Background: Patients with primary psychosis are the most common profile in inpatient psychiatry in Sweden. Treatment guidelines recommend that this patient group should have access to both pharmacological and psychological treatments. However, psychological treatments are not routinely offered in inpatient psychiatry in Sweden. There is also a lack of research on psychological treatments for this vulnerable population in the inpatient setting. Aim: the first aim was to explore hypothetical mediators to evaluate how change in them related to change in outcomes valued by the patient over time. Method: three patients with primary psychosis were treated with 2-4 sessions of a brief contextual-behavioral intervention. The study was a single case experimental design with multiple baselines across subjects. Data were analyzed using visual inspection, non-overlap Tau-U analysis and cross-lag correlation analysis. The Personal Questionnaire (PQ) was used as primary outcome, symptom preoccupation and symtom believability as proposed mediators. Results: all participants improved significantly on PQ and symptom preoccupation measures. All participants except one improved significantly on symptom believability measures. Cross-lag correlation analyses showed that change in proposed mediators preceded changes in the primary outcome or changed concomitantly. Changes ocurred at different rates and patterns for different individuals. Conclusions: 2-4 sessions of brief contexual-behavioral therapy for inpatients with psychosis seem to be effective in terms of outcomes valued by the patient. The results support the hypothesis that the intervention is effective by targeting symptom preoccupation and symptom believability.