Miller 2004

Methods	Methodological quality: low 1. Randomisation method: adequate 2. Dropouts: numbers, and reasons in each group were described: intention‐to‐treat analysis performed 3. Outcome assessment: a. Assessors: patients are not blinded to treatment status, no self‐report b. Measures: clearly defined and valid c. Follow‐up duration: appropriate ‐ at least 6 months after assessment 4. Baseline characteristics: reported, and comparable 5. Entry criteria: clearly defined
Participants	92 participants (aged 18‐65) with current bipolar I disorder, recruited in a university‐affiliated psychiatric hospital. Exclusions: no DSM‐III‐R alcohol or drug dependence within 12 months of enrolment. One psychiatrist, one therapist with a master's degree in social work and extensive clinical experience, and two psychotherapists. Treatments were monitored and evaluated at weekly meetings attended by all of the study clinicians and investigators.
Interventions	RCT to compare three conditions. T1(n:33) = family therapy + pharmacotherapy ‐ with family therapy involving 6‐10 sessions of 50 minutes of Problem Centered Systems Therapy of the Family (focus on problem solving, communication roles, affective responsiveness, affective involvement, and behaviour control). T2(n:30) = multifamily psychoeducational group therapy + pharmacotherapy ‐ 6 sessions of 90 minutes. T3(n:29) = pharmacotherapy ‐ weekly medication management sessions in the first month. As patients improved, appointments were scheduled less frequently. A mood stabiliser was prescibed to each subject, other medications were used as well, based on the type, intensity, and duration of symptoms.
Outcomes	1. Measures a. Structured Clinical Instrument for DSM‐III‐R‐Patient version b. Modified Hamilton Rating Scale Depression (MHRSD) c. Bech/Rafaelsen Mania Scale Recovery was defined as two consecutive months with Bech/Rafaelsen scores<6 and Hamilton scores<7. 2. Follow‐up times Assessments at intake and monthly thereafter. The proportion of subjects within T1, T2 and T3 who recovered by month 28 did not significantly differ. Time to receovery did not differ significantly between the treatment groups. Neither adjunctive family therapy nor adjunctive multifamily group therapy significantly improves recovery from bipolar I mood episodes, compared to pharmacotherapy alone.
Notes	Non‐compliance: T1=21/33, T2=20/30, T3=19/29 Withdrawal: T1= 3, T3=2 No integrity check
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear