BackgroundPoor social skills are a core characteristic of schizophrenia and are highly associated with the progression of negative symptoms. While positive symptoms have a good response to antipsychotics, the treatment of negative symptoms remains an unmet need.MethodsA randomized controlled trial to assess the efficacy of a 20-week social skills training (SST) program for the improvement of negative symptoms in patients with treatment-resistant schizophrenia (TRS) with predominantly negative persistent symptoms, with a score > 4 (moderate) in at least 3 items of the Negative Symptom Factor Score (NSFS) (blunted affect, emotional withdrawal, poor rapport, passive social withdrawal, lack of spontaneity, motor retardation and active social avoidance).Each session lasted 60 minutes and included 6 to 9 participants. The SST sessions were conducted by trained psychologists, following topics previously outlined in a manual, and role-playing activities. The non-directive control group was conducted by nurses specialists in mental health, with the same duration but without role-playing activities. Control groups’ therapists were instructed not to give directions to the patients but to listen and redirect questions to the group.TRS was defined as the persistence of psychotic symptoms after at least two adequate trials with two different antipsychotics, All patients were taking clozapine. Blinded raters evaluated the patients at baseline, 20 weeks and after 6 months follow-up by the Positive and Negative Syndrome Scale (PANSS), Clinical Global Impression (CGI) and the Social Skills Inventory (SSI). Longitudinal comparisons between groups were carried out using a linear mixed-effects model at pre, post and follow-up to assess differences between groups. Cohen’s d effect size was computed at each time point. Primary outcome measure was the PANSS negative subscale score.Social skill were assessed with the Social Skills Inventory.Results62 patients were randomized to SST (N=29) and control group (N=33). Subjects were predominantly male (70.96%) and single (88.70%). At baseline, groups showed no differences in terms of demographic variables and illness duration. Patients were moderately ill, with mean CGI = 4.10 (SD 0.78) in SST group and 4.34 (0.90) in the control group, and had a high baseline PANSS score, with a mean total PANSS 71.90 (10.83) in the SST group and 70.4 (13.8) in the control group. The mean PANSS negative subscale score was 25.48 (4.56) in SST group and 25.13 (6.34) in control group; in the SST group 28 patients completed the 20-week intervention and 24 were assessed after 6 months; in the control group, 18 completed the treatment and 16 were assessed at the follow-up. After the 20-week intervention period, the PANSS negative scores were 24.57 (4.92) in the SST group and 22.67 (6.59) in the control group. At the follow-up, the negative score was 23.92 (5.85) in the SST group and 22.97 (5.32) in the control. There was no improvement at any timepoint (p= 0.162) or any difference between groups (p= 0.864). Patients remained clinically stable during the study. The only symptom which showed a significant improvement was the control of aggressiveness (Cohen’s d at week 20 = 0.445 [IC 95% -0.140; 1.030]), which was maintained at follow-up (Cohen’s d = 0.682 [IC 95% 0.037; 1.327]).DiscussionThe study has limitations: the SSI was not designed to assess social skills in patients with psychosis, and we have not assessed the adherence to pharmacological treatment. Our findings suggest that SST is not effective to improve negative symptoms in patients with TRS with predominantly negative symptoms, but may be potentially useful for the control of aggressiveness in situations such as criticism and mockery, which frequently occur in social interactions.