crisis house
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised for single centre. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - unclear. Secondary outcomes: clinician/participant mediated - rating - unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Number of lost to follow up reported, but no reasons for missing data provided. LOCF for continuous data
Free of selective reporting?	High risk	Listed outcomes of interest not reported (continuous data from scales not reported; days in hospital reported only for site A, no SD)
Free of other bias?	Low risk	Public funded (California Department of Mental Health, MIMH). No details. No evidence other bias are occurring

Chandler-California1 (A)

Methods	Site A: Long Beach. Methods: see above Chandler-California1.
Participants	Participants: see above Chandler-California1. N=256.
Interventions	Interventions: see above Chandler-California1. ICM= 127. SC= 129.
Outcomes	Service use: average number of days per month in hospital; not remaining in contact with psychiatric services, admitted to hospital. Global state: leaving the study early. Social functioning: employment status, arrested, imprisoned, accomodation status. Costs: direct costs of psychiatric hospital care.
Notes

Chandler-California1 (B)

Methods	Site B: Stanislaus County. Methods: see above Chandler-California1.
Participants	Participants: see above Chandler-California1. N=260.
Interventions	Interventions: see above Chandler-California1. ICM= 125. SC=135.
Outcomes	Service use: average number of days per month in hospital, not remaining in contact with psychiatric services, admitted to hospital. Global state: leaving the study early. Social functioning: employment, arrested, imprisoned, accomodation status. Costs: direct costs of psychiatric hospital care.
Notes

Curtis-New York

Methods	Allocation: randomised. Design: single centre. Duration: 14 months. Country: New York, USA. Follow up*: ranging from 18 to 52 months.
Participants	Diagnosis**: not stated, DSM III. N=292. Setting: Harlem Hospital Centre (HHC). Age: age 18-54 years, mean ± SD 35.9 ± 12.1 yrs (N=430). Sex: 59% F. Ethnicity: 91% Black. History***: i. about to be discharged from hospital, ii. local residents, iii. without a sole diagnosis of substance abuse or organic mental disorder, iv. in-patients for >7 days, and not eligible for the ‘community support system’ programme - that is no psychiatric admission of >6 months duration/three admissions of >10 days within last 2 years, v. informed consent given
Interventions	ICM: Intensive outreach case management from a multi-disciplinary team at HHC. This implemented a discharge treatment plan, and monitored clinical and social problems. The team did not ‘assume direct responsibility for care but [..] help[ed] the patient enrol in a day hospital programme, adult mental health clinic, rehabilitation programme, or alcohol treatment programme’. Caseload:1:17. N=147 Standard Care: routine aftercare, within the discharge treatment plan prescribed for each patient from Harlem Hospital Centre (HHC); “most received at least initial treatment form various divisions of the departments of psychiatry within the Health and Hospitals Corporation”. N=145
Outcomes	Service use: average number of days in hospital per month, admitted to hospital. Death: all causes and suicide. Unable to use - Use of ambulatory services: this outcome is not listed as outcome of interest for the review. Quality of life: measuring instrument written by trialists for this particular trial and was not published in peer-reviewed (EAF - Evaluation Aftercare Form)
Notes	*Follow up period variable. ** Schizophrenia 38%; alcohol or drug abuse or dependence 39%. *** Mean number previous admissions >1. Some more severely ill clients not included in this part of study as they were eligible for ‘Community Support System’ programme group
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: some are clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Missing data are not addressed.
Free of selective reporting?	Low risk	Listed outcomes are reported completely.
Free of other bias?	Low risk	Funded by public institution (‘New York City Health and Hospitals Corporation’ and foundations). No details. No evidence other bias are occurring

Drake-NHamp

Methods	Allocation: randomised. Design: multi-centre (7 sites: Community Mental Health Centres - 2 sites urban area; - 5 sites rural area). Duration: 36 months. Country: New Hampshire; USA.
Participants	Diagnosis: schizophrenia, schizoaffective or bipolar disorder (DSM-III-R) and active substance user disorder within past 6 months (DSM-III-R)*. N=225. Setting: Community Mental Health Centres. Age: 18-60 years, mean ~ 34 years. Sex: 74.4% M (167M, 58F) . Ethnicity: ethnic minority 3.4%. History: i. no mental retardation or medical conditions that would prevent participation, ii. written informed consent given
Interventions	ICM: ACT teams with special training in substance abuse treatment. Caseload:1:12. N=109. non-ICM: standard non-Intensive Case Management, incorporating most of the ACT principle, but providing less individual service for substance abuse. Caseload:1:25. N=114
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services. Death: all causes. Global state: leaving the study early. Mental state: BPRS-24 items. Social functioning: average days in stable accomodation, Alcohol Use Scale (AUS), Substance Abuse Treatment Scale (SATS), alcohol use days (Time Line Follow Back-TLFB) , remission from alcohol use disorder. Quality if life: Quality of Life Index (QOLI). Unable to use - Substance abuse: drug use scale (DUS), drug use days (Time Line Follow Back-TLFB) and not in remission: attrition >50%, Addiciton Severity Index (ASI) (data not reported). Service use: Service Utilization Interview (data not reported). Costs: direct costs of psychiatric hospital care (no SD), average total study cost (not listed as outcome of interest in the review).
Notes	* 53% schizophrenia, 22% schizoaffective, 24% bipolar disorder, 72.6% alcohol abuse, 41.8% drug abuse
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: some are clinician/participant mediated - rating - Unclear. Some are interviewer rated - rating - Unclear. Interviewer blind, not tested
Incomplete outcome data addressed? All outcomes	Unclear risk	Number of lost to follow up is reported, but reason for missing data is not reported for each randomised group of treatment (it is reported overall, information not usable)
Free of selective reporting?	Low risk	Pre-specified outcomes of interest are reported.
Free of other bias?	Low risk	No information available. No evidence other bias are occurring

Drake-NHamp (A)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=23.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=11. Standard care: N=12.
Outcomes	Service use: average number of days in hospital per month.
Notes

Drake-NHamp (B)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=23.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=10. Standard care: N=13.
Outcomes	Service use: average number of days in hospital per month.
Notes

Drake-NHamp (C)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=25.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=14. Standard care: N=11.
Outcomes	Service use: average number of days in hospital per month.
Notes

Drake-NHamp(D)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=60.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=30. Standard care: N=30.
Outcomes	Service use: average number of days in hospital per month.
Notes

Drake-NHamp (E)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=32.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=17. Standard care: N=15.
Outcomes	Service use: average days per month in hospital.
Notes

Drake-NHamp (F)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=17.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=9. Standard care: N=8.
Outcomes	Service use: average number of days in hospital per month.
Notes

Drake-NHamp (G)

Methods	Methods: see above Drake-NHamp.
Participants	Participants: see above Drake-NHamp. For this centre, sample providing data: N=16.
Interventions	Interventions: see above Drake-NHamp. For this centre, sample providing data: ICM: N=7 Standard care: N=9.
Outcomes	Service use: average number of days in hospital per month.
Notes

Essock-Connecticut1

Methods	Allocation: randomised. Design*: multi-centre, 3 sites, all in city area (data for single centre not available). Duration: 18 months. Country: Connecticut, USA.
Participants	Diagnosis**: major depression, bipolar disorder, schizophrenia, schizoaffective disorder according to DSM III-R. N=262. Age: mean ~ 41 years. Sex: 64% M. History***: high level of service use (defined as: i. ≥ 2 psychiatric hospitalisations in last 2 years, ii. 1 psychiatric hospitalisation longer than 180 days in last yr, or iii. ≥ 2 contacts with crisis services in last 2 years) and significant difficulties meeting the demand of everyday life (defined as: 1. homeless in past year, or ii. requiring extensive supervision or assistance at least weekly to meet personal car needs), informed consent given
Interventions	ICM: Assertive Community Treatment teams (Stein&Test model), but having a richer staff and achieving 24-hrs coverage. Caseload 1:5-7. N=130****. non-ICM: generalist model, but providing case managers mobile. Caseload 1:25-30. N-132****
Outcomes	Service use: average number of days in hospital per month. Death: all causes. Global state: leaving the study early. Unable to use - Mental state: modified version of SCL-90 (no peer-reviewed scale, no SD). Social functioning: accomodation status, number of days homeless (no SD), imprisoned (not reported). Quality of life: modified version of QOLI (no peer-reviewed scale, no SD). Carer satisfaction: Family Burden Interview (no SD). Costs: data reported incompletely.
Notes	*Authors state that the same treatments provided in different sites are not identical as “different styles of providing services were used at different sites”, but single centre data are not available. **67% schizophrenia or schizoaffective disorder. ***62% individuals have life-time history of substance abuse or dependence and 25% individuals are hospitalised at the study entry
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. Randomisation counterbalanced “so that within each site clients had 50% likelihood of being assigned to either ICM or non-ICM and counterbalanced so that clients hospitalised at the time of assignment had 50% likelihood of being assigned to either ICM or non-ICM.”
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: leaving the study early - clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Low risk	Number and reason for missing data clearly reported. Missing outcome data balanced across intervention group
Free of selective reporting?	High risk	Listed outcome of interest are not reported or are reported incompletely
Free of other bias?	Low risk	Funded by public health institutes and private foundation. No other details provided. No evidence other bias are occurring

Essock-Connecticut2

Methods	Allocation: randomised. Design: multi-centre, 2 urban sites (data for single centre not available). Duration: 36 months. Country: Bridgeport, Connecticut, USA.
Participants	Diagnosis*: schizophrenia, schizoaffective disorder, depression with psychotic features, bipolar disorder - DSM III-R. N**=198 (randomised to site A: N=100; to site B: N=98). Setting: 2 state-operated out-patients CMHCs. Age: mean 36.5 years (SD 7.8 years). Sex: 72% M (143M, 55F). Ethnicity: African-American 55%; Hispanics 14%. History: i. active substance use disorder (abuse or dependence on alcohol or other drugs previous 6 months), ii. high service use previous 6 months (≥ 2 between: psychiatric hospitalisation, stays in a psychiatric crisis or respite programme, emergency department visit or incarceration), iii. homeless or unstably housed, iv. poor independent living skills, v. no pending legal charge, vi. no medical condition or mental retardation, vii. written informed consent given
Interventions	ICM***: Assertive Case Management. Caseload: 1:10-15. N=99. non-ICM***: Clinical Case Management provided by clinicians. Caseload: 1:20-30. N=99
Outcomes	Service use: average number of days in hospital per month. Unable to use - Global state: leaving the study early, GAS (not reported N for single intervention group) Social functioning: days in stable accomodation (not reported N for single intervention group), imprisoned (days in jail reported together to days in hospital). Social functioning: SATS, AUS, DUS (scales rated by clinicians), days of alcohol/drug use (not reported N for single intervention group). Mental state: BPRS - 24 items (not reported N for single intervention group). Client satisfaction: General Life Satisfaction (not reported N for single intervention group). Costs: not reported.
Notes	* 76% schizophrenia or schizoaffective disorder; 17% mood disorder; 74% alcohol disorder; 81% substance use disorder. ** Original randomised sample N=205. 7 participants left for administrative reason. Authors did not report what group they were assigned to. *** Both interventions addressed substance use disorders.
Risk ofbias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised using separate computer-generated randomisation stream for each site
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	No details.
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of missing data (number and reason for missing data are reported for the total sample, not for each intervention group). Seven randomised subjects were lost for administrative reason, but their intervention allocation was not reported
Free of selective reporting?	High risk	Listed outcomes of interest are fully reported for each site, but authors do not provide sample size data are referred to
Free of other bias?	Low risk	Public funded. No further details. No evidence other bias are occurring

Ford-UK

Methods	Allocation: randomised. Design: single centre (it is a multi-site study, but randomisation is used only in one site, we report data from this site). Duration: 18 months. Country: North Southwark, London, UK.
Participants	Diagnosis*: psychotic illness, severely and persistently mentally ill (criteria not reported) N=77. Setting: Centre for Mental Health. Age: mean ~ 46 years. Sex: 47% M. Ethnicity: 30% ethnic minority. History: i. either: recent in-patient admission, or impairment in social functioning, or problems in compliance with medication/treatment regimens, or problem in receiving multi-agency; ii. no primary diagnosis of organic psychosis, personality disorder, drug/ alcohol abuse, learning difficulty
Interventions	ICM: multidisciplinary team, providing assertive community care not following any specific model of case management. Caseload: <15 clients per case manager. ICM provided in addition to standard mental health service. N=39. Standard care: provided from psychiatric services. N=38.
Outcomes	Service use: average number of days in hospital per month, admitted to hospital. Death**: all causes and suicide. Global state: leaving the study early, compliance with medication. Social functioning: imprisoned. Mental state: BPRS-18 items. Quality of life: objective quality of life through QOLI. Costs: direct costs of psychiatric hospital care, direct costs of all care Unable to use - Social functioning: Life Skill Profile (LSP) (rated by the therapist, not independent rater) Social support: data not reported due to “high level of missing data”
Notes	* Schizophrenia 81%. **Difference of 1 death (all causes) between two reports (3 to 4 deaths reported in ICM group). (Number of suicide matches between reports). Here reported lowest number of death
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised. Sequence of random number generated by computer programme
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: some are clinician/participant mediated - rating - Unclear. Some are interviewer rated - rating - Unclear. No details
Incomplete outcome data addressed? All outcomes	High risk	Number and reason for missing dichoto-mous outcome data reported; imbalance in number and reason for missing data across intervention group. For continuous outcome data, this issue is not addressed
Free of selective reporting?	Low risk	Listed outcome of interest reported and if not reported, reason is provided (i.e. social support data are not reported due to high level of missing data)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Hampton-Illinois

Methods	Allocation: randomised. Design: multi-centre (2 centres: Site A - Bridge West; Site B - Bridge South). Duration: 12 months. Country: Chicago, USA.
Participants	Diagnosis*: serious mental illness. N=165. Setting: community mental health service. Age: ≥ 18 years, mean ~ 37 years. Sex: 77% M (127M, 38F). Ethnicity: 55.7% black. History: i. admitted to inpatient units of 2 state hospitals, ii. homeless on admission or at risk of homelessness on discharge, iii. informed consent given (site A)
Interventions	ICM: Assertive Case Management (Stein&Test model). Caseload: ~ 1:10. N-82. Standard care: provided from psychiatric services. N-83. (Site A included traditional office-based outpatient care and case management)
Outcomes	Outcomes from different centres are reported separately, as we assumed a single centre randomisation procedure. See below in Hampton-Illinois (A) and Hampton-Illinois (B).
Notes	*Schizophrenia 42%.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details. Although not explicit, authors of review assumed an independent randomisation for each centre. Authors reported: “Project staff recruited clients on varying days to reduce bias” - authors of this review interpreted this as meaning that different days were used for recruitment to the whole study and not that group of allocation was determined by day
Allocation concealment?	Unclear risk	Described as “blinded randomisation”. No further details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: some clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of attrition/exclusions to permit judgement
Free of selective reporting?	High risk	In both centres, Hampton-Illinois (A) and Hampton-Illinois (B), Some listed outcomes of interest are not reported or reported incompletely (i.e. days spent in hospital: no SD)
Free of other bias?	Unclear risk	Public funded (National Institute of Mental Health and Developmental Disabilities). In Hampton-Illinois (B) It is unclear whether the study was interrupted earlier. Authors reported “The study sample for Site B was inadequate due to late start, implementation problems and high dropout rate” because of this, they reported data only by 6 months

Hampton-Illinois (A)

Methods	Site A - Bridge West. Methods: see above Hampton-Illinois.
Participants	Participants: see above Hampton-Illinois Site A provided following information: Diagnosis: 78% (74/95) primary diagnosis of major mental illness (schizophrenia, affective disorder, other psychosis) , ~13% (12/95) have primary diagnosis of alcohol or substance abuse. N=95. Sex: 80% M (76M;19F). Ethnicity: non-white: 50.5% (48/95). Age: average ~ 38 years. History: average age at first psychiatric contact ~ 23.5 years, average duration homeless before entering the study ~ 44.6 months
Interventions	Interventions: see also above Hampton-Illinois. Site A provided following information: ICM: N=48. Standard care: provided from psychiatric services, including traditional office-based outpatient care and case management. N=47
Outcomes	Service use: average number of days in hospital per month*, not remaining in contact with psychiatric services. Death: all causes. Global state: leaving the study early. Social functioning: accomodation status. Unable to use - Service use: mean number of admissions (data not reported). Social functioning: imprisoned (data not reported).
Notes	* Variance not reported - data from another study used.

Hampton-Illinois (B)

Methods	Site B - Bridge South. Methods: see above Hampton-Illinois.
Participants	Participants: see above Hampton-Illinois. SIte B provided following information: Diagnosis: 54% (38/70) have primary diagnosis of schizophrenia, 27% (19/70) have primary diagnosis of affective disorder, 1% (1/70) have primary diagnosis of substance abuse. N=70. Sex: 71% M (M59/70) Age: average ~ 35 years. History: average age at first psychiatric contact ~ 23 years, average homeless before entering the study ~ 18 months
Interventions	Interventions: see above Hampton-Illinois. Site B provided following information: ICM: N-34. Standard care: N-36.
Outcomes	Service use: average number of days in hospital per month*. Global state: leaving the study early. Unable to use - ** Service use: not remaining in contact with psychiatric services. Death: all causes. Social functioning: accomodation status, imprisoned.
Notes	* Variance not reported - data from another study used. ** Data excluded due to a incomplete report (figures in project report do not add up - > than total N in study)

Harrison-Read-UK

Methods	Allocation: randomised. Design: single centre. Duration: 24 months. Country: London, UK.
Participants	Diagnosis*: “heavy psychiatric service users”, no diagnostic criteria reported. N=193. Setting: community psychiatric services. Age: 16-64 years, mean ~ 39 years. Sex: 53% M (102M, 91F). History: admitted within last the last 3 years and had at least 2 admissions in the last 6. 5 years. Excluded: participants who were continuously hospitalised during 8 months recruitment
Interventions	ICM: enhanced community management on ACT principles (Stein model) provided by dedicated multi-professional team. Caseload: 1:8-15. N=97. non-ICM: locality-based community psychiatric services using the UK Care Programme Approach. Caseload: 1≥20. N=96
Outcomes	Service use: average number of days in hospital per month. Death: all causes, suicide. Global state: leaving the study early, compliance, Rating of Medication Influences (ROMI). Social functioning: Social Functioning Questionnaire (SFQ). Mental state: Krawiecka Scale (KS)‘Health of the National Outcome Scale (HoNOS), Hospital Anxiety Depression Scale (HAD-S). Participant satisfaction: Camberwell Assessment of Need (CAN). Costs: direct costs of psychiatric hospital care Unable to use - Mental state: general symptoms, Well-being Questionnaire-WQBQ (not peer-reviewed scale and modified from the original)
Notes	*Schizophrenia ~65%. **Median number of 5 admissions over 6.5 years.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Random sequence generated by computer programme.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: some are interviewer rated - rating - No. Interviewers are not blind to treatment assignment
Incomplete outcome data addressed? All outcomes	Unclear risk	Number of lost to follow up reported, but reason for attrition is not reported. Number of lost to follow up balanced between two group. Some participants are excluded after randomisation, but reason for exclusion is not stated
Free of selective reporting?	Low risk	Listed outcomes of interest are fully reported.
Free of other bias?	Low risk	Public funded (Nation Health Service). No further details. No evidence of other bias

Herinckx-Oregon

Methods	Allocation: randomised. Design: single centre. Duration: 29 months. Country: Portland, Oregon, USA.
Participants	Diagnosis: schizophrenia, major affective disorder, paranoid disorder or another severe mental disorder; diagnostic criteria not reported. N=178. Setting: community. Age: >18 years, mean 36.5 ± 10.3 years (N-163). Sex: 61% M (N-163). Ethnicity: 18% black. *History: i. chronically mentally ill, ii. history of persistent psychotic symptoms not due to substance abuse, iii. impaired functioning in > 2 of (i) social role, (ii) daily living, (iii) social acceptability, iv. no mental retardation. In process of being discharged from hospital or transferring to new service providers within community
Interventions	ICM**: Assertive Community Treatment from combined teams staffed by consumers (N-58) and not staffed by consumers (N=59). ACT following the Stein &Test model. Caseload: 1:10N-117. Standard care***: provided by 1 of 4 CMHCs and a number of smaller, more specialised agencies (none providing assertive outreach). Average caseload ~ 1:27. N-61
Outcomes	Service use: not remaining in contact with psychiatric services****, admitted to hospital, number visits to ER. Social functioning: accomodation status, arrests. Unable to use - Social functioning: employment status, illicit drug use (not reported). Mental state: general symptoms (not reported). Quality of life: measurement instrument not specified (data not reported). Satisfaction with services: measurement instrument not specified (not reported)
Notes	*60% psychotic disorders, 40% affective disorders, 33% severe alcohol or drug use comorbidity, 61% 2+ admissions in last 6 months ** Staff members were self identified mental health consumers DSM-R axis I diagnosis (~ 50% of staff had a diagnosis of bipolar disorder) . *** In the Standard Care group: caseload ~ 1:15 is provided to ~ 33% participants. **** Disengagement does not include who moved out, who refused to be re-interviewed, death
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. Ratio randomisation between interventions: ICM:SC = 2:1. No further details
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Missing data are not addressed.
Free of selective reporting?	High risk	Not all listed outcomes of interest are reported.
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Holloway-UK

Methods	Allocation: randomised. Design: single centre. Duration: 18 months Country: East Lambeth, London, UK.
Participants	Diagnosis*: hospital diagnosis of functional psychosis (no diagnostic criteria stated). N=70. Setting: community psychiatric service. Age: 16-64 years, mean - 35 years. Sex: 66% M (46M, 24F). Ethnicity: not available. History**: referred by teams as being “hard to treat” (previous non-compliance with treatment, frequent readmission or poor symptomatic response to conventional management), ii. live locally, iii. informed consent given
Interventions	ICM: Intensive team-based Case Management. “Continuing care team” providing Assertive Case Management (according to the “Clinical Case Management Model” developed by Kanter): Caseload: 1:8. N-35. Standard care: provided by community psychiatric nursing service (CPNS). N-35
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services, admitted to hospital. Death: all causes and suicide. Global state: leaving the study early. Social functioning: Disability Assessment Scale (DAS). Mental state: Comprehensive Psychopathological Rating Scale (CPRS), Schedule for the Assessment of Negative Symptoms (SANS), depression, Beck Depression inventory (BDI). Quality of life: Lancashire Quality of Life Profile (LQoLP). Behaviour: Social Behaviour Scale (SBS). Unable to use - Participant satisfaction: satisfaction interview (not peer reviewed, scale developed by the research team)
Notes	*66% schizophrenia/schizoaffective disorder; 26% bipolar disorder. **All had one or more psychiatric admission, years since onset of illness ≃ mean 11 years
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	Randomisation by sealed envelops (not stated if opaque).
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: most are interviewer rated - rating - No. Interviewers are not blind to treatment condition
Incomplete outcome data addressed? All outcomes	Low risk	YES - Primary outcomes: average number of days in hospital per month; not remaining in contact with psychiatric services. No missing data NO - Secondary outcomes: imbalance in numbers for missing data across intervention groups
Free of selective reporting?	Low risk	All listed outcomes of interest are fully reported.
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Jerrell-SCarolina1

Methods	Allocation: randomised. Design: single centre trial. Duration: 18 months. Country: South Carolina, USA.
Participants	Diagnosis*: psychotic or major affective disorders (DSM-III R, according to the Computer based Diagnosis Interview Schedule Revised C-DIS-R). N=122. Age: 18-59 years, 55% > 34 years. Sex: 68M, 54F. Ethnicity: ~28% non white. Setting: large urban mental health service. History: i. Participants were being discharged from the most recent of 2+ in-patient admissions in last year or subacute care episodes or lengthy residential treatment & repeated emergency psychiatric visits; ii. 2+ of poor work history, eligible for public assistance, poor living skills, poor social support, history of inappropriate behaviour
Interventions	**ICM (1): Programme Assertive Community Treatment (PACT) adaptation. Caseload: 1:15/20. N=40. **ICM (2): Intensive Broker Case management Model. Caseload: 1:15/18. N=42. Standard Care: from one of 4 multi-disciplinary psychiatric teams. Clinical approach according to a generalist model with supplemental case management provided to ≃ 25% of the most unstable clients. Caseload: 1:35 to 1:45. N=40
Outcomes	Service use: average number of days in hospital per month***, average visit to emergency room***. Social functioning: Social Adjustment Scale-II (SAS-II), Role Functioning Scale (RFS) Unable to use - Mental state: not reported the psychometric measurement instrument. Social functioning: use of alcohol and drug, legal system involvement (not reported). Participant satisfaction: satisfaction with service (instrument not stated, data not reported), Satisfaction with Life scale (not reported). Costs: direct costs of psychiatric inpatient care (no SD).
Notes	*Schizophrenia ~75%. **Both Intervention (1) and Intervention (2) would fit into the definition of Intensive Case Management as described in this review, hence they could be considered as a single intervention. But as they are reported separately in the original study, it was not possible to present data from these two samples summed up together (it is not possible sum up SD data). We decided to present only data from Intervention (1) vs Standard Care. *** Variance not reported - data from another study used.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details
Allocation concealment?	Unclear risk	No informations.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated -rating - Unclear. No informations
Incomplete outcome data addressed? All outcomes	Unclear risk	Number of randomised participants is not stated, as it is reported only number of randomised participants completing the study period
Free of selective reporting?	High risk	Listed outcome of interest not reported or reported incompletely (i.e. service use: no SD)
Free of other bias?	Low risk	Public funded (National Institute of Mental Health). No details. No evidence other bias are occurring

Johnston-Australia

Methods	Allocation: randomised. Design: single centre. Duration: 12 months. Country: Sydney, Australia. Setting: Eastern Suburb Mental Health Service (ESMHS).
Participants	Diagnosis*: schizophrenia, bipolar disorder (diagnostic criteria not reported). N=73. Setting: Eastern Suburb Mental Health Service (ESMHS). Age: 16-70 years, mean - 42 years. Sex: 56% M, (41M, 32F). History: at least three of: i. high relapse rate over previous 2 years, ii. poor compliance, iii. disturbing behaviour, iv. frequent changes of accommodation, v. poor budgeting skills, vi. low quality of life, vii. difficulty to manage in existing service. Resident of the ESMHS catchment area. No primary diagnosis of substance misuse, organic brain disorder or intellectual disability
Interventions	ICM**: Caseload: 1:8-10. N-37. non-ICM: Caseload: 1:20-40. N-36.
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services, admission to hospital. Death: all causes. Global state: leaving the study early, compliance with medication. Social functioning: accomodation status, number living in supported accomodation, employment, participant spending at least 1 day employed, participant on paid employment, number of participant having contact with police or legal system. Behaviour: number of participant having incident of self-harm or harm to others. Costs: direct costs of all care. Unable to use - Service use: number admitted to hospital (not reported), use of general practitioner (not listed as review outcome of interest). Global state: clinically significant improvement (as LSP improvement −18 points/12 months) (scale assessment completed by the therapist, not reported what measurement used). Social functioning: accomodation changes (not listed as review outcome of interest), Life Skill Profile (LSP) (assessment completed by the therapist). Costs: direct costs of psychiatric hospital care (no SD).
Notes	*Schizophrenia 89%. **Main difference between teams is in ratio staff to patient. Both are multidisciplinary, coordinate and provide variety of service, have access to inpatient, rehabilitation and 24 hrs crisis service
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: most are clinician/ participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Low risk	YES - Primary outcomes: average number of days in hospital per month, not remaining in contact with psychiatric services. Number and reason for missing data clearly reported and balanced between group NO - Secondary outcomes: imbalance in numbers for missing data across intervention groups
Free of selective reporting?	Low risk	All listed outcome of interest are fully reported
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Lehman-Maryland1

Methods	Allocation: randomised. Design: single centre. Duration: 12 months. Country: Baltimore, Maryland, USA.
Participants	Diagnosis*: severe mental disorder (schizophrenia, schizoaffective disorder, or any other diagnosis Axis I and extensive prior hospitalisation history - DSM III-R). N=152. Setting: CMHCs. Age: 18-64 years, mean ~ 37 years. Sex: 67% M (102M, 50F). History**: i. homeless***, ii. severe mental disorder****, iii. written consent given
Interventions	ICM: Programme Assertive Community Treatment (Stein&Test model). Caseload: 1:10-12. N=77. Standard care: care: from community mental health centres and emergency facilities, though also a small amount of non-Intensive Case Management. N=75
Outcomes	Service use: average number of days in hospital per month, admitted to hospital, average number of admission to ER (mean adjusted for race as covariate). Global state: leaving the study early. Social functioning: not living independently, days in stable accomodation. Mental state: Colorado Symptoms Index (CSI) (mean adjusted for race as covariate). Quality of life: Quality of Life Index (QOLI), satisfaction with general well-being (mean adjusted for race as covariate). Costs: direct cost of psychiatric hospital, direct costs of all health care Unable to use - Social functioning: days in prison (reported data are not complete). Quality of life: specific items reported from QOLS (not global assessments). Social functioning: objective QOLS (no data), days homeless (split reporting of different types of homelessness, no SD). General Health: Medical Outcomes Study 36-Items Short Form Health Survey (SF36) (mean adjusted for race as covariate) (not listed as outcome of interest for the review
Notes	*Schizophrenia-like disorder: 58% bipolar disorder 20%, major depression 8%, comorbid for substance use disorders 71%. **74% homeless for at least 1 year total; 34 %homeless for ≥4 years. ***Homeless defined as: on street or shelter for ≥5 days last 45 or ≥14 last 180, or in temporary accommodation with ≥ 2 residential moves in last 6 months. ****Severe mental disorder: defined as diagnosis of schizophrenia or schizophrenia-like illness or receiving benefit because of mental disorder or had another axis I disorder & either > 2 hospitalisations of > 21 days in past 3 years or a total of > 42 days prior to current hospitalisation or =90 days in psychiatric hospital or nursing home in past 3 years or mental disability lasting > 1 year during which not able to spend > 75% of time in some gainful activity. Note complex inclusion criteria.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised (stratified random assignment). No further details
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: those interviewer rated - rating - Unclear. No details
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of attrition/exclusion (no reasons for missing data provided)
Free of selective reporting?	High risk	Not all listed outcomes of interest are reported completely.
Free of other bias?	Low risk	Public funded (Centre for Mental Health Services, Maryland). No details. No evidence other bias are occurring

Macias-Utah

Methods	Allocation: randomised. Design: single centre. Duration: 18 months. Country: Utah, USA.
Participants	Diagnosis*: serious and persistent mental disorder. N=41. Setting: mental health centre. Age:not reported. Sex: 56% M (21M, 16F). N-37. Ethnicity: 100% Caucasian. History: unclear, no primary diagnosis of mental retardation or substance abuse
Interventions	ICM: psychosocial rehabilitation programme at CMHC + Case Management (CM is modelled as Strengths CM). Caseload: 1:20. N-20. Standard care: psychosocial rehabilitation programme** at CMHC. N-21
Outcomes	Service use: admitted to hospital. Global state: leaving the study early. Unable to use - Mental state: Brief Psychological Well-Being Index (BPWI) (unpublished scales, designed especially by authors for the population under study). Global state: Self Report Inventory (SRI) (unpublished scales, designed especially by authors for the population under study). Carer satisfaction: Utah Family Burden Scale (unpublished, designed especially by authors for the population under study). Social functioning: Utah Case Management Consumer Assessment Record (not independently rated, no summary score, no SD)
Notes	*Schizophrenia 46%, major depression 22%. ** described as “a high quality rehabilitation programme that informally provides many services typical of case management”. It is not a specific package of care and it does not refer to a specific intervention
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of attrition/exclusion (no reasons for missing data provided)
Free of selective reporting?	Unclear risk	One listed outcome of interest is not reported completely (Utah Case Management Consumer Assessment Record: no summary report, no SD)
Free of other bias?	Low risk	Funded by public institution (NIMH). No details. No evidence other bias are occurring

Marshall-UK

Methods	Allocation: randomised. Design: single centre. Duration: 14 months. Country: Oxford, UK.
Participants	Diagnosis*: severe persistent psychiatric disorder. N=80. Setting: Oxford Social Service. Age: 20 to over 60 years, mean ~ 48 years. Sex: 85% M (68M, 12F). Ethnicity: not reported. History**: i. either homeless, at risk of homelessness, living in supported, temporary or poor-quality accommodation, experiencing social isolation or causing disturbances, ii. not already receiving case management, iii. informed consent given
Interventions	ICM: Case Management from team of social services case managers (case-managers are free to choose how much time to offer each subject, but provided some intervention as minimum). Caseload: 1:10. N-40. Standard care: provided by CMHTs. N-40.
Outcomes	Service use: average number of days in hospital per month, admitted to hospital. Death: all causes. Global state: leaving the study early. Social functioning: REHAB scale, imprisonment, employment. Quality of life: Quality of Life Index. Costs: costs of all care per week (including accomodation). Unable to use - Need for care: MRC Needs for Care Schedule (version modified by authors). Behaviour: Social Integration Questionnaire (not published in peer-reviewed). Social functioning: accomodation status (not stated clearly the definition of “better” and “worse” accommodation status), employment (no mean and SD reported). Costs: direct cost of psychiatric hospital care and of health care (no SD)
Notes	* Schizophrenia and related disorder 74%. ** 40 % illness > 1 yr = 40%, 85% previous psychiatric admission
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomisation by permuted block. No further details.
Allocation concealment?	Unclear risk	Randomisation by sealed envelopes (not stated if opaque)
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: those interviewer rated - rating - Unclear. No details
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of attrition/exclusion (no reasons for missing data provided). Lost to follow up reported on 7 months, not on 14 months. Reason for attrition reported only for the experimental sample
Free of selective reporting?	Low risk	Listed outcomes reported completely.
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

McDonel-Indiana

Methods	Allocation: randomised. Design: multi-centre (5 rural sites). Duration: 24 months. Country: Indiana, USA.
Participants	Diagnosis*: severe mental illness (DSM-III-R coded between 295 and 301.99). N=200 (40 participants for each site). Setting: 5 rural CMHCs. Age: >18 years, mean ~ 38.1 years (SD11.1). Sex: 43% M (86M, 114F). Ethnicity: Caucasian 98%. History**: i. poor utilisation of CMHS and frequent use of psychiatric hospital or emergency room, ii. difficulties with the legal system or in maintaining stable housing, iii. more than one episode of intensive psychiatric care lasting > 2 months, iv. impaired role functioning on a continuing or intermittent basis for at least 2 years
Interventions	ICM: Assertive Community Treatment (one site had addition of Rhinelander model to ACT). Caseload: 1:10. N=100. non-ICM: provided by the mental health services: office-based; subscribed to the tradition of individual Case Management (including day treatment, partial hospitalisation, out-patient therapy, residential services). Caseload: 1:30 to 1:60. N=100
Outcomes	Service use: average number of days in hospital per month (provided for two group of centres, see below in McDonel-Indiana A and McDonel-Indiana B). Global state: leaving the study early. Unable to use - Service use: admission (not reported sample size). Global functioning: Global Assessment of Functioning (GAF) (rated by the therapist), compliance with medication on 11-item client rated checklist (not peer-reviewed scale) Social functioning: Indiana Level of Functioning (Indiana LOF) (rated by the therapist) , accomodation quality and employment scale (not peer-reviewed scale). Social functioning: days in jail, number police contacts (not reported sample size). Mental state: Brief Psychiatric Rating Scale (BPRS) (rated by the therapist). Quality of life: scale modified by the trialist (not peer-reviewed scale). Satisfaction with services: satisfaction with service scale (not peer-reviewed scale)
Notes	*Schizophrenia 48%, affective disorder 32%. N=153. **Mean life-time hospitalisation: 8.4 (SD 7.5), mean hospitalisation in the previous years: 1.3 (SD 1.1).N-153
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcome: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of attrition/exclusion (number of missing data is reported, but no reasons is provided). Not clearly reported number of randomised to each group (some participants refused to participate and their number was not clearly accounted in each group)
Free of selective reporting?	High risk	Some listed outcomes of interest are reported incompletely (sample size not reported for social functioning and admission)
Free of other bias?	Low risk	Public funded (NIMH grant). No details. No evidence other bias are occurring

McDonel-Indiana (A)

Methods	McDonel-Indiana (A) refers to 4 centres grouped together, providing data on service use Methods: see above McDonel-Indiana.
Participants	Participants: see above McDonel-Indiana. N=160.
Interventions	Interventions: see above McDonel-Indiana. ICM: N-80. non-ICM: N-80.
Outcomes	Service use: average number of days in hospital per month.
Notes

McDonel-Indiana (B)

Methods	McDonel-Indiana (B) refers to the 5th centre alone, providing data on service use. Methods: see above McDonel-Indiana.
Participants	Participants: see above McDonel-Indiana. N=40.
Interventions	Interventions: see above McDonel-Indiana. ICM: N=20. Control group: N=20.
Outcomes	Service use: average number of days in hospital per month.
Notes

Morse-Missouri1

Methods	Allocation: randomised. Design: single site. Duration: 12 months. Country: Missouri, USA.
Participants	Diagnosis*: psychiatric diagnosis (DSM-III-R). N=178** Setting: community (drop-in daytime centres for homeless in St Louis area, mental health clinic operated by Missouri Department of Mental Health). Age: mean ~ 33.7 years. Sex: 58% M (103M, 75F). Ethnicity: 52.5% non-white, mostly African-American. History***: i. serious psychiatric disorder defined as a) previous psychiatric hospitalisation or b) above 90th centile on Global Severity Index or c) above 90th centile on psychoticism, paranoid ideation, or depression subscales of the Brief Symptoms Inventory Severity Index, ii. currently homeless, iii. plans to stay in the study area for its duration, iv. no serious violent behaviour
Interventions	ICM: clinical case management based on ACT principles (TCL model). Caseload: 1: 10. N=52 Standard care: traditional outpatient treatment provided at local mental health clinic (offered psychotherapy, medication and assistance in obtaining social services). N=64 Drop in centres****: Two daytime centres made available, one centre exclusively for women. Centre offered respite when other emergency shelters closed. Provided food, clothing, showers, some recreational activity. Social workers available for referrals to other social services, staff-client ratio 1:40. N=62
Outcomes	Global state: leaving the study early. Unable to use - Social functioning: Personality and Social Network Adjustment Scale (N for treatment groups not presented). Social functioning: mean number of days spent homeless in past month (N for treatment groups not presented), monthly quantity and frequency of alcohol consumption based on form developed by National Institute on Alcohol Abuse and Alcoholism (N for treatment groups not presented). Mental state: global severity index of Brief Symptom Inventory (N for treatment groups not presented). Participant satisfaction: measuring instrument not reported (N for treatment groups not presented). Self-esteem: Rosenberg’s Self-esteem Scale (not peer-reviewed, N for treatment groups not presented). Costs:mean monthly income (not described as a review outcome of interest, N for treatment groups not presented)
Notes	* 30.1% schizophrenia, 21% major depression, 8% bipolar, 5% other psychotic disorder, 12% alcohol abuse, 4% other drug abuse, 15% other axis I disorder, 5% no diagnosis. Of those with major axis I disorder 23% had dual diagnosis with substance abuse. N-155 (as for remaining 23 participants it was not possible to assess diagnosis). ** Initially 50 people assigned to each treatment group but if they were lost to follow-up within first month after entering the study, they were replaced by random assignment. *** Mean length of time since last stable address 16.6 months, 71.9% had previous psychiatric hospitalisation. **** Drop in centre group - data from this arm were not presented as the intervention provided did not fit in the inclusion criteria for any of the interventions addressed in this review
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details. Participants lost to follow-up within first month were replaced, replacement was performed through random allocation
Allocation concealment?	Unclear risk	Not described.
Blinding? All outcomes	Unclear risk	Primary outcome: not available. Secondary outcomes: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Number and reasons for lost to follow up incompletely reported
Free of selective reporting?	High risk	All data presented but not usable due to N for treatment groups missing
Free of other bias?	Low risk	Public funded (grants from National Institute of Mental Health). No further details. No evidence other bias are occurring

Morse-Missouri3

Methods	Allocation: randomised. Design: single site. Duration: 24 months. Country: Missouri, USA.
Participants	Diagnosis*: severe mental illness and substance abuse disorder (according to DSM-IV). N=196. Age: 18- 66 years, mean - 40 years. Sex: 80% M (119M, 30F). N=149. Ethnic:73% African American, 25% Caucasian, 2% other minority. History**: currently homeless and severely mentally ill, not already enrolled in an Intensive Case Management programme
Interventions	ICM: clinical team trained to deliver psychiatric care package following ACT principles and practices. Caseload: 1:10. Treatment of participant’s substance abuse via referral to other community providers for outpatient or individual substance abuse services, 12-step group. N=46 Standard Care: participants shown a list of community agencies that provided mental health and/or substance abuse treatment***. Team also provided linkage assistance to help participants access these services. N=49 Integrated Community Treatment: Combination of Integrated Treatment Services and Assertive Community Treatment (IACT), clinical team trained to deliver psychiatric care package following ACT principles and practices. Also trained to follow Integrated Treatment principles and practices. Substance abuse services provided directly via counselling and bi-weekly treatment groups. Substance abuse specialist part of IACT team. N=54****
Outcomes	Social functioning: number of days homeless per month, mean number of days used substances. Costs: inpatient psychiatric costs, health care costs, costs of all care Unable to use - Service use: contacts with treatment programme - not clearly defined (not described as a review outcome of interest). Global state: leaving the study early (overall loss given, no individual group data). Mental state: 24-item BPRS (data reported are not likely to be obtained through the stated measurement instrument, as rating scores reported are not consistent to the BPRS 24-items scale). Satisfaction: client satisfaction (modified scale used; not peer-reviewed). Social functioning: substance abuse rating score (scale not described or referenced). Costs: emergency shelter costs (not described as a review outcome of interest)
Notes	* 48% schizophrenia, 19% schizoaffective disorder, 11% atypical psychotic disorder, 11% bipolar, 9% major depression-recurrent disorder, 2% delusional disorder. ** Mean 12.5 days homeless in previous month. *** Most agencies specialised in either mental health treatment or substance abuse treatment, but not both. **** Data from this group not used in final analysis, as ACT plus another treatment
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised, but no description given.
Allocation concealment?	Unclear risk	Not described.
Blinding? All outcomes	Unclear risk	Primary outcome: not available. Secondary outcomes: interviewer rated -rating - Unclear. Not described
Incomplete outcome data addressed? All outcomes	Unclear risk	Attrition described for the total sample and analysed for effect on outcome. Number and reason for loss to follow up not reported for single arm
Free of selective reporting?	Low risk	All outcomes reported.
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Muijen-UK2

Methods	Allocation: randomised. design: single centre. Duration: 18 months. Country: Greenwich Health District, London, UK.
Participants	Diagnosis*: severe mental illness: psychotic disorder (schizophrenia or affective psychosis) (diagnostic criteria not reported). N=82. Setting: Health District. Age: 18-64 years, mean - 37 years. Sex: 56% M (46M, 36F). Race: 23.1% African/Afro-Caribbean. History**: schizophrenia or affective psychosis lasting >2 years, ≥2 hospital admissions last 2 years, about to be discharged, no primary organic disorder
Interventions	ICM: case management approach provided by a Community Support Team (community psychiatric nurses and team leader). The team is acting as advocate, practical assistance with welfare benefits and housing, no discharge policy. Caseload: 1:8-11 for the first 15 months (until April 1990), then increased up to caseload: 1:20-25 for the last 3 months, until the end of the trial. N=41. Standard Care: provided by Community Psychiatric Nurse (CPNs), working independently and based in primary care. N=41
Outcomes	Service use: average number of days in hospital per month. Death - all causes and suicide. Global state: leaving the study early. Mental state: BPRS 24-items, PSE. Social functioning scale: GAS, SAS, imprisoned. Participant satisfaction (by short term). Costs: direct costs of all care. Unable to use - Social functioning: accomodation (authors reported data on “patients using hostel accommodations”, it is not clear what is included in this definition). Participant satisfaction (by medium and long term) Client Satisfaction Questionnaire (CSQ) (attrition > 50%). Carer satisfaction: attrition >50%. Costs: other costs (no SD).
Notes	*Schizophrenia-like disorder 83%; mania 12%; psychotic depression: 0.5%. **Baseline mean number of admissions 5.7.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: those interviewer rated - rating - Unclear. No details
Incomplete outcome data addressed? All outcomes	Unclear risk	Number of lost to follow up is stated, but reason for attrition is reported generically, referred to the entire sample size and not the single intervention sample
Free of selective reporting?	High risk	All listed outcomes of interests are fully reported (but some economic outcomes missing any variance measurement)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Muller-Clemm-Canada

Methods	Allocation: randomised. Design: multi-centre (2 sites: Site A: New West, Site B: Surrey - individual centre data not reported). Duration: 24 months. Country: British Columbia, Canada.
Participants	Diagnosis*: majority of participants affected by schizophrenic disorder, others any DSM III Axis I or Axis II (including dual diagnosis). N=123. Setting: CMHC. Ethnicity: data not available. Age: 19 - 64 years, (randomised sample age not reported). Sex: 49.5% M (61M, 62F). History: i. serious and persistent mental illness with impaired role functioning, ii. about to be discharged from hospital, iii. high risk of re-hospitalisation, iv. no primary diagnosis of substance abuse, organic brain disease or developmental disorder, v. no recent history of severe violence, vi. informed consent to participate
Interventions	ICM: care from a CMHC plus additional Assertive Case Management according to Stein&Test model. Caseload: 1:10. N-63. Standard care: care from a CMHC. N-60.
Outcomes	Service use: average number of days in hospital per month**, number of admissions**. Death: all causes and suicide. Global state: leaving the study early. Unable to use - Quality of life: scale (not peer-reviewed).
Notes	*Schizophrenia-like disorder (60.2%). ** Variance not reported - data from another study used.
Risk ofbias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcome: those clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Number randomised not clearly reported, as authors declared that “Clients who withdrew from the study within the first 6 months were replaced by other clients”
Free of selective reporting?	High risk	Outcome length of hospitalisation reported incompletely (no SD)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Okpaku-Tennessee

Methods	Allocation: randomised. Design: single centre. Duration*: 4 months. Country: Nashville, Tennessee, USA.
Participants	Diagnosis**: mental illness causing significant impairment. No further information provided. N=152. Setting: urban mental health centres. Age:18 - 55 years, mean - 36.8 years (SD 9.1). Sex: 59% M. Ethnicity: 40% non-white. History: i. clients of mental health centres, ii. serious mental illness as judged by eligibility for disability benefits
Interventions	ICM*** : employment oriented case management provided by psychiatric vocational rehabilitation specialists, supervised by the multidisciplinary team, vocational rehabilitation specialists. Caseload: 1:10. N=73 non-ICM: standard case management services from CMHC. Caseload: 1:40 to 90. N=79
Outcomes	Global state: leaving the study early. Unable to use - Service use: admission data (not reported). Social functioning: employment status (data are reported referring to the full trial length, although intervention was provided only for first 4 months).* Costs: insufficient data reported.
Notes	* Variable follow up period - all received 4 months intervention and one 3 month follow up interview, some participants were followed up as long as 24 months (intervention duration was 4 months). Review authors reported data only from 4 months follow up. **Schizophrenia 23%, mood disorder 21%, 33% no current diagnosis according to SCID diagnostic criteria. *** The psychiatric vocational rehabilitation specialist intervention was provided for 4 months. The project specialist engaged the clients in therapeutic and rehabilitation activities, they assisted mental health workers to obtain services more expeditiously, the psychiatrist was available 24 hours a day for consultation
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: not available. Secondary outcome: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of attrition/exclusion to permit judgement (number for missing to follow up is reported, but reason is not reported)
Free of selective reporting?	High risk	More listed outcomes of interest are reported incompletely.
Free of other bias?	Low risk	Public funded (grant from the Social Security Administration). No further details. No evidence other bias are occurring

OPUS-Denmark

Methods	Allocation: randomised. Design: multi-centre (5 centres, in Copenhagen and Aarhus, data not available for single centres). Duration: 24 months. Country: Copenhagen, Denmark.
Participants	Diagnosis*: schizophrenia or schizophrenia like psychosis (according to Research Criteria, codes in the F2 category in ICD-10) (main diagnosis and comorbidity based on SCAN 2.0 and SCAN 2.1). N=547. Setting: public mental health service. Age: 18 - 45 years, mean 26.6 years. Sex: 59% M (323M, 224F). Ethnicity: not reported. History**: i. prior treatment of mental disorder has not been adequate (i.e. ≥12 weeks of continuous antipsychotic medication in antipsychotic dosage), ii. absence of learning disability, organic mental disorder and psychotic condition due to psychoactive substance use, iii. danish speaker, iv. written informed consent, v. legal residence in the catchment area, vi. the use of psychoactive drug did not cause exclusion
Interventions	ICM***: modified Assertive Community Treatment (including individual case manager, recommendation of antipsychotic medications, psychoeducational family treatment**** and social skill training). Caseload: 1:15. N-275. Standard care***: treatment in a CMHC. Caseload: 1:20-30. N-272
Outcomes	Service use: average number of days in hospital per month, not in contact with service, defined as “an unplanned break of at least 30 days in treatment or between treatment regimens or status (i.e. from discharge to outpatient status)”, admitted to hospital (during previous 12 months). Death: all causes, suicide. Global state: leaving the study early. Social functioning: not working or in education, not living independently, alcohol and drug abuse diagnosed with SCAN. Mental state: Scale for Assessment of Positive Symptoms (SAPS), Scale for Assessment of Negative Symptoms (SANS), comorbidity with depression. Participant satisfaction: Client Satisfaction Questionnaire (CSQ). Behaviour: specific behaviour, self-harm. Unable to use - Global state: compliance with medications: defined “good compliance” as having taken the prescribed antipsychotic medications in the recommended doses regularly during previous 3 months (not reported), GAF (reported only sub-scale data, not global score) Social functioning: Social Network Size (not published measure instrument), Social Network Schedule (SNS). Quality of life: not reported. Participant satisfaction: Camberwell Assessment of Need (CAN) (not reported). Relative satisfaction: Client Satisfaction Questionnaire 8-items, adapted version (not reported). Behaviour: Social Behaviour Assessment Schedule (SBAS) (reported only score on sub-scale). Use of coercive measure: not listed as a review outcome of interest. Assessment of Expressed emotion: not listed as a review outcome of interest
Notes	*Schizophrenia: 66%; schizotypal disorder ~ 15%; schizoaffective disorder ~ 5%. **Median duration of untreated psychosis ~ 50 weeks. ***In both interventions use of antipsychotics followed the guidelines from the Danish Psychiatry Society (recommending a low-dose strategy for patients with first-episode psychosis and the use of second generation drugs as first choice). ****Family treatment followed McFarlane’s manual for psychoeducational treatment for multiple family group (18 months treatment, 1.5 hours twice a week, in a multiple family group with two therapists and 4 to 6 patients with their families). Focus on problem solving and development of skills to cope with illness
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised. In Copenhagen, allocation sequence was computer generated. In Aahrus a secretary drew a lot among five red and five white lots from a black box. Central randomisation. Randomisation was 1: 1, in block of 6, stratified for each of the 5 centres
Allocation concealment?	Low risk	In Copenhagen, randomisation was carried out through centralised telephone allocation. In Aahrus, researchers were informed on the randomisation assignment after they had finished the entry assessment. Block sizes were unknown to investigator
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcome: those interviewer rated - rating - No. Reported: “Investigators were not blind to treatment allocation”
Incomplete outcome data addressed? All outcomes	Low risk	Number and reason for missing data are clearly reported. Analysis were conducted on an ITT basis
Free of selective reporting?	High risk	Authors reported the change of primary outcome stated in the protocol (from “relapse and positive symptoms” to “psychotic and negative symptoms”) as high attrition occurred in the former outcome measurements. Some listed outcome of interest are not reported
Free of other bias?	Low risk	Public funded. None declaration of interest. no further details. No evidence other bias are occurring

Pique-California

Methods	Allocation: randomised. Design: single centre. Duration: 24 months. Country: San Francisco, USA.
Participants	Diagnosis: severe and persistent mental illness (psychiatric disorder as the primary source of disability). N=37. Setting: Department of Psychiatry, San Francisco General Hospital. Age: =18 years. Sex: not reported. Ethnicity: European or African American. History: high user of intensive treatment care who had experienced an unsatisfactory quality of life in the community, i. recently hospitalised; ii. no primary diagnosis of (a) organic brain disease, (b) substance abuse disorder with no other psychiatric disorder, or (c) learning disability, iii. no history of violence not due to treatable psychiatric symptoms, iv. written informed consent
Interventions	ICM: standard care + Assertive Community Treatment according to Stein&Test model, culturally-focused on needs of Afro-American population. Caseload: 1:10. N= 22. Standard care: Caseload: 1:30. N=15.
Outcomes	Global state: leaving the study early. Unable to use - Service use: average number of days in hospital per month (not reported). Global functioning: GAF (attrition 57%). Costs: reported incompletely (no mean, no SD).
Notes
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised. Stratified by sex, ethnicity, recruitment centre, assignment to intervention 2:1. Randomisation sequence constructed by independent investigator using a table of random permutation
Allocation concealment?	Unclear risk	Randomisation sequence sealed in opaque envelopes, unused assignment envelopes kept in a locked container (available to the recruiter psychiatrists)
Blinding? All outcomes	Unclear risk	Primary outcome: not available. Secondary outcome: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Author did not address this outcome. Reason for missing outcome data is not reported
Free of selective reporting?	High risk	Listed outcome of interest reported incompletely (costs: no mean, no SD)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Quinlivan-California

Methods	Allocation: randomised. design: single centre. Duration: 24 months. Country: San Diego, California, USA.
Participants	Diagnosis*: major disorder DSM-III-R axis I. N=90. Setting: San Diego County Mental Health Service. Age: >18 years, 33% > 40 years, mean - 37 years. Sex: 44% M (40M, 50F). Ethnicity: 43% non-white (18% African-American). History: =3 hospitalisations last 2.5 years.
Interventions	ICM: Assertive Community Treatment model Stein & Tests. Caseload: 1:15. N-30. non-ICM: traditional CM programme, no team approach. Caseload: 1:40-60. N-30. Standard Care: services offered by the public mental health system. N-30
Outcomes	Service use: average number of days in hospital per month. Costs: direct costs of psychiatric hospital care. Unable to use - Service use: other service use than hospital (reported incompletely). Global state: leaving the study early (not reported). Costs: direct costs of other psychiatric care (not one of the three cost outcomes examined by this study)
Notes	*56% schizophrenia; bipolar disorder: 23%. This is a three arm study, data from the study are included in both comparison addressed by the review: Intensive Case Management vs Non-Intensive Case Management and Intensive Case Management vs Standard Care
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised, no further details.
Allocation concealment?	Unclear risk	No information.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: - rating - Yes. No information provided, but available outcomes are not likely to be influenced by lack of blinding
Incomplete outcome data addressed? All outcomes	Unclear risk	The study did not address this outcome.
Free of selective reporting?	Low risk	Listed outcomes of interest are reported completely.
Free of other bias?	Low risk	No data provided. No evidence other bias are occurring.

REACT-UK

Methods	Allocation: randomised. Design: multi-centre (2 centres, information for single centre not available). Duration: 18 months. Country: London, UK.
Participants	Diagnosis*: SMI (schizophrenia, schizoaffective disorder, other chronic psychosis, bipolar affective disorder). N=251. Setting: community services in 2 inner London boroughs (Camden and Islington). Age: mean 39 years (SD 11). Sex: 58% M. Ethnicity: African-Caribbean 36%. History: i. living independently or in low supported accomodation, ii. under the care of CMHT = 12 month and having difficulties engaging with standard community care, iii. recent high use of inpatient care (i.e. = 100 consecutive inpatient days or = 5 admissions previous 2 years, or = 50 consecutive inpatient days or = 3 admissions previous 1 years) , iv. substance misuse or personality disorder eligible if these were secondary diagnosis, v. no organic brain damage
Interventions	ICM: Assertive Community Treatment (as described by McGrew 1995). Caseload: 1: 12. N=127. non-ICM: services offered by CMHT (according to Care Programme Approach). Caseload: 1:35. N=124
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services, admitted to hospital. Death: all causes and suicide. Global state: leaving the study early, Health of the Nation Outcome Scale (HoNOS), Rating of Medication Influence scale (ROMI). Social functioning: arrested, imprisoned, number homeless; Life Skill Profile (LSP), substance abuse: assessed through various scales (Alcohol Use Scale - AUS, Drug Use Scale - DUS, Substance Abuse Treatment Scale - SATS), but reported as binary outcome. Mental state: Brief Psychiatric Rating Scale (BPRS-24 items). Behaviour: self-harm, injury to others. Quality of life: Manchester Short Assessment of Quality of Life (MANSA). Participant satisfaction: Client Satisfaction Questionnaire modified version (CSQ-modified), Camberwell Assessment Need abbreviated form (CAN) Unable to use - Use of Mental Health Act (not listed as review outcome of interest). Quality of engagement: adapted form of Homeless Engagement Acceptance Scale (HEAS) (not listed as review outcome of interest)
Notes	*53% schizophrenia, 13% schizoaffective, 4% bipolar; 26% illicit drug misuse; 25% alcohol misuse
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised: permuted block randomisation with a block size of eight ensuring parity between CMHT in proportions randomised to ICM
Allocation concealment?	Low risk	The interviewer contacted an administrator at the trial centre who opened the appropriate numbered envelope communicating the outcome of randomisation. Participants and referrers were informed of the treatment assignment by letter
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated -rating - No. Interviewers were independent of clinical care, but not blind
Incomplete outcome data addressed? All outcomes	Low risk	YES - Primary outcomes: average number of days in hospital per month. No missing data (except death, balanced in number across groups) YES - Secondary outcomes: number and reason for missing data are reported. Analysis carried out on an ITT basis
Free of selective reporting?	Low risk	All the listed outcomes of interest are completely reported.
Free of other bias?	Low risk	Public funded (Camden and Islington Health Authority; King’s Fund; Department of Health). Competing interesting declared: none. No further details. No evidence other bias are occurring

Rosenheck-USA

Methods	Allocation: randomised. Design: multi-centre. 10 sites: 6 General Medical and Surgical centres (GMS) and 4 Neuropsychiatric (NP) centres. Data available both for single centre and for two pooled centre groups (GMS and NP). See below. Duration: 24 months. Country: Northeastern United States, U.S.A.
Participants	Diagnosis*: primary psychiatric disorder. N=873. Setting: community-based psychiatric care - Department of Veteran Affairs (VA). Age: 48% > 45 years, mean 47.6 years. Sex: 100% M. Ethnicity: 20 % non-white. History: i. current inpatient in VA psychiatric unit, ii. no primary diagnosis of substance abuse or organic brain disease, iii. recent high user of psychiatric care (definition varied between GMS and NP centres), iv. written consent
Interventions	ICM: Intensive Psychiatric Community Care programme, providing ACT intervention according to Stein&Test model. Caseload: average 1:7-15. N-454. Standard Care**: routine care from psychiatric services provided by the Veteran Affairs, including inpatient and outpatient psychiatric treatment, psychopharmacological treatment and rehabilitation service. N=419
Outcomes	Service use: average number of days in hospital per month (both available pooled data for GMS and NP centres and data for single centre. See below)***. Global state: GAS (authors reported results pooled for GMS and NP centres; see below) Mental state: BPRS-18 items, BSI (authors reported results pooled for GMS and NP centres. See below). Costs: total health care cost (authors reported results pooled for GMS and NP centres. See below) Unable to use - Global state: self-reported measure (not published, not peer-reviewed). Participant satisfaction: satisfaction with service (measurement instrument not published, not peer-reviewed). Social functioning: Addiction Severity Index - ASI (reported only a sub-scale, not the overall score). Costs: non health care costs (not listed as an outcome of interest in the review)
Notes	*Schizophrenia 50.5%, dual diagnosis 28%, bipolar disorder 10%. **Standard Care provided by both type of site did not differ programme wise. *** Pooled data for GMS and NP entered the meta-analysis; data for single centre entered in meta-regression
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised through coin tossing.
Allocation concealment?	Unclear risk	No details (it is just reported that the assignment by coin tossing was performed by an independent researcher)
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated -rating - Unclear. Authors report that interviewers are independent, but it is not stated whether they are blind to participant assignment
Incomplete outcome data addressed? All outcomes	Unclear risk	The study did not address this outcome.
Free of selective reporting?	High risk	Some listed outcomes of interest are not reported completely (i.e. substance abuse: it is reported only as a subscale score and not the general score)
Free of other bias?	Low risk	No details. No evidence other bias are occurring.

Rosenheck-USA-GMS

Methods	Rosenheck-USA-GMS is referring to 6 General medical and Surgical centres (GMS) pooled together (centre A, B, D, F, I, J) Methods: see above Rosenheck-USA.
Participants	Participants: see above Rosenheck-USA. Following data are specific for Rosenheck-USA-GMS: N=528. Setting: Department of Veteran Affairs (VA) - 6 GMS centres enrolled in the Intensive Psychiatric Community Care Programme. GPS centres are located in urban centres, provide shorter-term, crisis-oriented, inpatient care. History: i. current inpatient in VA psychiatric unit, ii. no primary diagnosis of substance abuse or organic brain disease, iii. recent high user of psychiatric care (defined as i. = 40 days in hospital or ii. ≥ 2 admission in the previous yr), iv. written consent
Interventions	ICM*: Intensive Psychiatric Community Care programme, providing ACT intervention according to Stein&Test model. Caseload: average 1:7-15. N=271. Standard Care: routine care from psychiatric services provided by the Veteran Affairs, including inpatient and outpatient psychiatric treatment, psychopharmacological treatment and rehabilitation service. N=257
Outcomes	Service use: average number of days in hospital per month**. Global state: GAS. Mental state: BPRS-18 items, BSI. Costs: costs of health care.
Notes	* Centre J: Caseload 1:44. ** Entered in the meta-analysis.

Rosenheck-USA-GMS (A)

Methods	Rosenheck-USA-GMS (A) is a GMS centre. Methods: see above Rosheneck-USA and Rosenheck-USA-GMS.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-GMS. N=79.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-GMS. ICM: N-44. SC: N=35.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-GMS (B)

Methods	Rosenheck-USA-GMS (B) is a GMS centre. Methods: see above Rosenheck-USA and Rosenheck-USA-GMS.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-GMS. N=94.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-GMS. ICM: N-47. SC: N=47.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-GMS (D)

Methods	Rosenheck-USA-GMS (D) is a GMS centre. Methods: see above Rosenheck-USA and Rosenheck-USA-GMS.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-GMS. N=102.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-GMS. ICM: N=49. SC: N=53.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-GMS (F)

Methods	Rosenheck-USA-GMS (F) is a GMS centre. Methods: see above Rosenheck-USA and Rosenheck-USA-GMS.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-GMS. N=78.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-GMS. ICM: N=43. SC: N=35.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-GMS (I)

Methods	Rosenheck-USA-GMS (I) is a GMS centre. Methods: see above Rosenheck-USA and Rosenheck-USA-GMS.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-GMS. N=88.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-GMS. ICM: N=44. SC: N=44.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-NP

Methods	Rosenheck-USA-NP is referring to 4 Neuropsychiatric centres (NP) pooled together (centre C, E, G, H) Methods: see above Rosenheck-USA.
Participants	Participants: see above Rosenheck-USA. Following data are specific for Rosenheck-USA-NP: N=345. Setting: Department of Veteran Affairs (VA) - 4 NP centres enrolled in the Intensive Psychiatric Community Care Programme. NP centres are located in suburban or rural settings, are large facilities providing long term mental health care. History*: i. current inpatient in VA psychiatric unit, ii. no primary diagnosis of substance abuse or organic brain disease, iii. recent high user of psychiatric care (defined as i. >=180 days in hospital or ii. >= 4 admission in the previous yr), iv. written consent
Interventions	ICM: Intensive Psychiatric Community Care programme, providing ACT intervention according to Stein & Test model. Caseload: average 1:7-15. N-183. Standard Care***: routine care from psychiatric services provided by the Veteran Affairs, including inpatient and outpatient psychiatric treatment, psychopharmacological treatment and rehabilitation service. N=162
Outcomes	Service use: average number of days in hospital per month**. Global state: GAS. Mental state: BPRS-18 items, BSI. Costs: direct costs of health care.
Notes	*Characteristic at baseline differs between NP sites regarding to “inpatient days before programme entry”. Difference was attributable to site C. To compensate for this imbalance authors made adjustment. The adjusted value was used in all the calculations, but ‘the results they yielded did not differ substantially from those obtained with unadjusted value’. ** Entered in the meta-analysis.

Rosenheck-USA-NP (C)

Methods	Rosenheck-USA-NP (C) is a NP centre. Methods: see above Rosenheck-USA and Rosenheck-USA-NP.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-NP. N=93.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-NP. ICM: N-50. SC: N=43.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-NP (E)

Methods	Rosenheck-USA-NP (E) is a NP centre. Methods: see above Rosenheck-USA and Rosenheck-USA-NP.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-NP. N=67.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-NP. ICM: N=34. SC: N=33.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-NP (G)

Methods	Rosenheck-USA-NP (J) is a NP centre. Methods: see above Rosenheck-USA and Rosenheck-USA-NP.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-NP. N=71
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-NP. ICM: N=40. SC: N=31.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Rosenheck-USA-NP (H)

Methods	Rosenheck-USA-NP (H) is a NP centre. Methods: see above Rosenheck-USA and Rosenheck-USA-NP.
Participants	Participants: see above Rosenheck-USA and Rosenheck-USA-NP. N=114.
Interventions	Interventions: see above Rosenheck-USA and Rosenheck-USA-NP. ICM: N=59. SC: N=55.
Outcomes	Service use: average number of days in hospital per month*.
Notes	* Entered in the meta-regression.

Salkever-SCarolina

Methods	Allocation: randomised. Design: single centre. Duration: 18 months. Country: Charleston, South Carolina, USA.
Participants	Diagnosis*: schizophrenia, schizoaffective disorder, bipolar disorder or other psychotic disorder (DSM-III-R). N=173. Setting: Charleston and Darchester CMHCs. Age: 18-65 years, mean ≃ 35.5 yrs. Sex: 54.8% M. (N-114). Ethnicity: 62.5% non white. (N-114). History: i. history or high risk for high services use patterns (i.e. long term or multiple hospitalisation), ii. difficulty with treatment compliance, independent living or activities of daily living, iii. no primary diagnosis of personality disorder or substance abuse or organic brain syndrome, iv. no need for 24 hrs supervision, v. no assault behaviour in the previous year not associated with psychosis
Interventions	ICM**: Programme of Assertive Community Treatment (PACT) provided by two different location (treatment based in CMHC and treatment run by non-profit organisation located near CMHC). Caseload: ranging during the trial duration from 1:6.5 to 1:13. N=104. non-ICM: standard care from Community Mental Health Centre, providing primarily office-based case management programme. Caseload: decreased during the trial duration from 1:68 to 1:34. N-69
Outcomes	Service use: average number of days in hospital per month***; admitted to hospital. Global state: leaving the study early. Unable to use - Costs: direct costs of psychiatric hospital care (no SD).
Notes	* N-114, 81.5% schizophrenia or schizoaffective disorder, 50.8% secondary diagnosis of substance abuse disorder. ** PACT provided by the two sites do not differ. *** Variance not reported - data from another study used.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Insufficient reporting of missing data (number for missing data is provided, but reason is provided for the whole sample, it is not stated for single intervention group)
Free of selective reporting?	High risk	Listed outcome of interest are fully reported, but any variance measurement are missing
Free of other bias?	Low risk	Public funded (by NIMH and grant from universities and university centres for research). No further details. No evident other bias are occurring

Shern-USA1

Methods	Allocation: randomised. Design: single-centre. Duration: 24 months. Country: New York, USA.
Participants	Diagnosis*: severe mental illness causing severe disability (definition not clearly reported) N=168. Setting: community services. Age: ≥18 years, mean 40 years, range 21-66 yrs. Sex: 76% M (128M, 40F). Ethnicity: black 61%. History: i. homeless for 7 of last 14 nights, ii. no primary diagnosis of substance abuse or learning disability, iii. not considered dangerous to selfor others, iv. informed consent given
Interventions	ICM: from the Choices Programme which aims to develop the “housing readiness” (i.e. compliance with psychiatric treatment and period of sobriety). The main features are: i. drop-in centre, ii. Case Management intervention modelled according to main principles of the ACT model. Caseload: ~ 1:13. N=91. Standard care**: provided from a variety of agencies homelessness and specialty mental health services (including drop-in centres, outreach services, mental health and health services, soup kitchen, shelters). It could involve non-ICM. N=77
Outcomes	Social functioning: not living in stable accomodation. Mental state: Colorado Symptoms Index. Quality of Life: Quality of life Index (QOLI). Unable to use - Service use: hospital use (data partially reported). Social functioning: police contact (imprisoned, arrested) (data partially reported), change in proportion of time spent in residential setting (incompletely reported). Self-esteem: Rosenberg’s Self-esteem Scale (not peer-reviewed). Coping: Pearlin and Schooler’s Mastery Scale (not listed as a review outcome of interest) Participant satisfaction: unmet needs (not described measurement instrument)
Notes	*~9% no major mental illness. **People randomised to SC were just “provided informations by the research interviewers about local homelessness service programmes”. No attempts were made to arrange the first contact to available local services
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. No further details.
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	High risk	Primary outcome: not provided. Secondary outcomes: interviewer mediated - rating - NO. Not clearly stated, but it is implicitly not blind
Incomplete outcome data addressed? All outcomes	Unclear risk	Authors declared “using alternative techniques for accommodate missing observations”. Main concern regarding the high attrition rate declared by authors, but not clearly reported as presented data were already transformed through statistician techniques accounting for missing observation
Free of selective reporting?	High risk	Some listed outcomes of interest are not usable due to incomplete reporting (service use, social functioning, quality of life outcomes)
Free of other bias?	Low risk	Public funded (NIMH). No further details. No evident other bias are occurring

Solomon-Pennsylvania

Methods	Allocation: randomised. Design: single centre. Duration: 12 months. Country: Philadelphia, Pennsylvania, USA. Not entering meta-regression.
Participants	Diagnosis*: seriously mentally ill (schizophrenia, affective or personality disorder according to DSM III R). N=200. Setting: CMHC. Age: mean 35.2 years (SD 9.4). Sex: 86.5% M. Ethnicity: Afro-American 81% History: i. about to be released from prison, ii. homeless (i.e. situational, episodically or chronically without a domicile or if jail detention resulted in the loss of a stable housing situation), iii. state hospitalisation ≥60 days in previous 2 years or a significant amount of outpatient treatment, iv. GAF≤40 or GAF≤60 if age ≤ 35 years, v. written informed consent
Interventions	**. ICM: Assertive Community Treatment according to the Stein&Test model. Caseload: 1: 8-12. N=60. **. ICM: Intensive Case Management provided from an individual forensic case manager who worked at CMHC, but as individual rather than as part of a treatment team. Caseload: not available. N=60. . Standard care: from local CMHC (2 clients received ICM services sometimes during the year in the study). N=80
Outcomes	Global state: leaving the study early. Social functioning: imprisonment***. Unable to use - Service use: days in hospital, hospitalisation (data not available), drug and alcohol use Addiction Severity Index (data not available), accomodation, employment, arrest (data not available), Pattison Social Network Interview (data not available). Mental state: BPRS (data not available). Quality of Life: Lehman Quality of Life Interview (objective and subjective components)
Notes	*schizophrenia 82.5%, major affective disorder 10%, other (unspecified psychotic disorder and personality disorder) 3%, substance abuse disorder 52%. N=200. **We considered both arm as a single intervention. *** Attrition in the control group > 50%.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised. Authors reported that “slightly higher number of clients were assigned to the control treatment”, but it is not clear if it was a stratified randomisation. No further details provided
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: not available. Secondary outcome: clinician/participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Incomplete reporting of missing data (number of missing data is reported, but reason is not provided)
Free of selective reporting?	High risk	Some listed outcomes of interest are not reported (i.e. service use, quality of life, etc. )
Free of other bias?	Low risk	No evidence other bias are occurring.

Sytema-Netherlands

Methods	Allocation: randomised. Design: single centre. Duration*: follow-up variable (minimum 12 months, maximum 24 months). Country: Netherlands.
Participants	Diagnosis**: severe mental illness (diagnostic criteria not reported). N=118***. Setting: rural catchment area, community mental health service. Age: mean 41.5 yrs (SD 11 years). Sex: 68.6% M. Ethnicity: not reported. History: Health of the Nation Outcome Scale (HoNOS) ≥ 15.
Interventions	ICM: ACT model according to a fidelity scale (DACT scale). Caseload: 1:10. N=59. Standard care: provided by CMHT. Caseload:1:40. N=59.
Outcomes	Service use: average number of days in hospital per months, not remaining in contact with psychiatric services (defined by authors as “not having any registered contact with the mental health services during the last 12 months of observation”), average number of admission per month. Death: all causes, suicide. Social functioning: homeless at the end of the trial, average days per month in sheltered houses, Dartmouth Assessment of Lifestyle Interview (DALI). Mental state: BPRS-24 items. Participant satisfaction: Client Satisfaction Questionnaire (CSQ), Camberwell Assessment of needs - short version (CANSAS) Unable to use - Social functioning: Social Functioning Scale (SFS) (data not clearly reported)
Notes	*Participant inclusion in the study started April 2204 and was closed at 1 June 2005. All participants were followed up until August 2006. **Schizophrenia 51.7%, major depression 13.5%, bipolar 3%. ***Randomised 119, but one excluded from sample because moved to another area directly after randomisation
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Low risk	Randomised, based on block design (block size of 5).
Allocation concealment?	Low risk	Participant included in the study received id-number from service administration. The new id-number was emailed to a researcher who had a pre-arranged list of id-numbers randomised
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated -rating - No. Open-label
Incomplete outcome data addressed? All outcomes	Low risk	YES - Primary outcomes: service use data. No missing data (but one participant, excluded from sample because moved to another area directly after randomisation) UNCLEAR - Secondary outcomes: number of missing data is provided, but reason is not stated
Free of selective reporting?	Low risk	All listed outcomes of interest are completely reported.
Free of other bias?	Low risk	Public funded (grant from the Netherlands Organization for Health Research and Development). No further details. No evidence other bias are occurring

Test-Wisconsin

Methods	Allocation: randomised. Design: single centre. Duration: total duration of the trial 12 years, participants were followed at least 5 years extended to 12 years for those who were first entered in the study. Available some data at 4 years, some data at 24 months. Country: Dane County, Wisconsin, USA.
Participants	Diagnosis*: schizophrenia or schizoaffective disorder (according to RDC), or schizotypal personality (according to DSM III). N=122. Setting: CMHS. Age: 18-30 years. Sex: 67.2% M (82M, 40F). Ethnicity: white 95.9%. History**: i. resident in Dane County, ii. < 12 months total prior time spent in psychiatric and penal institutions, iii. no primary diagnosis of mental retardation, organic brain syndrome or alcoholism, iv. informed consent
Interventions	ICM: Assertive Community Treatment according to Stein&Test model. Caseload: ~ 1:9. N=75. Standard care: Routine care from Dane County psychiatric services - included an unspecified degree of case management. N=47
Outcomes	Service use: average number of days in hospital per month, not remaining in contact with psychiatric services, admitted to hospital. Death: all causes, suicide. Global state: leaving the study early. Social functioning: number homeless or living in sheltered accomodation (at least 1 day) , not living independently, imprisoned Unable to use - Mental state: BPRS, BSI (no data). Social functioning - days homeless (no SD), days in jail (no SD), Community Adjustment Form (scale not peer reviewed, no data). Quality of life: Satisfaction with Life Scale (scale not peer reviewed, no data)
Notes	*Schizophrenia: 73.8%; schizoaffective disorder ~ 23%. **Average age first contact with mental health system: 19.02 yrs
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised (randomised assignment to experimental and control intervention on a ratio 6:4), no further details
Allocation concealment?	Unclear risk	No details.
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: most are clinician/ participant mediated - rating - Unclear
Incomplete outcome data addressed? All outcomes	Unclear risk	Number and reason for missing data are not clearly reported.
Free of selective reporting?	High risk	Listed outcomes of interest not reported or reported incompletely, BPRS and BSI (not reported), days in jail and days in hospital (no SD)
Free of other bias?	Low risk	Public funded (grant by NIMH). No further details. No evidence other bias are occurring

UK700-UK

Methods	Allocation: randomised. Design: multi-centre (4 sites, 3 in London and 1 in Manchester). Duration: 24 months. Country: UK.
Participants	Diagnosis*: psychotic illness (diagnosed through OPCRIT). N=708. Setting: inner-city mental health services. Age: 18-65 years, mean - 37.3 years. Sex: 57% M (404M, 304F). Ethnicity: Afro-Caribbean 28%. History: i. psychotic illness for at least 2 years, ii. ≥ 2 previous hospital admissions, one within past 2 years, iii. no organic brain disease or primary diagnosis of substance misuse, iv. written informed consent
Interventions	Intensive Case Management**: Caseload: 1:10-15. N=353. non-ICM: Caseload: 1:30-35. N=355.
Outcomes	Service use: average number of days in hospital per month (available for single centre, see below), not remaining in contact with psychiatric services (report as not remaining in contact with case-manager), admitted to hospital, mean number of admissions. Death: all causes and suicide. Global state: leaving the study early. Social functioning: time in independent living accomodation (reported as time spent in stable accomodation), imprisoned, Camberwell Assessment of Need (CAN). Mental state: general symptoms CPRS, negative symptoms (SANS). Behaviour: self-harm, harm to others. Quality of life: Lancashire Quality of Life (LQoLP). Participant satisfaction: Patients’ satisfaction with health services questionnaire.. Costs: direct costs of all care. Unable to use - Social functioning: DAS (scale adapted by the trialist, not peer-reviewed). Mental state: sub-scale of CPRS for depression (MADRS) and for psychotic symptoms (no SD), for anxiety and for behavioural disturbance (no SD, not peer-reviewed scale). Carer satisfaction: Experience of Care giving Inventory, 12-items General Health Questionnaire (attrition 83.6%). Costs: direct costs of health care and of psychiatric hospital care (no SD). Adverse drug effects: Abnormal Involuntary Movement Scale (not listed as outcome of interest of the review)
Notes	*Schizophrenia 38%, schizoaffective 49%, mania or bipolar 5%. **The team organisation slightly varied across centres.
Risk of bias
Bias	Authors’ judgement	Support for judgement
Adequate sequence generation?	Unclear risk	Randomised, randomisation stratified by centre. No further details
Allocation concealment?	Low risk	Randomised allocation assigned by telephone or fax by an independent statistical centre
Blinding? All outcomes	Unclear risk	Primary outcome: clinician/participant mediated - rating - Unclear. Secondary outcomes: interviewer rated -rating - No. (Researchers were not masked to treatment allocation)
Incomplete outcome data addressed? All outcomes	Low risk	Number and reason for attrition stated. Analysis performed on an ITT basis
Free of selective reporting?	Unclear risk	Most of the listed outcomes are reported completely (the only exception is for CPRS sub-scale: SD missing)
Free of other bias?	Low risk	Public funded (grants from the UK Department of Health and NHS Research and Development programme). No further details. No evidence other bias are occurring

UK700-UK (A)

Methods	Centre: St. Georges - London. Methods: see above UK700-UK.
Participants	Participants: see above UK700-UK. N=196.
Interventions	Methods: see above UK700-UK. Sample size providing data: ICM: N-96. non-ICM: N=99.
Outcomes	Service use: average number of days in hospital per month.
Notes

UK700-UK (B)

Methods	Centre: King’s - London. Methods: see above UK700-UK.
Participants	Participants: see above UK700-UK. N=153.
Interventions	Methods: see above UK700-UK. Sample size providing data: ICM: N=77. non-ICM: N=74.
Outcomes	Service use: average number of days in hospital per month.
Notes

UK700-UK (C)

Methods	Centre: Manchester. Methods: see above UK700-UK.
Participants	Participants: see above UK700-UK. N=158.
Interventions	Methods: see above UK700-UK. ICM: N=79. non-ICM: N=79.
Outcomes	Service use: average number of days in hospital per month.
Notes

UK700-UK (D)

Methods	Centre: St. Mary * - London. Methods: see above UK700-UK.
Participants	Participants: see above UK700-UK. N=201.
Interventions	Methods: see above UK700-UK. Sample size providing data: ICM: N=99. non-ICM: N=101.
Outcomes	Service use: average number of days in hospital per month.
Notes	*In St. Mary centre staff were randomly assigned to ICM or non-ICM position

ACT - Assertive Community Treatment

CMHC - Community Mental Health Centre

CMHT - Community Mental Health Team

CPA - Care Programme Approach: the CPA is a combination of non-Intensive Case Management and care from a CMHT, introduced in England in the mid-1990s and becoming standard care thereafter

CPNS - Community Psychiatric Nursing Service

DSM-III-R- Diagnostic Statistical Manual, 3rd Edition, revised (APA 1987)

DSM IV - Diagnostic Statistical Manual, 4th Edition (APA 1994)

F - female

ICD-10 - International Statistical Classification of Diseases and Related Health Problems (WHO 1992)

ICM - Intensive Case Management

ITT - Intention to Treat

M - male

N - number

non-ICM - Non-Intensive Case Management

OPCRIT - Operational Criteria (McGuffin 1991)

OPD - outpatient department

PACT - Programme of Assertive Community Treatment

P/T - part-time

RDC - Research Diagnostic Criteria (Spitzer 1978)

SADS - Schedule for Affective Disorders and Schizophrenia interview (Endicott 1978)

SCAN 2.0 - 2.1: Schedule for Clinical Assessment in Neuropsychiatry (SCAN 2.0 in 1998, SCAN 2.1 since 1999) (WHO 1998)

SCID - Structured Clinical Interview for DSM-IV (First 1997)

SD - standard deviation

SMI - severe mental illness

TCL - Training in Community Living

VA - Veterans administration

Yrs - years

Scales

ACL - Adjective Checklists

ASI - Addiction Severity Index

AUS - Alcohol Use Scale

BDI - Beck Depression Inventory

BPRS - Brief Psychiatric Rating Scale

BSI - Brief Symptom Inventory

CAN - Camberwell Assessment of Need Interview

CPRS - Comprehensive Psychopathological Rating scale

C-DIS-R - Computer based Diagnosis Interview Schedule Revised

CSI - Colorado Symptom Inventory

CSQ - Client Satisfaction Questionnaire

CDRI - Client Services Receipt Inventory

DAS - Disability Assessment Scale

DUS - Drug Use Scale

GAF - Global Assessment of Functioning Scale

GAS - Global Assessment Scale

GSI - Global Severity Index

HAD-S - Hospital Anxiety Depression Scale

HEAS - Homeless Engagement Acceptance Scale

HoNOS - Health of the National Outcome Scale

IMPS - Inpatient Multidimensional Psychiatric Scale

Indiana LOF - Indiana Level of Functioning

ISSI: Interview Schedule for Social Interaction

KS - Krawiecka Scale

LQoLP - Lancashire Quality of Life Profile

LSP - Life Skill Profile

MANSA - Manchester Short Assessment of Quality of Life

PANSS - Positive and Negative Syndrome Scale

PSE - Present State Examination

PSNAS - Personality and Social Network Adjustment Scale

QOLI - Lehman’s Quality of Life Interview

REHAB - a scale of social functioning (REHAB GB: general behaviour); (REHAB DB: deviant behaviour)

ROMI - Rating of Medication Influences

RSES - Rosenberg Self Esteem Scale

SAI - Scale for the Assessment of Insight

SAI-E - Scale for the Assessment of Insight-expanded

SANS - Schedule for the Assessment of Negative Symptoms

SAS - Social Adjustment Scale

SATS - Substance Abuse Treatment Scale

SBS - Social Behaviour Scale

SCL-90: Hopkins Symptoms Check List-90

SCRS - Short Clinical Rating Scale

SFQ- Social Functioning Questionnaire

SLS - Satisfaction with Life Scale

TLFB - Time Line Follow Back

WQBQ - Well-being Questionnaire

Other

NIMH: National institute of Mental Health (USA).

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Bao-China	Allocation: randomised. Participants: people with schizophrenia. Interventions: Family intervention, community-based. Not ICM. Standard care.
Bigelow-Oregon	Allocation: not randomised, quasi-experimental design.
Bond-Chicago2	Allocation: not randomised, matched groups design. Interventions: two types of crisis housing.
Bond-Indiana2	Allocation: not randomised, allocation to ACT and reference group was not random in one of the three participating centres. The study could be included if separate data can be obtained from the two centres where randomisation took place
Borland-Washington	Allocation: not randomised.
Burns-UK	Allocation: randomised. Intervention: multi-disciplinary team home treatment, not ICM
Champney-Ohio	Allocation: randomised. Intervention: all 4 comparison groups received some form of case management, no ACT. Not Standard care or non-ICM in the comparison group
Chandler-California2	Allocation: randomised. Participants: current serious mental illness and substance use disorder. Intervention: In-custody standard care + brief aftercare + Integrated Dual Disorders Treatment. Post custody treatment is including Intensive Case Management and specific intervention addressing substance abuse disorder. Not exclusively community setting; not only ICM Service as usual: In custody standard care + usual post-custody services + 60 days of post release case management and housing assistance. Not exclusively community setting
Chen-China	Allocation: randomised. Participants: adults with schizophrenia (non-acute). Intervention: 1. Antipsychotic + Digital Network + Community Support Network . 2. Antipsychotic + Community Support Network. Digital network involves providing online consultation to patients, sending them email periodically, provide online health education; Community Support Network involves providing physical, psychological and rehabilitation treatment to non-acute patients. Not ICM
COAST-UK	Allocation: randomised. Participants: those with any functional psychosis. Interventions: ICM: Assertive Outreach Team model; caseload: 1:12. Including also individual therapy for residual positive symptoms of psychosis (CBT) and family intervention, if appropriate. Not only ICM. non-ICM: usual services available to a local multidisciplinary team (according to Care Programme Approach -CPA); caseload: 1:35. Not including any specialised psychological interventions
COMO-UK	Allocation: cluster randomisation. Participants: case-manager providing care to patients with psychosis and comorbid substance misuse (“dual diagnosis”). Interventions: Training in dual diagnosis intervention (i.e.: i. treatment manual; ii. 5-days training course in assessment and management of dual diagnosis; iii. subsequent monthly supervision). Not ICM. CMHT management as usual.
Cosden-California	Allocation: randomised. Participants: offenders with serious mental illness. Interventions: ICM + Non-adversarial court proceeding. non-ICM + Adversarial court proceeding. Two arms are provided with different intervention between groups in addition to ICM and non-ICM
CRIMSON-UK	Allocation: randomised. Participants: diagnosis of psychotic illness; ≥ 1 admission to a psychiatric in-patient service during previous 2 yrs. Interventions: Joint Crisis Plan: it aims to empower the client and to facilitate early detection and treatment of relapse. It contains client's treatment preferences for any future psychiatric emergency, when the client might be too unwell to express clear views. Not ICM. Standard treatment: as provided by CMHT * Status of trial: ongoing.
De Cangas-Canada	Allocation: randomised. Intervention: hospital based care for those in control group, not community based ICM
Dean-UK1	Allocation: not randomised.
Dean-UK2	Allocation: not randomised, quasi-experimental study.
Dekker-Netherlands	Allocation: randomised, but randomisation is compromised because after initial randomisation control group were combined with control group from another study that experienced problems. Participants: chronic psychiatric patients with a history of several admissions. Interventions: ACT; caseload: 1:30. Not ICM. Standard care
Deng-China	Allocation: quasi-randomised (randomisation according to hospital admission number). Participants: admitted to hospital for a first onset schizophrenia. Interventions: hospital based, not community based.
Dharwadkar-Victoria	Allocation: not randomised, before and after design.
Fenton-Canada	Allocation: randomised. Participants: people with schizophrenia, acutely ill, requiring immediate hospital admission. Intervention: home-care community based treatment versus hospital care. Not ICM in the experimental group; comparison group not community based
Franklin-Texas	Allocation: randomised. Intervention: Assertive Community Treatment caseload over 30. Not ICM
Glick-New York	Allocation: randomised. Intervention: day hospital care vs out-patient group therapy, not ICM
Godley-Illinois	Allocation: randomised. Participants: major psychiatric and substance abuse diagnosis. Intervention: ICM: Specialised case-management. Caseload 1;15. vs Standard care. Outcome: incomplete data reporting. Not providing usable outcomes
Goering-Canada	Allocation: not randomised, used historical controls.
Gold-SCarolina	Allocation: randomised. Participant: adults with severe mental illness. Intervention: ACT + Individual Placement and Support (integrated supported employment programme) . Not only ICM. 2. Traditional programme: providing mental health and brokered case management services in parallel to vocational services. Not only standard care services. Two arms are provided with different interventions between groups in addition to ACT and standard care
Grawe-Norway	Allocation: randomised. Participant: people with recent - onset schizophrenia. Interventions: Integrated Treatment: package of care provided by a multidisciplinary team, providing case management caseload 1: 10; cognitive-behavioural family treatment and home-based crisis management. Not only ICM. Standard Treatment: clinical-based case management caseload 1: 10 and crisis in-patient treatment
Hargreaves-California	Allocation: not randomised.
Havassy-California	Allocation: randomised. Participants: seriously mentally ill adults. Interventions: Intensive Clinical Case Management community-based versus Non-Intensive Case Management hospital-based (providing intensive support during the initial post discharge period, with a maximum of 60 days). Not community setting in the comparison group. Not Standard care or non-ICM in the comparison group
He-China	Allocation: quasi-randomised (randomised allocation according to register number)
Herz-New York	Allocation: randomised. Interventions: brief hospitalisation vs standard hospital care, not ICM
Hornstra-Kansas	Allocation: not randomised, historical controls.
Hoult-Australia	Allocation: randomised. Participants: people with schizophrenia, acutely ill, requiring immediate hospital admission. Interventions: ICM vs acute admission to a psychiatric hospital
Hurlburt-California	Allocation: randomised. Participants: severe mental illness. Interventions: Non-Intensive Case Management: comprehensive case management; caseload 1:22. Not ICM. Non-Intensive Case Management: traditional case management; caseload 1:40. Not ICM. Non-Intensive Case Management: comprehensive case management; caseload 1:22 + high level access to independent housing. Not ICM. Non-Intensive Case Management: traditional case management; caseload 1:40 + high level access to independent housing. Not ICM
Jerrell-California	Allocation: randomised. Randomisation is compromised because it is not reported number of people excluded after randomisation (patients were excluded if they refused to participate after randomisation or if they had not been discharged from hospital within 6 months of entering the study). Participants: severe mental illness. Intervention: ICM: according to the ACT Stein&Test model. Standard Care: provided in the community setting
Jerrell-SCarolina2	Allocation: randomised. Interventions: ACT vs 12-step recovery programme and behavioural skills training + standard care. Not standard care only in the comparison group
Jones-New York	Allocation: randomised. Participants: severe mental illness. Interventions: Time-limited adaptation of Intensive Case Management: designed for transitions from various institutions to the community, the goal is to enhance continuity of care strengthening the individual's long -term ties to services. Not ICM. Standard Care: community-based services.
Knight-California	Allocation: not randomised, quasi-experimental design.
Kuldau-California	Allocation: randomised. Interventions: rapid discharge vs hospital care, not ICM.
Lafave-Canada	Allocation: randomised. Intervention: hospital based care for control group, not community based care. Not Standard care or non-ICM in the comparison group
Langsley-Colorado	Allocation: randomised. Interventions: out patient family crisis management vs hospital admission. Not ICM
Lehman-Maryland2	Allocation: randomised. Participants: dually diagnosed people. Interventions: ICM + experimental group treatment - Being Sober Group (addressing specifically problems of dually diagnosed adults) versus non-ICM. In the experimental group: not only ICM
LEO-UK	Allocation: randomised. Participants: people with non-affective psychosis. Interventions: ICM: Assertive Outreach model; caseload ~ 1:7. Including also: i. CBT based on manualised protocol; ii. family counselling and vocational strategies based on established protocols. Not only ICM. non-ICM: delivered by the sector community mental health teams (according to Care Programme Approach). No CBT, no family counselling and vocational strategies
Lichtenberg-Israel	Allocation: randomised. Participants: people with at least 3 psychiatric admissions during the previous 2 yrs. Diagnosis is not stated. Interventions: Non-Intensive Case Management; caseload 1:30. Not ICM. Standard Care: provided by the mental health care centres
Lin-China	Allocation: not randomised.
Malm-Sweden	Allocation: randomised. Participants: diagnosed according to DSM-IV schizophrenia, schizophreniform, schizoaffective or delusion disorders. Intervention: 1.“Integrated Care” vs 2. “Rational Rehabilitation”. Both of the compared treatment programmes were delivered in a context of clinical case management by CMHTs. They actively incorporated some key features of ACT; caseloads 1:40. The additional component in the “Integrated Care” programme is the continuous social network resource group for each patients. Not ICM
Martin-Delaware	Allocation: unclear if randomised. Interventions: ACT. Original report missing - full copy requested.
Martin-UK	Allocation: randomised. Participants: people with intellectual disability and mental illness
Marx-Wisconsin	Allocation: randomised. Interventions: hospital based care for control group, not community based care. Not Standard care or non-ICM in the comparison group
McFarlane-New York	Allocation: unclear if randomised. Interventions: ACT vs ACT plus family support (FACT). Not Standard care or non-ICM in the comparison group
McGowan-California	Allocation: unclear if randomised, control and treatment groups were “randomly selected” from a population already receiving ACT or standard care
McGrew-Indiana	Allocation: not randomised (it was a before and after design which exam effects of implementing ACT teams in 6 sites in Indiana)
McHugo-Washington DC	Allocation: randomised. Participants: adults with severe mental illness at high risk for homelessness. Intervention: Parallel housing services programme: including ICM - caseload 1:15; implemented by mobile assertive community treatment team and housing by routine community -based landlords. Not only ICM. Integrated housing services programme: including ICM - caseload 1:15; case management and housing services were provided by teams within a single agency and were closely coordinated. Not Standard care or non-ICM in the comparison group
MECCA-Europe	Allocation: cluster randomisation. Participants: people with functional psychosis. Intervention: Treatment as usual + “outcome management”. It is an innovative treatment where a key-worker assesses patient subjective quality of life, treatment satisfaction and wishes for additional/ different support using a questionnaire (MECCA), every two months. It is expected that the results will directly feed into the therapeutic dialogue and be discussed by the patient and key-worker together. Not ICM. Treatment as usual.
Meneghelli-Italy	Allocation: not randomised.
Merson-UK	Allocation: randomised. Interventions: multi disciplinary team home treatment vs emergency assessment at hospital. Not Standard care or non-ICM in the comparison group
Modcrin-Kansas	Allocation: randomised. Participants: chronically mentally ill. Interventions: strengths model of case management versus standard case management. As caseload is not reported in either groups, it is not clear if experimental intervention could be considered ICM and comparison group could be consider standard care or non-ICM. The study could be included if more data can be obtained on caseload
Morse-Missouri2	Allocation: randomised. Participants: homeless people with severe mental illness. Interventions: 1. ICM: ACT, caseload 1:10; vs 2. non-ICM: Brokered Case Management, caseload 1: 85; vs 3. ACT plus community worker support, caseload 1:10. Outcomes: unable to use data, number for treatment groups not presented
Mosher-California	Allocation: not randomised, alternative assignment.
Muijen-UK1	Allocation: randomised. Participants: those with severe mental illnesses, requiring immediate emergency admission. Interventions: ICM vs acute admission to a psychiatric hospital
Mulder-Missouri	Allocation: randomised, but data from randomised and non-randomised patients not reported separately. Participants: people with schizophrenia, acutely ill, presenting for psychiatric hospital admission. Interventions: ICM: modelled according to the PACT model, (as an alternative to current hospitalisation). Standard care: usual hospital admission procedure and at discharge the usual aftercare case-management services
Pai-India	Allocation: not randomised, alternative assignment.
Polak-Colorado	Allocation: randomised. Interventions: admission to small “community-based therapeutic environments” vs standard hospital care. Not ICM
PRiSM-UK	Allocation: non randomised.
Ren-China	Allocation: not clearly stated if randomised. Participants: people with chronic schizophrenia, admitted to rehabilitation hospital. Interventions: rehabilitation hospital based, not community based
Rossler-Germany1	Allocation: not randomised, case control study.
Rossler-Germany2	Allocation: not randomised, case control study.
Rutter-UK	Allocation: randomised. Participants. severe mental illness. Interventions: ICM: case management provided by a case-manager internal to the CMHT; caseload; 1:15. ICM: case management provided by a case-manager outside the CMHT; caseload; 1:15. In the comparison group not standard care or non-ICM
Santiago-Arizona	Allocation: randomised. Participants: those with serious mental illnesses, recently admitted to hospital and ready for discharge. Intervention: Treatment Network Team provided both in community and hospital-based setting. Caseload not stated. Not ICM. Standard care. Unable to use all outcomes.
Shern-USA2	Allocation: randomised. Participants: serious and persistent mental health disorder + homelessness. Interventions: Pathways to Housing: supporting housing programme, providing permanent, independent housing+ ACT (according to Stein&Test model); caseload: 1:10. Not only ICM. Standard Care: provided by social agencies (i.e. out-reach teams, drop-in centres). If participant has no current affiliation to service providers, information were given about where services could be obtained. No active engagement
Shern-USA3	Allocation: randomised. Participants: severe mental illness, co-occurring addiction and homelessness. Interventions: 1. Housing First programme: providing permanent, independent housing without prerequisites for sobriety and treatment + ACT; versus 2. Standard care. Not only ACT
Shern-USA4	Allocation: randomised. Participants: homeless mentally ill population. Interventions: 1. Pathways to Housing: supporting housing programme, providing permanent, independent housing; versus 2. Standard residential treatment. Not ICM
Solomon-Pennsylvania1	Allocation: randomised. Intervention: one type of case management vs another. Not standard care or non-ICM in the comparison group
Stein-Wisconsin	Allocation: randomised. Participants: those with severe mental illnesses, requiring immediate emergency admission. Intervention: ICM (according to the Assertive Community Treatment model) versus acute admission to a psychiatric hospital
Susser-New York	Allocation: randomised. Interventions: “Critical time intervention” - it is a time-limited approach aimed at stabilising the patient's social support network. Not ICM
Tao-China	Allocation: randomised. Participants: people with schizophrenia, admitted to hospital. Intervention: provided pre-discharging. Hospital based, not community based
Teague-New Hampshire	Allocation: not randomised.
Thornicroft-Maryland	Allocation: not randomised.
Toro-New York	Allocation: randomised. Participants: a minority of the participants suffered from severe mental illness, around 80% were simply homeless
Tyrer-UK	Allocation: randomised Participants: psychotic illness. Intervention: 1. ICM. 2. Standard treatment: provided by the social and psychiatric services. The case managers in the treatment group are also case managers for the control group (see Excluded studies section).
Vincent-Ohio	Allocation: not randomised, alternative assignment.
Wood-New Zealand	Allocation: not randomised, case control study.

ACT - Assertive Community Treatment

CBT - Cognitive Behavioural Treatment

CMHT: Community Mental Health Team

CPA - Care Programme Approach: the CPA is a combination of non-Intensive Case Management and care from a CMHT, introduced in England in the mid-1990s and becoming standard care thereafter

DSM - Diagnostic Statistical Manual

ICM - Intensive Case Management

non-ICM - Non Intensive Case Management

PACT - Programme of Assertive Community Treatment

Characteristics of studies awaiting assessment [ordered by study ID]

Agius-Croatia

Methods	Missing paper. Requesting full copy.
Participants
Interventions
Outcomes
Notes

Dick-UK

Methods	Not available as full report, not possible to contact author
Participants
Interventions
Outcomes
Notes

Guo-China

Methods	Allocation: randomised, no further details.
Participants	Participants: people with schizophrenia, living in the community
Interventions	Intervention: community-based. 1. Experimental intervention: provided by team composed by nurses, social worker; caseload 1:10. Providing: i. Expressed Emotion intervention; ii. Psycho-education (how to prevent relapse, how to deal with adverse effects); iii. Daily living activity. 2. Standard care. Awaiting for further information on comparison treatment
Outcomes Notes	Outcomes: awaiting for data extraction.

Huang-China

Methods	Allocation: stratified randomisation.
Participants	Participants: people with schizophrenia.
Interventions	Intervention: 1. Community-based, family intervention (addressed to patients living in the community). No further details. 2. Community-based, control group (addressed to patients living in the community). No further details. 3. Hospital-based (addressed to hospitalised patients)
Outcomes
Notes	Further informations are requested for completing intervention assessment

Johnson-UK

Methods	Written to author for further information. The available report regards to the ongoing trial (should have been completed January 2008)
Participants
Interventions
Outcomes
Notes

Kane-Virginia

Methods	Written to authors re: i. if randomised; ii. caseload.
Participants
Interventions
Outcomes
Notes

Klotz-California

Methods	Written to author for further informations. The available paper is not enough to assess the study
Participants
Interventions
Outcomes
Notes

Li-China

Methods	Missing paper. Requesting full copy.
Participants
Interventions
Outcomes
Notes

Linszen-Netherlands

Methods	Written to authors for more published reports.
Participants
Interventions
Outcomes
Notes

NCT00781079

Methods	Missing paper. Requesting full copy.
Participants
Interventions
Outcomes
Notes

O’Donnell-Australia

Methods	Written to authors asking for caseload.
Participants
Interventions
Outcomes
Notes

Rivera-New York

Methods	Written to authors.
Participants
Interventions	ICM vs peer-enhanced ICM vs clinic-based ICM.
Outcomes
Notes

Sells-Connecticut

Methods	Allocation: randomised. Design:multi-centre (two Connecticut cities, not stated which ones). Duration: 12 months. Country: Connecticut, USA. Not entering meta-regression.
Participants	Diagnosis*: severe mental illness (schizophrenia spectrum disorder, major mood disorder or both). N=137. Setting: public mental health centres, urban site. Age: 20-63 years, mean 41 yrs (SD 9 years). Sex: 61% M. Ethnicity: African American 28.5%. History: i. Treatment disengagement, ii. informed consent provided
Interventions	ICM**: Case-management services from peer providers partnered with ACT teams. Peer case-manager. Cas-load: 1:10-12. N=68. non-ICM***: regular case-management from regular providers. Caseload: 1:20-24. N=69
Outcomes	Unable to use - Service use: 26-item self-reported measure of service use (not peer review). Level of engagement: rated using 1 item of Level of Care Utilisation System (sub-scale not validated). Social functioning: sub-scale from Addiction Severity Index (ASI) (sub-scale not peer reviewed, data not reported). Client-counsellor relationship: modified version of Barrett-Lennard Relationship Inventory (BLRI) (not peer reviewed, modified by authors)
Notes	*61% psychiatric disorder; 63% major mood disorder; 72% substance abuse disorder; 70% co-occurring disorder (psychotic disorder, mood disorder or both plus substance abuse disorder). **All peer staff had publicly disclosed histories of severe mental illness and some of co-occurring drug use disorder. They received broad-based training concerning the provision of case-management service. ***Regular providers worked alongside peer providers on the same treatment teams One additional paper requested.

Tan-China

Methods	Allocation: randomised, no further details.
Participants	Diagnosis: schizophrenia.
Interventions	1. Experimental intervention: provided by team, caseload not reported. Providing: i. antipsychotic drug treatment; ii. family intervention; iii. occupational intervention; iv. rehabilitation intervention. 2. Control intervention: no details provided
Outcomes	Awaiting further information from author about experimental intervention caseload
Notes

Verhaegh-Netherlands

Methods	Written to authors.
Participants
Interventions
Outcomes
Notes

Characteristics of ongoing studies [ordered by study ID]

Walsh-Connecticut

Trial name or title	Specialized Treatment Early in Psychosis (STEP).
Methods	Allocation: randomised. Blinding: open label.
Participants	Diagnosis: schizophrenia spectrum psychosis or affective psychosis (DSM IV, SCID). N=expected 200. Age: 16-45 years. History: ≤ 8 weeks of received antipsychotic treatment lifetime, informed consent, no psychosis believed due to substance use
Interventions	Specialised early treatment: including individual case management, antipsychotic prescription, multifamily group therapy, group cognitive behavioural therapy and cognitive remediation Standard care: usual referral to CMHC.
Outcomes	Primary outcomes: Service use: rehospitalisation (measured every 6 months for 5 yrs). Secondary outcomes: relapse, overall functioning, quality of life, education and employment status, treatment satisfaction, adherence, substance use, adverse effects (including self-harm and violence), medication side effect, economic measures
Starting date	March 2006.
Contact information	barbara.walsh@yale.edu; vinod.srihari@yale.edu
Notes	ClinicalTrials.gov Identifier: NCT00309452 Expected completion: September 2011.

CMHC - Community Mental Health Centre

DSM - Diagnostic Statistical Manual

SCID - Structured Clinical Interview for DSM-IV (First 1997)

DATA AND ANALYSES

Comparison 1. INTENSIVE CASE MANAGEMENT versus STANDARD CARE.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Service use: 1. Average number of days in hospital per month - by about 24 months	24	3595	Mean Difference (IV, Random, 95% CI)	−0.86 [−1.37, −0.34]
1.1 skewed data (sample size ≥ 200)	5	1812	Mean Difference (IV, Random, 95% CI)	−0.46 [−0.95, 0.03]
1.2 skewed data (sample size < 200)	19	1783	Mean Difference (IV, Random, 95% CI)	−1.01 [−1.74, −0.28]
2 Service use: 2. Not remaining in contact with psychiatric services	9	1633	Risk Ratio (M-H, Random, 95% CI)	0.43 [0.30, 0.61]
2.1 by short term	1	95	Risk Ratio (M-H, Random, 95% CI)	0.54 [0.28, 1.05]
2.2 by medium term	3	1063	Risk Ratio (M-H, Random, 95% CI)	0.51 [0.36, 0.71]
2.3 by long term	5	475	Risk Ratio (M-H, Random, 95% CI)	0.27 [0.11, 0.66]
3 Service use: 3a. Admitted to hospital	15		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
3.1 by short term	2	244	Risk Ratio (M-H, Random, 95% CI)	0.61 [0.22, 1.69]
3.2 by medium term	5	1303	Risk Ratio (M-H, Random, 95% CI)	0.85 [0.77, 0.93]
3.3 by long term	11	1516	Risk Ratio (M-H, Random, 95% CI)	0.96 [0.74, 1.23]
3.4 by long term- during previous 12 months	1	547	Risk Ratio (M-H, Random, 95% CI)	0.67 [0.52, 0.86]
4 Service use: 3b. Average number of admissions per month (skewed data)			Other data	No numeric data
4.1 by medium term			Other data	No numeric data
4.2 by long term			Other data	No numeric data
5 Service use: 4a. Admitted to ER - by long term	1	178	Risk Ratio (M-H, Random, 95% CI)	1.13 [0.72, 1.76]
6 Service use: 4b. Average number of admissions to ER (skewed data) - by medium term			Other data	No numeric data
7 Adverse event: 1a. Death - any cause	14		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
7.1 by short term	2	161	Risk Ratio (M-H, Random, 95% CI)	1.04 [0.16, 6.91]
7.2 by medium term	6	901	Risk Ratio (M-H, Random, 95% CI)	0.78 [0.23, 2.62]
7.3 by long term	9	1456	Risk Ratio (M-H, Random, 95% CI)	0.84 [0.48, 1.47]
8 Adverse event: 1b. Death - suicide	12		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
8.1 by short term	2	127	Risk Ratio (M-H, Random, 95% CI)	0.35 [0.04, 3.27]
8.2 by medium term	4	819	Risk Ratio (M-H, Random, 95% CI)	0.98 [0.17, 5.60]
8.3 by long term	9	1456	Risk Ratio (M-H, Random, 95% CI)	0.68 [0.31, 1.51]
9 Global state: 1. Leaving the study early	21		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
9.1 by short term	5	598	Risk Ratio (M-H, Random, 95% CI)	0.79 [0.44, 1.41]
9.2 by medium term	8	1699	Risk Ratio (M-H, Random, 95% CI)	0.60 [0.51, 0.70]
9.3 by long term	13	1798	Risk Ratio (M-H, Random, 95% CI)	0.68 [0.58, 0.79]
10 Global state: 2. Average endpoint score (GAF, high = good)	5		Mean Difference (IV, Random, 95% CI)	Subtotals only
10.1 by short term	4	797	Mean Difference (IV, Random, 95% CI)	2.07 [0.28, 3.86]
10.2 by medium term	3	722	Mean Difference (IV, Random, 95% CI)	0.09 [−3.11, 3.28]
10.3 by long term	5	818	Mean Difference (IV, Random, 95% CI)	3.41 [1.66, 5.16]
11 Global state: 3. Not compliant with medication - by long term	1	71	Risk Ratio (M-H, Random, 95% CI)	0.35 [0.15, 0.81]
12 Social functioning: 1. Contact with legal system (various measurements)	10		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
12.1 by short term - contact with the police	1	61	Risk Ratio (M-H, Random, 95% CI)	2.57 [0.73, 9.04]
12.2 by medium term -arrested	3	604	Risk Ratio (M-H, Random, 95% CI)	1.08 [0.61, 1.90]
12.3 by medium term -contact with the police	1	88	Risk Ratio (M-H, Random, 95% CI)	0.23 [0.09, 0.55]
12.4 by medium term -imprisoned	4	804	Risk Ratio (M-H, Random, 95% CI)	0.80 [0.39, 1.64]
12.5 by long term - arrested	1	178	Risk Ratio (M-H, Random, 95% CI)	0.66 [0.32, 1.37]
12.6 by long term -imprisoned	4	361	Risk Ratio (M-H, Random, 95% CI)	0.72 [0.31, 1.67]
13 Social functioning: 2. Employment status (various measurements)	6		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
13.1 by medium term - not competitively employed at the end of the trial	1	88	Risk Ratio (M-H, Random, 95% CI)	1.00 [0.91, 1.10]
13.2 by medium term - not employed at the end of the trial	4	1136	Risk Ratio (M-H, Random, 95% CI)	0.89 [0.79, 1.00]
13.3 by long term - not employed at the end of the trial	4	1129	Risk Ratio (M-H, Random, 95% CI)	0.70 [0.49, 1.00]
14 Social functioning: 3a. Accommodation status (various measurements)	10		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
14.1 by short term -homelessness	1	95	Risk Ratio (M-H, Random, 95% CI)	0.04 [0.00, 0.70]
14.2 by medium term -homelessness	1	88	Risk Ratio (M-H, Random, 95% CI)	0.32 [0.03, 2.95]
14.3 by medium term - not living independently	5	1303	Risk Ratio (M-H, Random, 95% CI)	0.80 [0.66, 0.97]
14.4 by long term -homelessness	3	418	Risk Ratio (M-H, Random, 95% CI)	0.78 [0.34, 1.82]
14.5 by long term - not living independently	4	1185	Risk Ratio (M-H, Random, 95% CI)	0.65 [0.49, 0.88]
14.6 by long term - not living in stable accommodation	1	168	Risk Ratio (M-H, Random, 95% CI)	0.80 [0.65, 0.98]
15 Social functioning: 3b. Accommodation status (various measurements, skewed data)			Other data	No numeric data
15.1 by medium term -average days per month in stable accommodation			Other data	No numeric data
15.2 by long term - average days per month in sheltered homes			Other data	No numeric data
16 Social functioning: 4a. Substance abuse - by long term	1		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
16.1 alcohol abuse	1	547	Risk Ratio (M-H, Random, 95% CI)	0.55 [0.26, 1.17]
16.2 illicit drug abuse	1	547	Risk Ratio (M-H, Random, 95% CI)	0.96 [0.63, 1.47]
17 Social functioning: 4b. Substance abuse (DALI, skewness not detectable) - by medium term			Other data	No numeric data
17.1 alcohol abuse (DALI, −4 to + 6, high = worse)			Other data	No numeric data
17.2 drug abuse (DALI, − 4 to + 4, high = worse)			Other data	No numeric data
18 Social functioning: 4c. Substance abuse - days used per month (skewed data)			Other data	No numeric data
18.1 by medium term			Other data	No numeric data
18.2 by long term			Other data	No numeric data
19 Social functioning: 5a. Average endpoint score (various scales)	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
19.1 by short term (RFS, low = poor)	1	80	Mean Difference (IV, Random, 95% CI)	−0.62 [−2.23, 0.99]
19.2 by short term (SAS-adapted version, low = poor)	1	80	Mean Difference (IV, Random, 95% CI)	−3.34 [−7.55, 0.87]
19.3 by medium term - social role performance (DAS, high = poor)	1	55	Mean Difference (IV, Random, 95% CI)	0.10 [−0.40, 0.60]
19.4 by medium term (RFS, low = poor)	1	80	Mean Difference (IV, Random, 95% CI)	−0.86 [−2.72, 1.00]
19.5 by medium term (SAS-adapted version, low = poor)	1	80	Mean Difference (IV, Random, 95% CI)	−3.30 [−7.83, 1.23]
19.6 by long term - social role performance (DAS, high = poor)	1	58	Mean Difference (IV, Random, 95% CI)	−0.20 [−0.67, 0.27]
19.7 by long term (ISSI, low = poor)	1	62	Mean Difference (IV, Random, 95% CI)	3.20 [0.11, 6.29]
19.8 by long term (RFS, low = poor)	1	80	Mean Difference (IV, Random, 95% CI)	−2.35 [−4.05, −0.65]
19.9 by long term (SAS-adapted version, low = poor)	1	80	Mean Difference (IV, Random, 95% CI)	−2.75 [−7.13, 1.63]
19.10 by long term (Strauss-Carpenter Scale, low = poor)	1	60	Mean Difference (IV, Random, 95% CI)	0.10 [−1.17, 1.37]
20 Social functioning: 5b. Average endpoint score (various scales, skewed data)			Other data	No numeric data
20.1 by short term (SAS, high = poor)			Other data	No numeric data
20.2 by medium term (SAS, high = poor)			Other data	No numeric data
20.3 by long term (SAS, high = poor)			Other data	No numeric data
20.4 by long term (REHAB, high = poor)			Other data	No numeric data
21 Mental state: 1a. General symptoms - average endpoint score (various scales)	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
21.1 by short term (BPRS-18 items, high = poor)	2	668	Mean Difference (IV, Random, 95% CI)	−1.56 [−6.85, 3.73]
21.2 by short term (BSI, high = poor)	2	668	Mean Difference (IV, Random, 95% CI)	−0.06 [−0.19, 0.06]
21.3 by short term (CSI, low = poor)	1	125	Mean Difference (IV, Random, 95% CI)	0.56 [0.28, 0.84]
21.4 by medium term (BPRS-18 items, high = poor)	2	662	Mean Difference (IV, Random, 95% CI)	−0.96 [−2.42, 0.51]
21.5 by medium term (BSI, high = poor)	2	662	Mean Difference (IV, Random, 95% CI)	−0.02 [−0.15, 0.10]
21.6 medium term (CSI, low = poor)	1	125	Mean Difference (IV, Random, 95% CI)	0.35 [0.05, 0.65]
21.7 by long term (BPRS-18 items, high = poor)	2	647	Mean Difference (IV, Random, 95% CI)	−2.65 [−4.11, −1.20]
21.8 by long term (BSI, high = poor)	2	647	Mean Difference (IV, Random, 95% CI)	−0.18 [−0.31, −0.06]
22 Mental state: 1b. General symptoms - mean change from baseline (CSI, low = poor ) - by long term	1	168	Mean Difference (IV, Random, 95% CI)	−0.32 [−0.53, −0.11]
23 Mental state: 1c. General symptoms - average endpoint score (various scales, skewed data)			Other data	No numeric data
23.1 by short term (BPRS-24 items, high = poor)			Other data	No numeric data
23.3 by short term (PSE, high = poor)			Other data	No numeric data
23.4 by medium term (BPRS-24 items, high = poor)			Other data	No numeric data
23.7 by medium term (CPRS, high = poor)			Other data	No numeric data
23.8 by medium term (PSE, high = poor)			Other data	No numeric data
23.9 by long term (BPRS-18 items, high = poor)			Other data	No numeric data
23.11 by long term (BPRS-24 items, high = poor)			Other data	No numeric data
23.12 by long term (CPRS, high = poor)			Other data	No numeric data
23.13 by long term (PSE, high = poor)			Other data	No numeric data
23.14 by long term (SCL-90, high = poor)			Other data	No numeric data
24 Mental state: 2a. Specific symptoms - depression at follow up interview	1		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
24.1 by medium term	1	547	Risk Ratio (M-H, Random, 95% CI)	0.77 [0.56, 1.04]
24.2 by long term	1	547	Risk Ratio (M-H, Random, 95% CI)	0.83 [0.57, 1.21]
25 Mental state: 2b. Specific symptoms - average endpoint score (various scales, skewed data)			Other data	No numeric data
25.1 by medium term -depression symptoms (BDI, high = poor)			Other data	No numeric data
25.2 by medium term -negative symptoms (SANS, high = poor)			Other data	No numeric data
25.3 by long term - depression symptoms (BDI, high = poor)			Other data	No numeric data
25.6 by long term - negative symptoms (SANS, high = poor)			Other data	No numeric data
26 Behaviour: 1. Specific behaviour - self-harm	4		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
26.1 by medium term	2	620	Risk Ratio (M-H, Random, 95% CI)	0.99 [0.61, 1.59]
26.2 by long term	2	374	Risk Ratio (M-H, Random, 95% CI)	0.82 [0.38, 1.78]
26.3 by long term (during previous 12 months)	1	547	Risk Ratio (M-H, Random, 95% CI)	0.81 [0.47, 1.38]
27 Behaviour: 2. Social behaviour - average endpoint score (SBS, high = poor)			Other data	No numeric data
27.1 by medium term			Other data	No numeric data
27.2 by long term			Other data	No numeric data
28 Quality of life: 1a. Average endpoint score (various scales)	6		Mean Difference (IV, Random, 95% CI)	Subtotals only
28.1 by short term - general well-being (QOLI, high = better)	1	125	Mean Difference (IV, Random, 95% CI)	0.53 [0.09, 0.97]
28.2 by medium term (LQoLP, high = better)	1	52	Mean Difference (IV, Random, 95% CI)	0.09 [−0.60, 0.78]
28.3 by medium term (MANSA - range 1-7, high = better)	1	81	Mean Difference (IV, Random, 95% CI)	0.20 [−0.29, 0.69]
28.4 by long term (LQoLP, high = better)	2	113	Mean Difference (IV, Random, 95% CI)	−0.23 [−0.55, 0.08]
28.5 by long term (QOLI, high = better)	2	132	Mean Difference (IV, Random, 95% CI)	0.09 [−0.24, 0.42]
29 Quality of life: 1b. Mean change from baseline (QOLI, high = better, skewed data) - by long term			Other data	No numeric data
30 Participant satisfaction: 1a. Average endpoint score (CSQ, high = better)	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
30.1 by short term	1	61	Mean Difference (IV, Random, 95% CI)	6.20 [2.60, 9.80]
30.2 by medium term	2	500	Mean Difference (IV, Random, 95% CI)	1.93 [0.86, 3.01]
30.3 by long term	2	423	Mean Difference (IV, Random, 95% CI)	3.23 [2.31, 4.14]
31 Participants satisfaction: 1b. Average endpoint score (CSQ, high = better, skewed data) - by short term			Other data	No numeric data
32 Participants need: 1. Average endpoint score (various scales, skewed data)			Other data	No numeric data
32.2 by medium term - met needs (CANSAS, high = better)			Other data	No numeric data
32.3 by medium term - unmet needs (CANSAS, high = poor)			Other data	No numeric data
32.5 by long term (CANSAS, high = poor)			Other data	No numeric data
32.7 by long term (CAN, high = poor)			Other data	No numeric data
33 Costs: 1a. Direct costs of psychiatric hospital care - by medium term (Unit cost = US $, fiscal year 1990)	2	426	Mean Difference (IV, Random, 95% CI)	−143.74 [−272.40, −15.08]
34 Costs: 1b. Direct costs of psychiatric hospital care -skewed data			Other data	No numeric data
34.3 by medium term			Other data	No numeric data
34.4 by long term			Other data	No numeric data
35 Costs: 2a. Direct health care costs - by long term (Unit cost = US $, fiscal year 1988)	2	873	Mean Difference (IV, Random, 95% CI)	−529.24 [−2143.59, 1085.10]
36 Costs: 2b. Direct health care costs - by medium term (skewed data)			Other data	No numeric data
37 Costs: 3. Direct costs of all care			Other data	No numeric data
37.2 by short term			Other data	No numeric data
37.3 by medium term			Other data	No numeric data
37.4 by long term			Other data	No numeric data

Comparison 2. INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Service use: 1. Average number of days in hospital per month - by about 24 months	21	2220	Mean Difference (IV, Random, 95% CI)	−0.08 [−0.37, 0.21]
1.1 skewed data (sample size ≥ 200)	3	694	Mean Difference (IV, Random, 95% CI)	−0.58 [−1.93, 0.76]
1.2 skewed data (sample size < 200)	18	1526	Mean Difference (IV, Random, 95% CI)	−0.03 [−0.33, 0.28]
2 Service use: 2. Not remaining in contact with psychiatric services	4		Risk Ratio (M-H, Random, 95% CI)	Totals not selected
2.1 by medium term	1		Risk Ratio (M-H, Random, 95% CI)	Not estimable
2.2 by long term	3		Risk Ratio (M-H, Random, 95% CI)	Not estimable
3 Service use: 3a. Admitted to hospital - by long term	3	1132	Risk Ratio (M-H, Random, 95% CI)	0.91 [0.75, 1.12]
4 Service use: 3b. Average number of admissions (skewed data - sample size ≥ 200) - by long term	1	678	Mean Difference (IV, Random, 95% CI)	−0.18 [−0.41, 0.05]
5 Service use: 3c. Average number of admissions (skewed data) - by medium term			Other data	No numeric data
6 Adverse event: 1a. Death - any cause	7		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
6.1 by short term	1	193	Risk Ratio (M-H, Random, 95% CI)	Not estimable
6.2 by medium term	3	294	Risk Ratio (M-H, Random, 95% CI)	2.92 [0.12, 69.43]
6.3 by long term	5	1637	Risk Ratio (M-H, Random, 95% CI)	0.90 [0.46, 1.75]
7 Adverse event: 1b. Death -suicide	8		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
7.1 by short term	1	193	Risk Ratio (M-H, Random, 95% CI)	Not estimable
7.2 by medium term	6	929	Risk Ratio (M-H, Random, 95% CI)	1.61 [0.26, 9.85]
7.3 by long term	3	1152	Risk Ratio (M-H, Random, 95% CI)	0.88 [0.27, 2.84]
8 Global state: 1. Leaving the study early	9	2195	Risk Ratio (M-H, Random, 95% CI)	0.72 [0.52, 0.99]
8.1 by medium term	2	225	Risk Ratio (M-H, Random, 95% CI)	0.64 [0.13, 3.07]
8.2 by long term	7	1970	Risk Ratio (M-H, Random, 95% CI)	0.70 [0.52, 0.95]
9 Global state: 2a. Average endpoint score (HoNOS, high = poor) - by long term	1	239	Mean Difference (IV, Random, 95% CI)	−0.40 [−1.77, 0.97]
10 Global state: 2b. Average endpoint score (HoNOS, high = poor) - skewed data			Other data	No numeric data
10.1 medium term			Other data	No numeric data
10.2 long term			Other data	No numeric data
11 Global state: 3a. Not compliant with medication - by medium term	1	73	Risk Ratio (M-H, Random, 95% CI)	1.14 [0.42, 3.05]
12 Global state: 3b. Compliance with medication - average endpoint sub-scale score (ROMI) - by long term	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
12.1 compliance sub-scale (high = good)	1	239	Mean Difference (IV, Random, 95% CI)	0.60 [−0.05, 1.25]
12.2 non-compliance sub-scale (high = poor)	1	239	Mean Difference (IV, Random, 95% CI)	−0.60 [−1.63, 0.43]
13 Global state: 3c. Compliance with medication - average endpoint sub-scale score (ROMI, score 1-3, skewed data)			Other data	No numeric data
13.1 medium term -compliance sub-scale (high = good)			Other data	No numeric data
13.2 medium term - non-compliance sub-scale (high = poor)			Other data	No numeric data
13.3 long term - compliance sub-scale (high = good)			Other data	No numeric data
13.4 long term -non-compliance sub-scale (high = poor)			Other data	No numeric data
14 Social functioning: 1. Contact with legal system (various measurements)	3	1283	Risk Ratio (M-H, Random, 95% CI)	0.95 [0.65, 1.37]
14.1 by medium term -contact with the police	1	73	Risk Ratio (M-H, Random, 95% CI)	0.32 [0.04, 2.97]
14.2 by long term -imprisoned	2	959	Risk Ratio (M-H, Random, 95% CI)	1.15 [0.64, 2.08]
14.3 by long term - arrested	1	251	Risk Ratio (M-H, Random, 95% CI)	0.87 [0.53, 1.42]
15 Social functioning 2. Employment status - by medium term	1		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
15.1 spent > 1 day employed	1	73	Risk Ratio (M-H, Random, 95% CI)	1.46 [0.45, 4.74]
15.2 on paid employment	1	73	Risk Ratio (M-H, Random, 95% CI)	0.97 [0.14, 6.54]
16 Social functioning: 3a. Accommodation status (various measurements)	2		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
16.1 by medium term - living in supported accommodation	1	73	Risk Ratio (M-H, Random, 95% CI)	2.59 [0.75, 9.01]
16.2 by long term -homelessness	1	251	Risk Ratio (M-H, Random, 95% CI)	0.69 [0.34, 1.38]
17 Social functioning: 3b. Accommodation status -average days per month in stable accommodation	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
17.1 by short term	1	203	Mean Difference (IV, Random, 95% CI)	−0.20 [−2.48, 2.08]
17.2 by medium term	1	203	Mean Difference (IV, Random, 95% CI)	0.10 [−2.15, 2.35]
17.3 by long term	2	901	Mean Difference (IV, Random, 95% CI)	−0.19 [−1.37, 1.00]
18 Social functioning: 4a. Substance abuse - by long term	2		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
18.1 alcohol abuse	1	251	Risk Ratio (M-H, Random, 95% CI)	1.10 [0.67, 1.83]
18.2 illicit drug abuse	1	251	Risk Ratio (M-H, Random, 95% CI)	1.08 [0.69, 1.71]
18.3 alcohol - remission from alcohol use disorder (AUS score < 3)	1	223	Risk Ratio (M-H, Random, 95% CI)	0.86 [0.65, 1.14]
19 Social functioning: 4b. Substance abuse - average endpoint score (SATS, low = poor)	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
19.1 by short term	1	203	Mean Difference (IV, Random, 95% CI)	0.07 [−0.28, 0.42]
19.2 by medium term	1	203	Mean Difference (IV, Random, 95% CI)	−0.11 [−0.55, 0.33]
19.3 by long term	1	203	Mean Difference (IV, Random, 95% CI)	0.11 [−0.41, 0.63]
20 Social functioning: 4c. Alcohol - abuse (various measurements, skewed data)			Other data	No numeric data
20.1 short term - days using alcohol during previous 6 months (TLFB)			Other data	No numeric data
20.2 short term - average endpoint score (AUS, high = poor)			Other data	No numeric data
20.3 medium term - days using alcohol during previous 6 months (TLFB)			Other data	No numeric data
20.4 medium term - average endpoint score (AUS, high = poor)			Other data	No numeric data
20.5 long term - days using alcohol during previous 6 months (TLFB)			Other data	No numeric data
20.6 long term - average endpoint score (AUS, high = poor)			Other data	No numeric data
21 Social functioning: 5a. Average endpoint score (LSP, high = poor) - by long term	1	239	Mean Difference (IV, Fixed, 95% CI)	4.0 [−0.61, 8.61]
22 Social functioning: 5b.Average endpoint score (SFQ, high = poor) - skewed data			Other data	No numeric data
22.1 by medium term			Other data	No numeric data
22.2 by long term			Other data	No numeric data
23 Mental state: 1a. General symptoms - average endpoint score (various scales)	2		Mean Difference (IV, Random, 95% CI)	Subtotals only
23.1 by short term (BPRS-24 items, high = poor)	1	203	Mean Difference (IV, Random, 95% CI)	−0.65 [−3.99, 2.69]
23.2 by medium term (BPRS-24 items, high = poor)	1	203	Mean Difference (IV, Random, 95% CI)	−1.62 [−4.76, 1.52]
23.3 by long term (BPRS-24 items, high = poor)	1	203	Mean Difference (IV, Random, 95% CI)	−0.22 [−3.32, 2.88]
23.4 by long term (CPRS, high = poor)	1	595	Mean Difference (IV, Random, 95% CI)	0.40 [−1.83, 2.63]
24 Mental state: 1b. General symptoms - average endpoint scores (various scales, skewed data)			Other data	No numeric data
24.1 by medium term (Krawiecka Scale, high = poor)			Other data	No numeric data
24.2 by long term (Krawiecka Scale, high = poor)			Other data	No numeric data
24.3 by long term (BPRS 24-items, high = good)			Other data	No numeric data
25 Mental state: 2a. Specific symptoms: negative symptoms - average endpoint score (SANS, high = poor) - by long term	1	593	Mean Difference (IV, Random, 95% CI)	0.20 [−2.32, 2.72]
26 Mental state: 2b. Specific symptoms - average endpoint scores (various scales, skewed data)			Other data	No numeric data
26.1 medium term - anxiety (HADS, high = poor)			Other data	No numeric data
26.2 medium term -depression (HADS, high = poor)			Other data	No numeric data
26.3 long term - anxiety (HADS, high = poor)			Other data	No numeric data
26.5 long term - depression (HADS, high = poor)			Other data	No numeric data
27 Behaviour: 1. Specific behaviour (various measurements)	3		Risk Ratio (M-H, Random, 95% CI)	Subtotals only
27.1 by medium term - harm to self or others	1	73	Risk Ratio (M-H, Random, 95% CI)	0.88 [0.40, 1.90]
27.2 by long term - self-harm	1	708	Risk Ratio (M-H, Random, 95% CI)	1.06 [0.70, 1.61]
27.3 by long term - injury to others	2	959	Risk Ratio (M-H, Random, 95% CI)	1.09 [0.85, 1.40]
28 Quality of life: 1. Average endpoint score (various scales)	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
28.1 by short term - overall life satisfaction (QOLI, high = better)	1	203	Mean Difference (IV, Random, 95% CI)	−0.02 [−0.43, 0.39]
28.2 by medium term - overall life satisfaction (QOLI, high = better)	1	203	Mean Difference (IV, Random, 95% CI)	−0.04 [−0.43, 0.35]
28.3 by long term (LQoL, high = better)	1	526	Mean Difference (IV, Random, 95% CI)	0.03 [−0.10, 0.16]
28.4 by long term (MANSA, range 1-7, high = better)	1	166	Mean Difference (IV, Random, 95% CI)	0.10 [−0.19, 0.39]
28.5 by long term - overall life satisfaction (QOLI, high = better)	1	203	Mean Difference (IV, Random, 95% CI)	0.10 [−0.25, 0.45]
29 Participant satisfaction/need: 1a. Average endpoint scores (various scale) - by long term	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
29.1 Patient need: CAN (high = poor)	1	585	Mean Difference (IV, Random, 95% CI)	−0.29 [−0.69, 0.11]
29.2 Patient Satisfaction with Health Services (high = poor)	1	490	Mean Difference (IV, Random, 95% CI)	−0.40 [−1.25, 0.45]
30 Participant need: 1b. Average endpoint scores (CAN, high = poor) - skewed data)			Other data	No numeric data
30.1 by medium term			Other data	No numeric data
30.2 by long term			Other data	No numeric data
31 Participant satisfaction: 1. Average endpoint scores (CSQ-modified, high = better, skewed data) - by long term			Other data	No numeric data
32 Costs: 1. Direct costs of psychiatric hospital care (skewed data)			Other data	No numeric data
32.3 by medium term			Other data	No numeric data
32.4 by long term			Other data	No numeric data
33 Costs: 2a. Direct costs of all care - by long term (Unit cost £, fiscal year 1997/98)	1	667	Mean Difference (IV, Random, 95% CI)	77.0 [−66.63, 220. 63]
34 Costs: 2b. Direct costs of all care - by medium term (skewed data)			Other data	No numeric data

Analysis 1.1. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 1 Service use: 1. Average number of days in hospital per month - by about 24 months.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 1 Service use: 1. Average number of days in hospital per month - by about 24 months

Analysis 1.2. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 2 Service use: 2. Not remaining in contact with psychiatric services.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 2 Service use: 2. Not remaining in contact with psychiatric services

Analysis 1.3. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 3 Service use: 3a. Admitted to hospital.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 3 Service use: 3a. Admitted to hospital

Analysis 1.4. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 4 Service use: 3b. Average number of admissions per month (skewed data).

Study	Intervention	Mean	SD	Total	Note
by medium term
Bond-Chicagol	1. ICM	0.16	0.15	42
Bond-Chicagol	2. Standard care	0.26	0.23	40
by long term
Audini-UK	l. ICM	0.02	0.05*	33
Audini-UK	2. Standard care	0.03	0.06*	33	* Carried over from Sytema-Netherlands.
Muller-Clemm-Canada	l. ICM	0.09	0.05*	61
Muller-Clemm-Canada	2. Standard care	0.08	0.06*	57	* Carried over from Sytema-Netherlands.
Sytema-Netherlands	l.ICM	0.05	0.05	58
Sytema-Netherlands	2. Standard care	0.05	0.06	57

Analysis 1.5. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 5 Service use: 4a. Admitted to ER - by long term.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 5 Service use: 4a. Admitted to ER - by long term

Analysis 1.6. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 6 Service use: 4b. Average number of admissions to ER (skewed data) - by medium term.

Study	Intervention	Mean	SD	Total	Note
Jerrell-SCarolinal	1. ICM	0.85	1.7*	40
Jerrell-SCarolinal	2. Standard care	0.73	3.3*	40	* Carried over from Lehman-Maryland1.
Lehman-Marylandl	l. ICM	0.9	1.7	77
Lehman-Marylandl	2. Standard care	2	3.3	75

Analysis 1.7. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 7 Adverse event: 1a. Death - any cause.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 7 Adverse event: 1a. Death - any cause

Analysis 1.8. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 8 Adverse event: 1b. Death - suicide.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 8 Adverse event: 1b. Death - suicide

Analysis 1.9. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 9 Global state: 1. Leaving the study early.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 9 Global state: 1. Leaving the study early

Analysis 1.10. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 10 Global state: 2. Average endpoint score (GAF, high = good).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 10 Global state: 2. Average endpoint score (GAF, high = good)

Analysis 1.11. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 11 Global state: 3. Not compliant with medication - by long term.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 11 Global state: 3. Not compliant with medication - by long term

Analysis 1.12. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 12 Social functioning: 1. Contact with legal system (various measurements).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 12 Social functioning: 1. Contact with legal system (various measurements)

Analysis 1.13. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 13 Social functioning: 2. Employment status (various measurements).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 13 Social functioning: 2. Employment status (various measurements)

Analysis 1.14. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 14 Social functioning: 3a. Accommodation status (various measurements).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 14 Social functioning: 3a. Accommodation status (various measurements)

Analysis 1.15. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 15 Social functioning: 3b. Accommodation status (various measurements, skewed data).

Study	Intervention	Mean	SD	Total
by medium term - average days per month in stable accommodation
Lehman-Marylandl	1. ICM	17.5	9	77
Lehman-Marylandl	2. Standard care	13.34	9	75
Morse-Missouri3	1. ICM	5.77	7.42	54
Morse-Missouri3	2. Standard care	5.02	8.62	49
by long term - average days per month in sheltered homes
Morse-Missouri3	1. ICM	17.78	12.68	54
Morse-Missouri3	2. Standard care	12.59	13.27	49
Sytema-Netherlands	1. ICM	2.8	7.4	58
Sytema-Netherlands	2. Standard care	3.6	9.2	57

Analysis 1.16. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 16 Social functioning: 4a. Substance abuse - by long term.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 16 Social functioning: 4a. Substance abuse - by long term

Analysis 1.17. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 17 Social functioning: 4b. Substance abuse (DALI, skewness not detectable) - by medium term.

Study	Intervention	Mean	SD	Total
alcohol abuse (DALI, −4 to + 6, high = worse)
Sytema-Netherlands	1. ICM	−0.8	2.7	45
Sytema-Netherlands	2. Standard Care	−1.2	2.4	36
drug abuse (DALI, − 4 to + 4, high = worse)
Sytema-Netherlands	1. ICM	−1.4	1.3	45
Sytema-Netherlands	2. Standard Care	−1.8	1.3	36

Analysis 1.18. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 18 Social functioning: 4c. Substance abuse - days used per month (skewed data).

Study	Intervention	Mean	SD	Total
by medium term
Morse-Missouri3	1. ICM	6.25	7.84	54
Morse-Missouri3	2. Standard care	6.34	7.52	49
by long term
Morse-Missouri3	1. ICM	6.77	8.86	54
Morse-Missouri3	2. Standard care	6.42	7.84	49

Analysis 1.19. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 19 Social functioning: 5a. Average endpoint score (various scales).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 19 Social functioning: 5a. Average endpoint score (various scales)

Analysis 1.20. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 20 Social functioning: 5b. Average endpoint score (various scales, skewed data).

Study	Intervention	Mean	SD	Total
by short term (SAS, high = poor)
Muijen-UK2	1. ICM	3.9	1.1	35
Muijen-UK2	2. Standard care	3.9	1.6	29
by medium term (SAS, high = poor)
Muijen-UK2	1. ICM	4.3	1.6	29
Muijen-UK2	2. Standard care	3.7	1.5	25
by long term (SAS, high = poor)
Audini-UK	1. ICM	3.0	1.6	30
Audini-UK	2. Standard care	2.9	1.1	28
Muijen-UK2	1. ICM	3.6	1.4	24
Muijen-UK2	2. Standard care	4.2	1.4	22
by long term (REHAB, high = poor)
Marshall-UK	1. ICM	31.7	29.3	31
Marshall-UK	2. Standard care	40.83	19.65	30

Analysis 1.21. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 21 Mental state: 1a. General symptoms - average endpoint score (various scales).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 21 Mental state: 1a. General symptoms - average endpoint score (various scales)

Analysis 1.22. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 22 Mental state: 1b. General symptoms - mean change from baseline (CSI, low = poor ) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 22 Mental state: 1b. General symptoms - mean change from baseline (CSI, low = poor ) - by long term

Analysis 1.23. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 23 Mental state: 1c. General symptoms - average endpoint score (various scales, skewed data).

Study	Intervention	Mean	SD	Total
by short term (BPRS-24 items, high = poor)
Audini-UK	1. ICM	39.1	10.0	31
Audini-UK	2. Standard Care	39.5	12.0	30
Muijen-UK2	1. ICM	42.4	16.8	36
Muijen-UK2	2. Standard Care	43.1	15.2	32
by short term (PSE, high = poor)
Audini-UK	1. ICM	7.2	7.2	31
Audini-UK	2. Standard Care	8.4	9.3	30
Muijen-UK2	1. ICM	19.9	19.5	35
Muijen-UK2	2. Standard Care	17.3	15.8	32
by medium term (BPRS-24 items, high = poor)
Audini-UK	1. ICM	42.3	14.8	30
Audini-UK	2. Standard Care	41.4	12.2	28
Muijen-UK2	1. ICM	45.7	15.2	32
Muijen-UK2	2. Standard Care	43.1	12.7	26
Sytema-Netherlands	1. ICM	38	10	45
Sytema-Netherlands	2. Standard Care	42	11	36
by medium term (CPRS, high = poor)
Holloway-UK	1. ICM	20.6	12.1	22
Holloway-UK	2. Standard Care	21.3	14.0	22
by medium term (PSE, high = poor)
Muijen-UK2	1. ICM	18.7	15.9	35
Muijen-UK2	2. Standard Care	14.4	15	27
by long term (BPRS-18 items, high = poor)
Ford-UK	1. ICM	12.8	9.6	36
Ford-UK	2. Standard Care	13.5	11.9	32
by long term (BPRS-24 items, high = poor)
Muijen-UK2	1. ICM	44.4	13.3	31
Muijen-UK2	2. Standard Care	51.8	18.8	26
by long term (CPRS, high = poor)
Holloway-UK	1. ICM	21.6	12.9	21
Holloway-UK	2. Standard Care	22.4	14.5	19
by long term (PSE, high = poor)
Audini-UK	1. ICM	7.6	8.2	30
Audini-UK	2. Standard Care	10.6	12.2	28
Muijen-UK2	1. ICM	20.3	13.7	28
Muijen-UK2	2. Standard Care	27.6	23.5	25
by long term (SCL-90, high = poor)
Bjorkman-Sweden	1. ICM	102	68.5	27
Bjorkman-Sweden	2. Standard Care	81.4	55.1	33

Analysis 1.24. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 24 Mental state: 2a. Specific symptoms - depression at follow up interview.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 24 Mental state: 2a. Specific symptoms - depression at follow up interview

Analysis 1.25. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 25 Mental state: 2b. Specific symptoms - average endpoint score (various scales, skewed data).

Study	Intervention	Mean	SD	Total
by medium term - depression symptoms (BDI, high = poor)
Holloway-UK	1. ICM	11.5	8.9	23
Holloway-UK	2. Standard care	18.5	13.9	19
by medium term - negative symptoms (SANS, high = poor)
Holloway-UK	1. ICM	7.3	4	26
Holloway-UK	2. Standard care	6.3	4.4	22
by long term - depression symptoms (BDI, high = poor)
Holloway-UK	1. ICM	12.8	8.1	25
Holloway-UK	2. Standard care	14.8	11.5	17
by long term - negative symptoms (SANS, high = poor)
Holloway-UK	1. ICM	7.3	3.7	26
Holloway-UK	2. Standard care	7.1	4.1	20

Analysis 1.26. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 26 Behaviour: 1. Specific behaviour - self-harm.

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 26 Behaviour: 1. Specific behaviour - self-harm

Analysis 1.27. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 27 Behaviour: 2. Social behaviour - average endpoint score (SBS, high = poor).

Study	Intervention	Mean	SD	Total
by medium term
Holloway-UK	1. ICM	3.2	2.8	33
Holloway-UK	2. Standard Care	2.4	2.9	30
by long term
Holloway-UK	1. ICM	3.3	2.8	34
Holloway-UK	2. Standard Care	2.7	2.2	26

Analysis 1.28. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 28 Quality of life: 1a. Average endpoint score (various scales).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 28 Quality of life: 1a. Average endpoint score (various scales)

Analysis 1.29. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 29 Quality of life: 1 b. Mean change from baseline (QOLI, high = better, skewed data) - by long term.

Study	Intervention	Mean	SD	Total
Shern-USA1	1. ICM	1.19	1.99	91
Shern-USA1	2. Standard Care	−0.02	1.65	77

Analysis 1.30. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 30 Participant satisfaction: 1a. Average endpoint score (CSQ, high = better).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 30 Participant satisfaction: 1a. Average endpoint score (CSQ, high = better)

Analysis 1.31. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 3 1 Participants satisfaction: 1 b. Average endpoint score (CSQ, high = better, skewed data) - by short term.

Study	Intervention	Mean	SD	Total	Note
Muijen-UK2	1. ICM	26	5.4	30
Muijen-UK2	2. Standard Care*	22	7.5	13	* Attrition >50%.

Analysis 1.32. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 32 Participants need: 1. Average endpoint score (various scales, skewed data).

Study	Intervention	Mean	SD	Total	Note
by medium term - met needs (CANSAS, high = better)
Sytema-Netherlands	1. ICM	8.5	4.5	45
Sytema-Netherlands	2. Standard Care	8.6	4.7	36
by medium term - unmet needs (CANSAS, high = poor)
Sytema-Netherlands	1. ICM	1.4	1.9	45
Sytema-Netherlands	2. Standard Care	1.6	1.7	36
by long term (CANSAS, high = poor)
REACT-UK	1. ICM	3.3	2.7	91
REACT-UK	2. Standard Care	3.4	2.9	75
by long term (CAN, high = poor)
Bjorkman-Sweden	1. ICM	3.2	1.8	28
Bjorkman-Sweden	2. Standard Care	4.6	3.8	36

Analysis 1.33. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 33 Costs: 1a. Direct costs of psychiatric hospital care - by medium term (Unit cost = US $, fiscal year 1990).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 33 Costs: 1a. Direct costs of psychiatric hospital care - by medium term (Unit cost = US $, fiscal year 1990)

Analysis 1.34. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 34 Costs: 1 b. Direct costs of psychiatric hospital care - skewed data.

Study	Intervention	Mean	SD	Total	Note
by medium term
Lehman-Marylandl	1. ICM*	2,619	4,440	77
Lehman-Marylandl	2. Standard care*	4,662	6,034	75	* Unit cost US $, fiscal year 1994. t-value=2.34 Time period 12 months
Morse-Missouri3	l. ICM*	624	2,314	54
Morse-Missouri3	2. Standard care*	439	1,596	49	* Unit cost US $, fiscal year 2001. “No main effect of of treatment condition for inpatient costs, F (2, 146)=0.10, p=0.9,5 f2=0.01.” Time period 6 months.
by long term
Ford-UK	l. ICM*	378	846	39
Ford-UK	2. Standard care*	237	492	38	* Unit cost £, fiscal year not reported, study base year 1990. ** No statistical analysis available from the paper. Time period 18 months
Morse-Missouri3	1. ICM*	855	2,356	54
Morse-Missouri3	2. Standard care*	455	1,065	49	* Unit cost US $, fiscal year 2001. ** “No main effect of treatment condition for inpatient costs, F (2, 146)=0.10, p=0.9, I2=0.01.” Time period 6 months.
Quinlivan-California	1. ICM*	301	397	30
Quinlivan-California	2. Standard care*	1,636	2,593	30	* Unit cost US $, fiscal year not reported, but study was carried on from April 1990 to March 1992. ** “Costs significantly lower for the ICM group (F=4.32, df=2. 87, p=0.02.)” Time period 24 months

Analysis 1.35. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 35 Costs: 2a. Direct health care costs - by long term (Unit cost = US $, fiscal year 1988).

Review: Intensive case management for severe mental illness

Comparison: 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE

Outcome: 35 Costs: 2a. Direct health care costs - by long term (Unit cost = US $, fiscal year 1988)

Analysis 1.36. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 36 Costs: 2b. Direct health care costs - by medium term (skewed data).

Study	Intervention	Mean	SD	Total	Note
Lehman-Marylandl	1. ICM*	4,229	5,058	77
Lehman-Marylandl	2. Standard care*	5,540	6,368	75	* Unit cost US $, fiscal year 1994. ** ‘Total per-case cost did not reach statistical significance (p = 0.07). Transformation of total costs per case to account for non-normality (square root of total costs, t-test=0.77, df=1,134, NS) and non-parametric analysis (Wilcoxon test for ranks, Z=0.146, NS) also were non-significant.’ ***To Marina: Time period 12 months

Analysis 1.37. Comparison 1 INTENSIVE CASE MANAGEMENT versus STANDARD CARE, Outcome 3.

Study	Intervention	Mean	SD	Total	Note
by short term
Audini-UK	1.ICM*	4,264	1,768	33
Audini-UK	2. Standard care*	7,202	5,564	29	* Unit cost £, fiscal year 1996/7. ** ‘Bivariate cost comparisons (after log transformation) revealed significant advantage for ICM group (p=0.001)’ ***Time period missing XXX It is not the monthly cost per patient!!
by medium term
Marshall-UK	1. ICM*	1,044	425.3	31
Marshall-UK	2. Standard care*	1,108	530.4	30	* Unit cost £, fiscal year 1994. ** ‘No significant differences between two groups were found. ***To Marina: time period mean weekly cost;
Morse-Missouri3	1. ICM*	2,946.8	3,219.3	54
Morse-Missouri3	2. Standard care*	1,899.5	3,629.6	49	* Unit cost US $, fiscal year 2001. ** ‘There was a main effect of treatment condition on total costs, F(2, 146)=4.00, p=0.02, I2=0.05. Standard care condition had significantly lower costs than ICM.’ ***To Marina: time period 6 months
by long term
Audini-UK	1. ICM*	10,192	3,900	32
Audini-UK	2. Standard care*	15,288	17,160	28	* Unit cost £, fiscal year 1996/7. ** ‘Bivariate cost comparisons (after log transformation) did not revealed significant advantage for ICM group (p=0.09)’ ***Time period missing XXX It is not the monthly cost per patient!!
Ford-UK	1. ICM*	1,813	1,347	39
Ford-UK	2. Standard care*	717	768	38	* Unit cost £, fiscal year not reported, study base year 1990. ** ‘ANOVA analysis carried on, revealing significant advantage for ICM group (p<0.05).’ ***To Marina: time period 12 months
Marshall-UK	1. ICM*	996	398	31
Marshall-UK	2. Standard care*	1,088	562.4	30	* Unit cost £, fiscal year 1994. ** ‘No significant differences between two groups were found. ***To Marina: time period mean weekly cost
Morse-Missouri3	1. ICM*	3,190	3,441	54
Morse-Missouri3	2. Standard care*	1,467	2,173	49	* Unit cost US $, fiscal year 2001. ** ‘There was a main effect of treatment condition on total costs, F(2, 146)=4.00, p=0.02, fj 2=0.05. Standard care condition had significantly lower costs than ICM.’ ***To Marina: time period 6 months

Analysis 2.1. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 1 Service use: 1. Average number of days in hospital per month - by about 24 months.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 1 Service use: 1. Average number of days in hospital per month - by about 24 months

Analysis 2.2. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 2 Service use: 2. Not remaining in contact with psychiatric services.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 2 Service use: 2. Not remaining in contact with psychiatric services

Analysis 2.3. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 3 Service use: 3a. Admitted to hospital - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 3 Service use: 3a. Admitted to hospital - by long term

Analysis 2.4. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 4 Service use: 3b. Average number of admissions (skewed data -sample size ≥ 200) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 4 Service use: 3b. Average number of admissions (skewed data -sample size 200) - by long term

Analysis 2.5. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 5 Service use: 3c. Average number of admissions (skewed data) - by medium term.

Study	Intervention	Mean	SD	Total
Johnston-Australia	1. ICM	1.6	2	35
Johnston-Australia	2. non-ICM	1.9	2.4	33

Analysis 2.6. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 6 Adverse event: 1a. Death - any cause.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 6 Adverse event: 1a. Death - any cause

Analysis 2.7. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 7 Adverse event: 1b. Death - suicide.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 7 Adverse event: 1b. Death - suicide

Analysis 2.8. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 8 Global state: 1. Leaving the study early.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 8 Global state: 1. Leaving the study early

Analysis 2.9. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 9 Global state: 2a. Average endpoint score (HoNOS, high = poor) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 9 Global state: 2a. Average endpoint score (HoNOS, high = poor) - by long term

Analysis 2.10. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 10 Global state: 2b. Average endpoint score (HoNOS, high = poor) - skewed data.

Study	Intervention	Mean	SD	Total
medium term
Harrison-Read-UK	1. ICM	12	6.8	54
Harrison-Read-UK	2. non-ICM	11.4	6.4	64
long term
Harrison-Read-UK	1. ICM	11.9	5.9	60
Harrison-Read-UK	2. non-ICM	10.4	6.4	59

Analysis 2.11. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 11 Global state: 3a. Not compliant with medication - by medium term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 11 Global state: 3a. Not compliant with medication - by medium term

Analysis 2.12. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 12 Global state: 3b. Compliance with medication - average endpoint sub-scale score (ROMI) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 12 Global state: 3b. Compliance with medication - average endpoint sub-scale score (ROMI) - by long term

Analysis 2.13. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 13 Global state: 3c. Compliance with medication - average endpoint sub-scale score (ROMI, score 1-3, skewed data).

Study	Intervention	Mean	SD	Total
medium term - compliance sub-scale (high = good)
Harrison-Read-UK	1. ICM	1.8	0.4	49
Harrison-Read-UK	2. non-ICM	2.0	0.5	61
medium term - non-compliance sub-scale (high = poor)
Harrison-Read-UK	1. ICM	1.3	0.3	49
Harrison-Read-UK	2. non-ICM	1.2	0.3	61
long term - compliance sub-scale (high = good)
Harrison-Read-UK	1. ICM	1.8	0.4	62
Harrison-Read-UK	2. non-ICM	1.9	0.5	60
long term - non-compliance sub-scale (high = poor)
Harrison-Read-UK	1. ICM	1.2	0.3	63
Harrison-Read-UK	2. non-ICM	1.2	0.3	61

Analysis 2.14. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 14 Social functioning: 1. Contact with legal system (various measurements).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 14 Social functioning: 1. Contact with legal system (various measurements)

Analysis 2.15. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 15 Social functioning 2. Employment status - by medium term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 15 Social functioning 2. Employment status - by medium term

Analysis 2.16. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 16 Social functioning: 3a. Accommodation status (various measurements).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 16 Social functioning: 3a. Accommodation status (various measurements)

Analysis 2.17. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 17 Social functioning: 3b. Accommodation status - average days per month in stable accommodation.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 17 Social functioning: 3b. Accommodation status - average days per month in stable accommodation

Analysis 2.18. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 18 Social functioning: 4a. Substance abuse - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 18 Social functioning: 4a. Substance abuse - by long term

Analysis 2.19. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 19 Social functioning: 4b. Substance abuse - average endpoint score (SATS, low = poor).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 19 Social functioning: 4b. Substance abuse - average endpoint score (SATS, low = poor)

Analysis 2.20. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 20 Social functioning: 4c. Alcohol - abuse (various measurements, skewed data).

Study	Intervention	Mean	SD	Total
short term - days using alcohol during previous 6 months (TLFB)
Drake-NHamp	1. ICM	56.8	56.4	75
Drake-NHamp	2. non-ICM	47.5	58.4	68
short term - average endpoint score (AUS, high = poor)
Drake-NHamp	1. ICM	3.09	1.02	83
Drake-NHamp	2. non-ICM	2.91	1.08	73
medium term - days using alcohol during previous 6 months (TLFB)
Drake-NHamp	1. ICM	59.1	53.3	75
Drake-NHamp	2. non-ICM	42.8	52.9	68
medium term - average endpoint score (AUS, high = poor)
Drake-NHamp	1. ICM	3.11	1.05	83
Drake-NHamp	2. non-ICM	2.8	1.13	73
long term - days using alcohol during previous 6 months (TLFB)
Drake-NHamp	1. ICM	46.4	53.6	75
Drake-NHamp	2. non-ICM	43.6	57.3	68
long term - average endpoint score (AUS, high = poor)
Drake-NHamp	1. ICM	2.64	1.12	83
Drake-NHamp	2. non-ICM	2.77	1.18	73

Analysis 2.21. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 21 Social functioning: 5a. Average endpoint score (LSP, high = poor) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 21 Social functioning: 5a. Average endpoint score (LSP, high = poor) - by long term

Analysis 2.22. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 22 Social functioning: 5b.Average endpoint score (SFQ, high = poor) - skewed data.

Study	Intervention	Mean	SD	Tot
by medium term
Harrison-Read-UK	1. ICM	7.3	5.3	49
Harrison-Read-UK	2. non-ICM	7.5	5.1	62
by long term
Harrison-Read-UK	1. ICM	8.9	4.9	57
Harrison-Read-UK	2. non-ICM	7.9	4.9	58

Analysis 2.23. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 23 Mental state: 1a. General symptoms - average endpoint score (various scales).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 23 Mental state: 1a. General symptoms - average endpoint score (various scales)

Analysis 2.24. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 24 Mental state: 1b. General symptoms - average endpoint scores (various scales, skewed data).

Study	Intervention	Mean	SD	Total
by medium term (Krawiecka Scale, high = poor)
Harrison-Read-UK	1. ICM	8.8	5.6	57
Harrison-Read-UK	2. non-ICM	8	4.5	65
by long term (Krawiecka Scale, high = poor)
Harrison-Read-UK	1. ICM	9.2	5.5	47
Harrison-Read-UK	2. non-ICM	7.9	4.5	57
by long term (BPRS 24-items, high = good)
REACT-UK	1. ICM	32.9	9	91
REACT-UK	2. non-ICM	33.5	8.6	75

Analysis 2.25. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 25 Mental state: 2a. Specific symptoms: negative symptoms - average endpoint score (SANS, high = poor) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 25 Mental state: 2a. Specific symptoms: negative symptoms - average endpoint score (SANS, high = poor) - by long term

Analysis 2.26. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 26 Mental state: 2b. Specific symptoms - average endpoint scores (various scales, skewed data).

Study	Intervention	Mean	SD	Total
medium term - anxiety (HADS, high = poor)
Harrison-Read-UK	1. ICM	6.5	4.9	52
Harrison-Read-UK	2. non-ICM	6.7	4.6	61
medium term - depression (HADS, high = poor)
Harrison-Read-UK	1. ICM	6.4	5.4	52
Harrison-Read-UK	2. non-ICM	6.6	4.9	61
long term - anxiety (HADS, high = poor)
Harrison-Read-UK	1. ICM	7.5	5.3	56
Harrison-Read-UK	2. non-ICM	6.4	4.6	58
long term - depression (HADS, high = poor)
Harrison-Read-UK	1. ICM	7.3	5.4	56
Harrison-Read-UK	2. non-ICM	6.8	5.6	58

Analysis 2.27. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 27 Behaviour: 1. Specific behaviour (various measurements).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 27 Behaviour: 1. Specific behaviour (various measurements)

Analysis 2.28. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 28 Quality of life: 1. Average endpoint score (various scales).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 28 Quality of life: 1. Average endpoint score (various scales)

Analysis 2.29. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 29 Participant satisfaction/need: 1a. Average endpoint scores (various scale) - by long term.

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 29 Participant satisfaction/need: 1a. Average endpoint scores (various scale) - by long term

Analysis 2.30. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 30 Participant need: 1b. Average endpoint scores (CAN, high = poor) - skeweed data).

Study	Intervention	Mean	SD	Total
by medium term
Harrison-Read-UK	1. ICM	7.3	3.7	49
Harrison-Read-UK	2. non-ICM	6.1	4	60
by long term
Harrison-Read-UK	1. ICM	6.6	3.6	54
Harrison-Read-UK	2. non-ICM	5.6	3.8	59

Analysis 2.31. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 3 1 Participant satisfaction: 1. Average endpoint scores (CSQ-modified, high = better, skewed data) - by long term.

Study	Intervention	Mean	SD	Total
REACT-UK	1. ICM	77.2	20	91
REACT-UK	2. non-ICM	70	20.6	75

Analysis 2.32. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 32 Costs: 1. Direct costs of psychiatric hospital care (skewed data).

Study	Intervention	Mean	SD	Total	Note
by medium term
Harrison-Read-UK	1. ICM*	501	967.4	97
Harrison-Read-UK	2. Low ICM*	527	753	96	* Unit cost £, fiscal year 1995/6. ** ‘No significant difference between groups. Statistical analysis on non-parametric data were performed using bootstrap methods’ ***To Marina: time period 12 months.
by long term
Harrison-Read-UK	1. ICM*	414	777.8	97
Harrison-Read-UK	2. Low ICM*	478	890	96	* Unit cost £, fiscal year 1995/6. ** ‘No significant difference between groups. Statistical analysis on non-parametric data were performed using bootstrap methods’ ***To Marina: time period 12 months.
Quinlivan-California	1. ICM*	301	396.6	30
Quinlivan-California	2. Low ICM*	959	1,572.7	30	* Unit cost US $, fiscal year not reported, but study was carried on from April 1990 to March 1992. ** ‘Costs significantly lower for the ICM group (F=4.32, df=2. 87, p=0.02.)’ ***To Marina: Time period 24 months

Analysis 2.33. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 33 Costs: 2a. Direct costs of all care - by long term (Unit cost £, fiscal year 1997/98).

Review: Intensive case management for severe mental illness

Comparison: 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT

Outcome: 33 Costs: 2a. Direct costs of all care - by long term (Unit cost , fiscal year 1997/98)

Analysis 2.34. Comparison 2 INTENSIVE CASE MANAGEMENT versus NON-INTENSIVE CASE MANAGEMENT, Outcome 34 Costs: 2b. Direct costs of all care - by medium term (skewed data).

Study	Intervention	Mean	SD	Total	Note
Johnston-Australia	1. ICM*	2,408	2,581.4	33
Johnston-Australia	2. Low ICM*	1,762	1,872	25	* Unit cost Aus $, fiscal year 1991/2. ** ‘The significance test on the cost of care per patient was performed on transformed means. No significant differences were found between groups’ ***To Marina: time period 12 months

Appendix 1. Search method Assertive Community Treatment Review

1. Electronic searches

   • 1.1 CINAHL (January 1982 to May 1997) was searched using the CSG’s terms for randomised controlled trials and the CSG’s terms

     for schizophrenia combined with the phrase:

     [and ((case or care) near management) or CPA or (Care near1 Programme near1 Approach) or (Assertive near1 Community near1 Treatment) or PACT or TCL or (Training near (community near1 living)) or (Madison near4 model)]

   • 1.2 The Cochrane Schizophrenia Group’s Register (1997) was searched using the phrase:

     [and ((case or care) and management) or CPA or (Care and Programme and Approach) or (Assertive and Community and Treatment) or PACT or TCL or (Training and (community and living)) or (Madison and model)]

     1.3 EMBASE (January 1980 to May 1997) was searched using the CSG’s terms for randomised controlled trials and the CSG’s terms for schizophrenia combined with the phrase:

     [and ((case or care) near management) or CPA or (Care near1 Programme near1 Approach) or (Assertive near1 Community near1 Treatment) or PACT or TCL or (Training near (community near1 living)) or (Madison near4 model)]

   • 1.4 MEDLINE (January 1966 toMay 1997) was searched using the CSG’s terms for randomised controlled trials and the CSG’s terms for schizophrenia combined with the phrase:

     [and ((case or care) near management) or CPA or (Care near1 Programme near1 Approach) or (Assertive near1 Community near1 Treatment) or PACT or TCL or (Training near (community near1 living)) or (Madison near4 model)]

   • 1.5 PsycLIT (January 1974 to May 1997) was searched using the CSG’s terms for randomised controlled trials and the CSG’s terms for schizophrenia combined with the phrase:

     [and ((case or care) near management) or CPA or (Care near1 Programme near1 Approach) or (Assertive near1 Community near1 Treatment) or PACT or TCL or (Training near (community near1 living)) or (Madison near4 model)]

2. Searching other resources

   • 2.1 Reference searching

     Each of the randomised controlled trial identified was sought as a citation on the SCISEARCH database. Reports of articles that had cited these studies were inspected in order to identify further trials. Reference lists of all included trials and identified reviews were scanned for evidence of trials missed by the computerised search.

     It should be noted that in electronic searches the phrase ’ACT’ is not feasible as this common word generates a very large number of false positives.

Appendix 2. Search method Case Management Review

1. Electronic searches

The Cochrane Schizophrenia Group’s Register (1997), EMBASE (January 1980 to May 1995), MEDLINE (January 1966 to May 1995), PsycLIT (January 1974 to May 1995) and CINAHL were all searched for any text containing the following phrases:

[((case or care) and (management)) or CPA or (Care Programme Approach) or (Assertive Community Treatment) or (PACT) or (TCL) or (Training near2 community living) or (Madison near model)]

Each randomised controlled trial identified by the search was sought as a citation on the SCISEARCH database.

2. Searching other resources

2.1 Hand searching

Reports of articles that had cited these studies were inspected in order to identify further trials.

Citation lists of all included trials and identified reviews were scanned for evidence of trials missed by the computerised search.

Appendix 3. Dealing with missing data - Standard deviation mean number of days per month in hospital (Intensive Case Management Protocol)

3.2.2 Standard deviation mean number of days per month in hospital

For the primary outcome, mean number of days per month in hospital, if standard deviations were not reported and could not be calculated from available data, we asked authors for additional information. In the absence of data from authors, we calculated missing standard deviations using a regression analysis of standard deviation against mean, based on data from studies which did report these data. If the standard deviations calculated according to the above technique were available from a previous review (published and unpublished data) (Burns 2007), we used these data.

Appendix 4. Dealing with missing data - Incomplete data for meta-regression (Intensive Case Management Protocol)

3.4 Incomplete data for meta-regression

In some cases we anticipate that IFACT score variables will not all be available. Where these missing data are from multi-centre studies from which we do have relevant data we assumed the missing variable score to be the mode of the available data from the other centres that we used as a reference. Where there was no clear reference centres, we tried to match the study to another we felt to be closest and used those scores. As explained above, we will only make these assumptions if we are able to directly rate over 50% of the data within each multi-centre study and overall.

In some cases we anticipate that baseline hospital use data will not all be available. Where these missing data are from multi-centre studies from which we do have relevant data we assumed the missing information to be the mean of the available data from the other centres that we used as a reference. Where there was no clear reference centres, we tried to match the study to another we felt to be closest and used those means. As explained above, we will only make these assumptions if we are able to directly rate over 50% of the data within each multi-centre study and overall.

Appendix 5. R script used to impute data and perform meta-regression analysis

AOIP <- read.csv(’META-REG·2010·fmi.csv’)

AOIP <- AOIP[1:52,]

names(AOIP)

AOIPM <- AOIP

# initially replace missing SD’s by simple regression model on means

exp <- lm(AOIP$Exp·SD ~ poly(AOIP$Exp·MEAN,2))

exp2 <- predict(exp, as.data.frame(AOIP$Exp·MEAN), type = ’response’)

# plot(AOIP$Exp·SD, AOIP$Exp·MEAN, xlim = c(0,15))

# points(exp2, AOIP$Exp·MEAN, col = ’red’)

exp3 <- AOIP$Exp·SD

exp3[is.na(exp3)] <- exp2[is.na(exp3)]

AOIP$Exp·SD <- exp3

exp <- lm(AOIP$Con·SD ~ poly(AOIP$Con·MEAN,2))

exp2 <- predict(exp, as.data.frame(AOIP$Con·MEAN), type = ’response’)

# plot(AOIP$Con·SD, AOIP$Con·MEAN, xlim = c(0,15))

# points(exp2, AOIP$Con·MEAN, col = ’red’)

exp3 <- AOIP$Con·SD

exp3[is.na(exp3)] <- exp2[is.na(exp3)]

AOIP$Con·SD <- exp3

library(mi)

AOIP2 <- cbind(AOIP[,18], AOIP[,5:11], AOIP[,13], AOIP[,16:17])

names(AOIP2)[1] <- names(AOIP)[18]

names(AOIP2)[9] <- names(AOIP)[13]

info.aoip <- mi.info(AOIP2)

# info.aoip <- update(info.aoip, “include”,

# list(trial·no=“FALSE”,cent.no=“FALSE”,trial·id=“FALSE”,centre·id=“FALSE” ))

imp.aoip <- mi(AOIP2, info.aoip, n.iter = 30)

AOIP2 <- mi.data.frame(imp.aoip)

AOIP2 <- cbind(AOIP[,1:4],AOIP2)

# recreate Fideltot

Fideltot <- AOIP2$Fid.Org + AOIP2$Fidstaff

AOIP2 <- cbind(AOIP2, Fideltot)

# set OPUS baseline mean to missing

is.na(AOIP2[,13]) <- 19

AOIP2.mi <- AOIP2

AOIP <- AOIPM

# write.csv(AOIP2, ’metaregdataset·imputed2010.csv’)

# meta analysis

library(meta)

AO1 <- metacont(AOIP$Exp·N, AOIP$Exp·MEAN, AOIP$Exp·SD, AOIP$Con·N, AOIP$Con·MEAN,

AOIP$Con·SD, AOIP$centre·id)

AO2 <- metacont(AOIP2$Exp·N, AOIP2$Exp·MEAN, AOIP2$Exp·SD, AOIP2$Con·N, AOIP2$Con·MEAN,

AOIP2$Con·SD, AOIP2$centre·id)

AO3 <- metacont(AOIP2$Exp·N, AOIP2$Exp·MEAN, AOIP2$Exp·SD, AOIP2$Con·N, AOIP2$Con·MEAN,

AOIP2$Con·SD, AOIP2$centre·id, subset = (AOIP2$SCvsLICM == 1))

AO4 <- metacont(AOIP2$Exp·N, AOIP2$Exp·MEAN, AOIP2$Exp·SD, AOIP2$Con·N, AOIP2$Con·MEAN,

AOIP2$Con·SD, AOIP2$centre·id, subset = (AOIP2$SCvsLICM == 2))

summary(AO1)

summary(AO2)

# metaregression

# Control type

AO2rct <- lm(AO2$TE ~ AOIP2$SCvsLICM, weights = AO2$w.random)

summary(AO2rct)

# Total fidelity score

AO2r1 <- lm(AO2$TE ~ AOIP2$Fideltot, weights = AO2$w.random)

summary(AO2r1)

plot(AO2$TE ~ AOIP2$Fideltot)

abline(lm(AO2$TE ~ AOIP2$Fideltot, weights = AO2$w.random))

# Interaction

AO2rint <- lm(AO2$TE ~ AOIP2$Fideltot * AOIP2$SCvsLICM, weights = AO2$w.random)

summary(AO2rint)

plot(AO2$TE ~ AOIP2$Fideltot)

abline(lm(AO2$TE ~ AOIP2$Fideltot, weights = AO2$w.random))

# Organizational fidelity score

AO2r2 <- lm(AO2$TE ~ AOIP2$Fid.Org, weights = AO2$w.random)

summary(AO2r2)

plot(AO2$TE ~ AOIP2$Fid.Org)

abline(lm(AO2$TE ~ AOIP2$Fid.Org, weights = AO2$w.random))

# more fancy plot

plot(AO1$TE ~ AOIP$Fid.Org, pch = 21, bg = ’blue’, cex = AO1$w.random*6,

ylab = ’Treatment Effect (Days)’, xlab = ’Organizational Fidelity Score’)

abline(lm(AO2$TE ~ AOIP2$Fid.Org, weights = AO2$w.random))

# against baseline

AO2r4 <- lm(AO2$TE ~ AOIP2$Baseline·Mean, weights = AO2$w.random)

summary(AO2r4)

plot(AO2$TE ~ AOIP2$Baseline·Mean)

abline(lm(AO2$TE ~ AOIP2$Baseline·Mean, weights = AO2$w.random))

# fancy plot

plot(AO1$TE ~ AOIP$Baseline·Mean, pch = 21, bg = ’blue’, cex = AO1$w.random*6,

ylab = ’Treatment Effect (Days)’, xlab = ’Baseline Days in Hospital’)

abline(lm(AO2$TE ~ AOIP2$Baseline·Mean, weights = AO2$w.random))

# multivariate model

AO2r5 <- lm(AO2$TE ~ AOIP2$Baseline·Mean + AOIP2$SCvsLICM + AOIP2$Fid.Org, weights = AO2$w.random)

summary(AO2r5)

# Thin plate spline

library(fields)

x <- cbind(AOIP2$Baseline·Mean[−19], AOIP2$Fid.Org[−19])

fit <- Tps( x, AO2$TE[-19], weights = AO2$w.random[−19], df = 10)

par(mfrow=c(2,1))

surface(fit)

plot(x, pch = 21, bg = ’red’, cex = AO2$w.random[−19]*4)

WHAT’S NEW

Last assessed as up-to-date: 24 June 2010.

Date	Event	Description
19 January 2011	Amended	Contact details updated.

HISTORY

Protocol first published: Issue 3, 2009

Review first published: Issue 10, 2010

Date	Event	Description
6 October 2010	Amended	Contact details updated.

DIFFERENCES BETWEEN PROTOCOL AND REVIEW

The following paragraph follows the numbering system of the methods of the review.

1. Types of outcome measures (see Secondary outcomes)

Some social functioning and costs outcomes have been slightly amended in the review. We reported the protocol version below.

4. Social functioning

4.1 Imprisonment (i.e. police contacts & arrests)

9. Economic

9.1 Costs of psychiatric hospital care

9.2 Costs of health care

9.3 Costs of all care

2. Compliance with medication (see Secondary outcomes)

This secondary outcome had been listed twice in the protocol due to inattention (both as Global state and Behaviour). We amended it, listing Compliance with medication just under Global state outcome.

3. Data regarding criteria and outcomes (see Data extraction and management)

This section was further clarified in the review. We reported the protocol version below.

1. Extraction

1.1 Data regarding criteria and outcomes

Authors MD and CBI independently extracted data fromincluded studies.Again, any disagreementwas discussed, decisions documented and, if necessary, authors of studies were contacted for clarification. With remaining problemsMMhelped clarify issues and those final decisions were documented. Data presented only in graphs and figures were extracted whenever possible, but were included only if two reviewers independently had the same result. Attempts were made to contact authors through an open-ended request in order to obtain missing information or for clarification whenever necessary. Where possible, we extracted data relevant to each component centre of multi-centre studies separately.

4. Skewed data (see Data extraction and management)

The following clarification had not been anticipated in the protocol, but added in the review.

5.2 Specific - mean number of days in hospital

We implemented one exception to the above rule (5.1) in order to present more data, recognising that this is a post hoc decision, but also that the rules as regards management of skewed data and how robust skewed data are within meta-analysis is unclear (Higgins 2008). Where mean number of days in hospital data were skewed, and they were provided by studies of less than 200 participants, we nevertheless entered those data into a sub-group of the overall analysis. We also presented the overall effect from all data pooled.

5. Summary of findings table (see Data extraction and management, 9.)

Outcomes included in the Summary of findings were slightly rearranged in the way they were listed and presented, but they were not substantially modified. We reported the protocol version below. Besides, we decided not to apply the low and high control risk calculation as stated in the protocol, and reported below.

9. Summary of findings table

We anticipate including the following long term main outcomes in a summary of findings table.

1. Global state

1.1 Hospitalisation: mean number of days per month in hospital

1.2 Relapse

1.3 Leaving the study early (lost to follow up)

2. Service use

2.1 Hospital admission across time

3. Adverse effect

3.1 Death - suicide

4. Social functioning

4.1 Employment - unemployed at end of study

5. Mental state: general symptoms

5.1. Not improved to a clinically meaningful extent (as defined in trial)

Within the Summary of findings table we assumed for calculation of the low risk groups that the lowest control risk applied to all data. We did the same for the assumption of the highest risk groups.

6. Assessment of risk of bias in included studies (see Assessment of risk of bias in included studies)

This section was further clarified in the review. We reported the protocol version below.

Again working independently, MD and CBI assessed risk of bias using the tool described in the Cochrane Collaboration Handbook (Higgins 2008). This tool encourages consideration of how the sequence was generated, how allocation was concealed, the integrity of blinding at outcome, the completeness of outcome data, selective reporting and other biases. We would have excluded studies where allocation was clearly not concealed.

7. Standard deviation mean number of days per month in hospital (see Dealing with missing data, 3.2.2)

Substancially amended. For original protocol version please see Appendix 3.

8. Incomplete data for meta-regression (see Dealing with missing data, 3.4)

Substancially amended. For original protocol version please see Appendix 4.

9. Statistical heterogeneity (see Assessment of heterogeneity, 3.2)

The last paragraph was further clarified in the review. We reported the protocol version below.

3.2 Employing the I² statistic

Heterogeneity between studies was investigated by considering the I² method alongside the Chi2 ’p’ value. The I² provides an estimate of the percentage of inconsistency thought to be due to chance (Higgins 2003). The importance of the observed value of I² depends on i. magnitude and direction of effects and ii. strength of evidence for heterogeneity (e.g. ’p’ value from Chi2 test, or a confidence interval for I²).

I² estimate greater than or equal to 50%accompanied by a statistically significant Chi2 statistic, was interpreted as evidence of substantial levels of heterogeneity (Section 9.5.2 - Higgins 2008) and reasons for heterogeneity were explored. If the inconsistency was high and the clear reasons were found, data were presented separately.

10. Subgroup analyses (see Subgroup analysis and investigation of heterogeneity, 1.)

No sub-group analyses had been anticipated in the protocol.

11. Subgroup analyses (see Subgroup analysis and investigation of heterogeneity, 2.)

This section was further clarified in the review. We reported the protocol version below.

2. Investigation of heterogeneity

2.1 Unanticipated heterogeneity

Should unanticipated clinical or methodological heterogeneity be obvious we will simply state hypotheses regarding these for future reviews or versions of this review. We do not anticipate undertaking analyses relating to these.

12. Sensitivity Analysis (see Sensitivity analysis, 4.)

This section was substantially amended in the review. We reported the protocol version below.

4. Assumptions for incomplete data for meta-regression

Where assumptions had to be made regarding missing trial data for meta-regression (see Dealing with missing data) we compared the findings of the meta-regression on primary outcome when we used our assumption compared with completer data only. A sensitivity analysis was undertaken testing how prone result from meta-regression were to change when ’completed’ data only were compared to the imputed data using the above assumption. If there was a substantial difference, then only completed data were employed.