Predictors of Early and Late Treatment Dropout in Early Course Schizophrenia during a Multi-Site Randomized Trial of Cognitive Enhancement Therapy

Background: Engagement in psychosocial interventions during the early course of schizophrenia demonstrably mitigates long-term disability, yet dropout persists as a common barrier to treatment adherence and completion. A recent 18-month multi-site confirmatory trial of two social work-developed interventions for early course schizophrenia (N = 102), Cognitive Enhancement Therapy (CET; n = 58) and Enriched Supportive Therapy (EST; n = 44), had a study attrition rate of 49%. The largest benefits of treatment in this trial were among participants who completed the full protocols (Wojtalik et al., 2021). As such, it is critical to identify predictors of psychosocial treatment dropout and to understand if predictors vary based on when dropout occurs.

Methods: Using data from Wojtalik et al. (2021), the aim of this research was to identify baseline demographic (age, sex, race, IQ, education, employment, illness length, substance abuse history), medication (dose, adherence), study location, cognition, functional outcome, and symptom predictors of early and late dropout from treatment (CET and EST). Forty-nine (48%) participants completed the full treatment protocols, 40 (39%) dropped during the first half, and 13 (13%) during the latter half. ANOVA and Chi-square tests first examined differences in study predictors across attrition status (completers, early dropout, late dropout). Post-hoc pairwise t-tests using a false discovery rate p-value adjustment explored the directionality of any significant differences. Next, significant characteristics across attrition status were entered as predictors into a multinomial logistic regression using the mlogit package in R. Interactions with treatment assignment were not explored, as attrition status did not differ between CET and EST (p = .290).

Results: Significant differences related to attrition status were observed for age (F^2,99=4.35, p=0.016), IQ (F^2,99=5.16, p=0.007), illness duration (F^2,98=4.03, p=0.021), and antipsychotic dose (F^2,95=4.76, p=0.011). Relative to completers, early dropouts were significantly younger (p=0.021), had a lower IQ (p=0.008), shorter illness duration (p=0.047), and a lower prescribed antipsychotic dose (p=0.009). Similarly, significantly shorter illness duration was observed among late dropouts relative to completers (p=0.047). Post-hoc analyses between early and late dropouts were not significant. In the multinomial model, IQ and antipsychotic dose emerged as significant predictors, such that early dropouts had lower IQ (b=-.09, p=0.003) and lower antipsychotic dose (b=-0.00, p=0.011), relative to completers. The probability of dropping early from treatment decreased by an odds ratio of 0.91 (95% CI: 0.86, 0.97) for every unit increase in IQ and 0.99 (95% CI: 0.99, 0.99) for every unit increase in antipsychotic dose.

Discussion: Attrition in the parent study provided an important opportunity to examine factors associated with dropout during different phases of psychosocial treatment in early course schizophrenia. The findings indicated that individuals who have lower IQs and on lower doses of prescribed antipsychotics may require more support to become engaged with treatment from the start. Such information is vital for developing novel strategies to increase treatment engagement and address attrition challenges within future trials of CET and EST.

Wojtalik, J. A., et al. (2021). Confirmatory efficacy of Cognitive Enhancement Therapy for early Schizophrenia: Results from a multi-site trial. Psychiatric Services. doi: 10.1176/appi.ps.202000552