Open Access

Background Extensive research has been undertaken to predict treatment response (TR) to antipsychotics. Most studies address TR to antipsychotics in general and as monotherapy, however, it is unknown whether patients might respond favourably to a combination of antipsychotics. Aims This study aimed to identify differential predictors for TR to monotherapy with amisulpride or olanzapine compared to a combination of antipsychotics. Methods Post-hoc analysis was conducted of data collected from the COMBINE-study, a double-blind, randomized, controlled trial. Demographic and disease-related measures were gathered at baseline to predict TR after eight weeks defined by the Positive and Negative Syndrome Scale. Missing values were accounted for by a random replacement procedure. Attrition effects and multicollinearity were analysed and sets of logistic regression models were calculated for different treatment groups. Results Of the 321 randomized patients, 201 completed procedures until week eight and 197 were included in the analyses. For all treatment groups, early TR after two weeks and high subjective well-being under antipsychotics at baseline were robust predictors for TR. The propensity for early side effects also indicated a higher risk of later non-response. Specific parameter estimates were rather similar between treatment groups. Conclusion Early TR, drug-related subjective well-being, and early side effect propensity evolved as predictors for later TR whether to monotherapy or combination strategy. Accordingly, due to a lack of differential predictors, early and close monitoring of targeted and unwanted effects is indicated to guide respective treatment decisions.