The effect on relapse of Culturally-adapted Family Intervention (CaFI) compared to usual care among African & Caribbean people diagnosed with psychosis in the UK: a Randomised Controlled Trial

AIMS: i) Evaluate CaFI’s clinical and cost-effectiveness in Sub Saharan African and Caribbean populations. ii) Determine how to maximise facilitators/overcome barriers to successful implementation. DESIGN: Blind-rated RCT with internal Stop/Go pilot and process evaluation. SETTING: 6 NHS Mental Health Trusts: 3 Northwest, 2 Midlands, 1 London, 1 South West. POPULATION: African and Caribbean origin service users (including ‘Black British’ and ‘Mixed’) in inpatient, rehabilitation, community, and forensic settings. HEALTH TECHNOLOGY: 10 x 1 hour sessions of Culturally-adapted Family Intervention (CaFI) delivered within a 20 week ‘therapy window’. PRIMARY OUTCOME: Reduction in relapse (as defined by an exacerbation of symptoms lasting over two weeks(7)). SECONDARY OUTCOMES: Number of psychiatric and compulsory admissions, length of inpatient stays, service use, Perceived Coercion(8), Positive and Negative Syndrome Scale (PANSS)(9), Personal and Social Performance Scale (PSP)(10), Brief Illness Perception Questionnaire (Brief-IPQ)(11), Knowledge about Psychosis Interview (KAPI)(12), General Health Questionnaire (GHQ-12)(13), EQ-5D-5L(14), Working Alliance Inventory (WAI)(15), Service Engagement Scale (SES)(16). SAMPLE SIZE: Meta-analyses show a 0.55 - relative risk of relapse after family intervention without cultural adaptation(17); 40% of controls per year relapsed in the comparison arm. Reducing relapse risk from 40% during follow-up to 24% equates to a clinically-significant difference. Recruiting 386 participants across 6 NHS Trusts (193 in each arm) gives 90% power. Randomisation will be stratified by location and ethnicity. Within each stratum, participants will be randomly allocated (1:1) to the intervention or control arms in blocks of size 4, 6 or 8 with size randomly chosen. STATISTICAL ANALYSES: Will use an intention-to-treat approach. Log-rank tests will compare survival distributions of the two arms. If assumptions are met, Cox’s proportional hazards model will be fitted to adjust for covariates. ECONOMIC EVALUATION: Cost effectiveness acceptability analysis of bootstrapped net costs and QALYs to estimate the probability that CaFI is cost effective at 12 months. Service use data, collected from all participants, will be used to estimate costs from an NHS & Social Care perspective. Regression analysis will estimate net costs and QALYs, accounting for missing data and covariates. Sensitivity analyses will assess broader perspectives for cost, choice of outcome measure and impact of missing data. PROCESS EVALUATION: Semi-structured interviews with 30 staff (e.g. therapists, care coordinators, NHS senior leaders and service managers, commissioners) will explore potential implementation barriers and facilitators. We will purposively sample participants across all sites. Interviews will be analysed using Framework Analysis. We will assess fidelity to the therapy manual (= 10% of sessions), therapists’ ability to deal with culturally-specific issues and number of sessions completed. TIMETABLE: 54 months: 12m study set-up (including acceptability testing with Africans); 22m total recruitment period; 12m Stop/Go pilot; 12m follow-up per participant; 3m closedown, analysis and write-up (See Gantt chart).