What should prompt an urgent referral to a community mental health team?

Abstract

Background There is often little guidance to advise general practitioners on whether a referral to a community mental health team should be classified as ‘urgent’ or not.

Aims (1) To identify the proportion and appropriateness of referrals considered urgent by the referrer; (2) To develop a set of criteria to guide what should constitute an ‘urgent’ referral.

Methods One hundred consecutive referral letters to a community mental health team were analysed to determine the proportion that were considered urgent by the referrer compared to a consensus panel of psychiatrists. A Delphi group was then used to develop a set of criteria to guide referrers as to what should be regarded as an urgent referral.

Results Thirty-three percent of referrals were deemed urgent by the referrer, compared to 17% by the psychiatric consensus panel, with little agreement between the two (kappa = 0.021, P = 0.013). Referrals that were made using a single assessment process (SAP) form were significantly more likely to be inappropriately marked as being urgent (P < 0.001). A set of 12 criteria was developed using the Delphi technique.

Conclusions There was significant disagreement between the referrers and the assessing team as to which referrals required urgent attention. The findings justified the creation of guidelines, and this paper outlines a set of 12 criteria to guide what should prompt an urgent referral.

Introduction

In the UK, referrals to mental health services are made to a community mental health team (CMHT). This is a multidisciplinary team composed of doctors, psychologists, occupational therapists, social workers and nurses. The response to any referral to secondary care is dependent upon the accuracy and comprehensiveness of information supplied by the referrer. A study by Burbach found that only 19% of referrals mentioned the degree of urgency.¹ Furthermore, 25% of referrals were felt to underestimate the severity of symptoms. Cubbin et al found that the agreement with regard to urgency between a consultant psychiatrist and referrer varied depending on who was making the referral.² In another study, referrers were asked to complete a rating of mental health problem severity as part of the referral process. This was completed in only 25% of referrals and led to no significant improvement in appropriateness of referrals or in the receiving team's ability to identify urgent referrals.³ We observed frequent incongruity between the degree of urgency described by the referrer and that considered by the CMHT.

Our literature search failed to identify any guidance on what should constitute an urgent referral to a CMHT. Receiving an excessive number of referrals marked as being urgent renders the CMHT less able to distinguish patients truly requiring urgent attention. We were also concerned that uncertainty about the appropriateness or stated urgency of referrals could lead to mutual erosion of trust.

Referrals to our service either come in the form of a letter or a standardised referral proforma based on the single assessment process (SAP). The single assessment process was introduced as part of the National Service Framework for Older People, in an attempt to streamline communication between health and social care agencies,⁴ but its utility in psychiatry has been challenged.⁵ The SAP form introduced locally in 2004 has several headings including one requiring an assessment of urgency. The form contains the question ‘do you feel this requires urgent attention?’ and a space for the referrer to tick ‘yes’ or ‘no’. If the referrer considers the referral urgent they are then asked to state a reason. In common with many CMHTs we accept referrals from any health or social care professional – not just the general practitioner (GP). In this service, ‘urgent’ indicates that some action (usually a home visit) is required within 24 hours of receipt of referral.

This study had two main aims;

1. to identify the proportion and appropriateness of referrals to our CMHT considered urgent by the referrer

2. to develop criteria to guide what should constitute an urgent referral.

Methods

One hundred consecutive referrals to the older adults CMHT between November 2006 and June 2007 were examined and data collected on whether each referral was marked as being urgent and whether a reason for this was given. A consensus was reached by a minimum of two members of the research panel as to whether or not they considered the referral to be urgent, based on information supplied by the referrer. The research panel members were blinded to whether the referrer had marked the referral as requiring urgent attention. In cases of disagreement, a third member of the research team was consulted, and a majority decision taken. Additional information was collected including demographic data, the background of the referrer (e.g. GP or social worker) and whether or not the referral was made on a SAP form. Data were pseudonomysed and were analysed using SPSS 15.0 using Chi-squared tests, Fisher's exact test and kappa coefficient as appropriate.

The Delphi technique is an iterative process to collate and refine the opinions of experts with a view to developing a consensus.⁶ We used this technique to develop a set of criteria on what should constitute an urgent referral. A group of 12 stakeholders were selected to represent those involved in the referral and assessment process. Stakeholders included three GPs, two consultant psychiatrists, one community psychiatric nurse, one Admiral nurse (who provides support and advocacy for carers of people with dementia), two social workers, one social work manager, one day hospital manager and one psychologist. The stakeholders were unaware of the results of the first part of the study.

Communication took place via email. Two specialist registrars (senior trainees) in psychiatry facilitated the responses received. Each participant was initially asked to ‘brainstorm’ ideas as to what they felt should constitute an urgent referral. Following this, the criteria were collated maintaining respondent anonymity, and the list was circulated to all Delphi group members, requesting them to refine each idea, commenting on the strengths and weakness, and identify new ideas. Iterations of the above process were performed until no new ideas emerged, and until broad themes had been refined by the panellists into more specific criteria. Resolution occurred by the nominal group technique of voting, i.e. each member was asked to vote for the top five ideas, rating the most relevant urgent criterion as five. The results were tallied and rank ordered, based on the evaluation.

Results

Evaluation of urgency

Box 1 shows demographic data for the sample. Of the 100 referrals examined, 33 were classified as urgent by the referrer, and 17 were felt to be urgent by the consensus panel (see Table 1). There was a low level of agreement between the referrer and consensus panel as to which referrals were urgent (kappa = 0.21, P = 0.013).

Box 1 Demographic data for the sample of 100 referrals

Sex

41% male, 59% female

Known to services

5% known to services, 95% new referrals

Residency

89% live independently, 8% in residential homes, 2% sheltered accommodation, 1% of no fixed abode

Referrers

64% GPs, 20% hospital doctors, 10% social workers, 2% community matrons, 1% duty workers, 1% Admiral nurses, 1% liaison psychiatrists, 1% psychologists

SAP

40% on SAP forms, 60% referral letters

Table 1.

Comparison of referrals marked as being urgent with those felt to be urgent by the consensus panel

View of referrer	View of consensus panel
	Urgent	Not urgent
Urgent	10	23
Not urgent	7	60

Of the 40 referrals made on SAP forms, 24 (60%) were marked urgent, as opposed to nine (15%) referral letters (P < 0.001). The referrer and consensus panel disagreed regarding urgency for 17 (42.5%) referrals made on SAP forms, compared with 13 (21.7%) of referral letters (P = 0.026). Table 2 shows where these disagreements arose. Twenty-two (66%) of the referrals marked as being urgent explicitly stated a reason.

Table 2.

Agreement of consensus panel and referrer against whether or not a SAP form was used

	Agreed not to be urgent	Felt to be urgent by the referrer but not the consensus panel	Felt to be urgent by the consensus panel but not the referrer	Agreed to be urgent
SAP form	15	16	1	8
Referral letter	45	7	6	2

Delphi group

All 12 panellists (100%) in the Delphi group completed round one and provided a total of 65 criteria. The results were collated and after elimination of duplicate responses 35 suggested criteria remained. Six of the 12 panellists (50%) completed round two, and 11 of 12 panellists (91.7%) completed round three. At the end of three rounds, 12 urgent criteria were rated from 1–5 by the panellists. Box 2 shows the final results of the Delphi process.

Box 2 Results of the Delphi process

Delphi criteria for an urgent referral to a community mental health service

1. Reason to believe that there is a risk of significant harm (immediate and serious risk to self and/or others) occurring as a result of a mental disorder, or suspected mental disorder, if no action is taken before the next community referrals allocation meeting, i.e. within one week

2. Level of cognitive functioning that places them at immediate risk of harm to self or others including carers (e.g. due to wandering, self-neglect, abuse) and physical/medical causation has been excluded

3. Patient will have been ill enough to have seen by the ‘referrer’ who has knowledge of the patient and considers the risk for urgency to be linked with a perceived mental health problem

4. Support network not coping or sudden loss of social support network resulting in risk to patient

5. Request to provide a medical assessment for a Mental Health Act assessment

6. Patient is at risk of abuse/vulnerable and is in a location that is not safe

7. Delirium where no cause is found in order to rule out an acute mental illness

8. Patient needs to be seen within a specified time frame with good ‘clinical’ reasons (e.g. not because this is causing inconvenience to other services). However, even if they are causing an inconvenience, it may be indicative of ‘something going on’ which may require urgent review

9. Referral from acute hospital where there is a capacity issue around consenting to urgent medical intervention, where the Responsible Medical Officer has been unable to assess capacity because of mental health complications.

10. Assess urgently if challenging behaviour might improve with immediate treatment/medication. Urgency may vary depending on the nature of the behaviour, i.e. physical or antisocial

11. Not eating or drinking. Level of urgency may depend on duration and what attempts have been made to address this before

12. Assessment requested as part of a Safeguarding Adult enquiry (Protection of the Vulnerable Adult) where urgent protection of vulnerable adult is needed. Level of urgency may depend upon nature of risk and need for a protection plan

Discussion

This study demonstrates a low level of agreement between the psychiatric team and referrers as to what should constitute an urgent referral. A high proportion of referrals were felt to be inappropriately marked as being urgent, making it more difficult to triage referrals and to respond appropriately. Perhaps more worryingly there was also a significant proportion of referrals that were not marked as being urgent but which the psychiatric team felt required urgent attention. If a referral is marked as being urgent it will be reviewed by a senior member of the team on the day of receipt and a decision taken on what action is needed, but non-urgent referrals wait to be discussed at the weekly CMHT meeting. Failure to identify cases requiring urgent assessment and to communicate this to the CMHT may put patients at risk.

Referrals were more likely to be marked as being urgent, and were more likely to be felt to be inappropriately urgent, if made on a SAP form. Conversely, referrals were more likely to be deemed urgent by the team but not marked urgent if made via a referral letter. The presence of a ‘tick box’ identifying urgency on the SAP form may serve to remind referrers to provide this important information, but appears also to lead to indiscriminate use.

The findings above justified the creation of a set of criteria to guide referrers. To our knowledge this paper presents the first set of guidelines on what should prompt an urgent referral to a CMHT. The criteria established by our study have been further validated in a workshop involving the research group and a group of local GPs.

The 12 criteria established by the Delphi group covered a wide range of perspectives, and illuminate the thought processes that influence both the referrer and assessor. Risk emerged as the factor that most drives urgency, more so than some other important elements, for instance patient suffering or strain on carers. That the patient should have been seen and personally assessed by the referrer ranked highly (many of the referrals we receive are referred following requests by a third party and have not been seen by the referrer), and it is likely that referrals where the patient has not been seen by the referrer will receive less urgent attention. During the process it became apparent that many factors influence the referrer's assessment of urgency. These include experience and general approach, the use of paradigms and guidelines and the relationship between the referrer and the service being referred to. Previous research has found that referrals may be influenced not just by clinical urgency but by the personal threshold and perceived personal level of competence of the referrer.⁷

The observational component of the study examined a large number of unselected case notes. It was necessary to search for 202 consecutive referrals in order to locate 100 sets of notes, and it may be that the low yield has introduced an element of bias to the study. It is possible that the notes we were unable to find were for patients requiring long-term follow-up, and so this study may in fact underestimate the number of urgent referrals. The research team was blinded to whether the referral was marked as requiring urgent attention, but there were occasions when a member of the team had prior knowledge of the patient.

The Delphi technique is a well-recognised and widely used process that has been employed successfully to create clinical guidelines in other specialties.^8,9 The use of the Delphi group to formulate guidelines allowed the sharing of ideas and decision making by a large number of stakeholders who were geographically distanced, and who had diverse training, backgrounds and roles in the referral process. The anonymising of opinions and use of independent facilitators allowed a transparent and democratic decision to be reached, where everyone's opinion carried equal weight.

In line with previous research, this study identified confusion with regard to which referrals should warrant urgent attention, and significant levels of disagreement between the referrers and the receiving team.^1,2

In a previous study by Ferriter et al, the SAP form was assessed as providing less information than the traditional referral letter.⁵ Perceived difficulty in referring to CMHTs may lead to ‘gaming’ and make referrers more likely to tick an urgent box in order that a referral will be accepted.⁷ Although our results suggest that the presence of an ‘urgent’ tick box encourages referrers to make a larger number of inappropriately urgent referrals, it does also mean that genuine urgency is more likely to be communicated, and so acted upon appropriately by the CMHT. In this case the potential dangers associated with a low sensitivity may mean that it is acceptable to compromise on specificity.

This paper identifies a lack of consensus between referrers and members of the assessing team as to which referrals should be marked as urgent, and presents a set of guidelines that clinicians can refer to in order to help them to decide whether or not a referral should be urgent. It is anticipated that the use of these will improve the standard of referrals and help improve communication between referrers and CMHTs. Further research is required to validate these criteria in other CMHTs and to confirm their utility in clinical practice.