The predictive power of expressed emotion and its components in relapse of schizophrenia: a meta-analysis and meta-regression

Background Schizophrenia is a longstanding condition and most patients experience multiple relapse in the course of the condition. High expressed emotion (HEE) has been found to be a predictor of relapse. This meta-analysis and meta-regression examined the association of global EE and relapse specifically focusing on timing of relapse and EE domains. Methods Random-effects model was used to pool the effect estimates. Multiple random-effects meta-regression was used to compute the moderator analysis. Putative effect moderators including culture, EE measurements, age, length of condition and study quality were included. Results Thirty-three prospective cohort studies comprising 2284 patients were included in the descriptive review and 30 studies were included for meta-analysis and meta-regression. Findings revealed that global HEE significantly predicted more on early relapse (<= 12 months) [OR 4.87 (95% CI 3.22-7.36)] than that on late relapse (>12 months) [OR 2.13 (95% CI 1.36-3.35)]. Higher level of critical comments (CC) significantly predicted relapse [OR 2.22 (95% CI 1.16-4.26)], whereas higher level of warmth significantly protected patients from relapse [OR 0.35 (95% CI 0.15-0.85)]. None of the moderators included significantly change the results. Conclusions These findings indicate that there is a dynamic interaction between EE-relapse association with time, and CC and warmth are the two important EE domains to influence relapse among patients with schizophrenia. Results also confirmed the foci of family interventions on reducing CC and improving warmth in relationship.