Malakouti 2015.

Study characteristics
Methods	Study design: RCT, single‐blind (outcome assessors were blinded to participant randomisation)   Duration of study: 12 months. Participants were recruited between December 2007 and March 2008 and were followed up until 12 months post baseline
Participants	Country: Iran  Income classification: lower‐middle income between 2007 and 2008; upper‐middle income in 2009 Geographical scope: urban  Healthcare setting: participants’ homes  Mental health condition: schizophrenia or bipolar disorder  Population (patient, carer)  1. Age: 18 and older  2. Gender: both  3. Socioeconomic background: more than 50% of participants were unemployed  4. Inclusion criteria  Patients  (1) Diagnosis of bipolar or schizophrenia spectrum disorder (2) Hospitalised at least twice in the past 2 years  (3) Poor compliance  5. Exclusion criteria  Patients  (1) Acute phase of illness (2) Mental retardation (3) Organic brain problem (4) Addiction to psychoactive substances (5) Receiving the same services contemporarily from other sources
Interventions	Stated purpose: to compare the effectiveness and cost‐effectiveness of home visit clinical case management services provided by trained nurses and general practitioners (GPs) compared to usual treatment   INTERVENTION 1 (n = 46) Name: home visit by GP  Delivered by: PHP Title/name of PW and number: general practitioners (doctors) ‐ 3  1. Selection: after training, written and oral evaluation was performed, and 3 out of 5 trainees were selected  2. Educational background: medical degree  3. Training: trained in multiple courses for 24 hours a. Manual provided b. Trainer not specified     4. Supervision: every 2 weeks for first 3 months, then once every month, by main investigator (psychiatrist)  Intervention details: case management services and GP care   1. Duration/frequency: monthly home visits, 45 minutes per session, for 1 year   2. Content of intervention (by types of health workers and per patients/carers): check for symptoms and drug side effects, educate patients and family members, inject depot antipsychotic if prescribed, serve as point of contact for patients or family in case of emergency, inform psychiatrist or study supervisors when necessary or during emergency. Psychotherapeutic (new or existing) medicine could be prescribed   INTERVENTION 2 (n = 52)   Name: home visit by nurse  Delivered by: nurse  Title/name of PHW/CW and number: nurses ‐ 3  1. Selection: after training, written and oral evaluation was performed and 3 out of 6 trainees were selected  2. Educational background: Bachelor’s degree  3. Training: trained for 33 hours; trainer not specified  4. Supervision: every 2 weeks for first 3 months, then once every month, by main investigator (psychiatrist)  Intervention details: case management  1. Duration/frequency: monthly home visits, 45 minutes per session, for 1 year   2. Content of intervention (by types of health workers and per patients/carers): check for symptoms and drug side effects, educate patients and family members, inject depot antipsychotic if prescribed, serve as point of contact for patients or family in case of emergency, inform psychiatrist or study supervisors when necessary or during emergency   CONTROL (n = 54) Referred to outpatient clinic; during exacerbations, referred to psychiatric hospitals or wards for admission  CO‐INTERVENTIONS: nil
Outcomes	Patients  1. Young Mania Rating Scale  2. PANSS  3. Kohlman Evaluation of Living Skills (KELZ)  4. QALY (calculated from SF‐36) 5. Hospital admission during past 12 months 6. Client questionnaire satisfaction#  Carers  1. GHQ‐28  2. Caregiver burden 3. Knowledge questionnaire#  Process/health workers Nil  Economic outcomes  1. Cost of home visit  2. Total costs of study participants  3. Cost‐effectiveness ratio   4. ICER (Table 4 and 7th page of article, first 2 paragraphs) 5. Mean difference in utility score# 6. Incremental mean cost# 7. Incremental effectiveness#  (asterisk for study's primary outcomes; star: outcomes that we have not reported in this review)  Time point post intervention: 0 months
Notes	Source of funding: Shahid Beheshti University, Research of University of Welfare and Rehabilitation Sciences  Notes on validation of instruments: validated  Additional information: declarations of interest ‐ study authors declared no conflicts of interest  Handling the data: nil  Prospective trial registration number: IRCT138807251959N3
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "who attended selective hospitals according to the study criteria and were randomly allocated into three groups by block randomization method and using random..."
Allocation concealment (selection bias)	Low risk	Quote: "the study instruments were completed by trained psychiatric residents or clinical psychologists with master’s degree who were blind to the randomization of the participants before and after the intervention" Judgement comment: randomisation was performed at the start of the study and was concealed from outcome assessors
Blinding of participants and personnel (performance bias) All outcomes	Low risk	No blinding due to nature of the intervention; unlikely to have influenced outcomes
Blinding of outcome assessment (detection bias) all outcomes	Low risk	Study instruments were completed by trained psychiatric residents or clinical psychologists with master’s degree who were blind to randomisation of participants before and after the intervention
Baseline outcome measurements similar	Low risk	Baseline outcome measures were similar in all 3 groups
Baseline characteristics similar?	Low risk	The 3 groups were similar in all demographic features (Tables 1 and 2), but the educational level of participants in GP and nurse groups was higher than that of participants in the control group
Incomplete outcome data (attrition bias) Efficacy data	High risk	20/66 = 30.3% in the GP group were lost to follow‐up; this may have impacted the results
Incomplete outcome data (attrition bias) Safety data (e.g. adverse events)	High risk	Significant rate of attrition in the GP group may have influenced hospital admission data
Protection against contamination	Low risk	Judgement comment: intervention was delivered at home. Control group received usual care; therefore unlikely to have communicated with nurses or GPs
Selective reporting (reporting bias)	Low risk	All outcomes described in trial registry (IRCT138807251959N3) are reported in the paper
Other bias	Low risk	No other risks of bias present