Evidence-based psychosocial treatments for schizophrenia founded on Western belief systems and values may not be efficacious in different cultures without adaptation. This systematic review analyses the nature and outcomes of culturally-adapted psychosocial interventions in schizophrenia, examining how interventions have been adapted, their efficacy and what features drive heterogeneity in outcome.

Articles identified by searching electronic databases from inception to 3 March 2016, reference lists and previous reviews were independently screened by two authors for eligible controlled trials. Data on the nature of adaptations was analysed inductively using thematic analyses. Meta-analyses were conducted using random effects models to calculate effect sizes (Hedges’ g) for symptoms.

Forty-six studies with 7828 participants were included, seven adapted for minority populations. Cultural adaptations were grouped into nine themes: language, concepts and illness models, family, communication, content, cultural norms and practices, context and delivery, therapeutic alliance, and treatment goals. Meta-analyses showed significant post-treatment effects in favour of adapted interventions for total symptom severity (n = 2345, g: −0.23, 95% confidence interval (CI) −0.36 to −0.09), positive (n = 1152, g: −0.56, 95% CI −0.86 to −0.26), negative (n = 855, g: −0.39, 95% CI −0.63 to −0.15), and general (n = 525, g: −0.75, CI −1.21 to −0.29) symptoms.

The adaptation process can be described within a framework that serves as a benchmark for development or assessment of future adaptations. Culturally adapted interventions were more efficacious than usual treatment in proportion to the degree of adaptation. There is insufficient evidence to show that adapted interventions are better than non-adapted interventions. Features of context, intervention and design influenced efficacy. Investigating whether adaptation improves efficacy, most importantly amongst ethnic minorities, requires better designed trials with comparisons against unadapted interventions.