Predicting Participation in Psychiatric Randomized Controlled Trials: Insights From the STEP-BD

Area of Science:

• Clinical Psychology
• Psychiatric Research
• Medical Sociology

Background:

• Differences between participants and non-participants in randomized controlled trials (RCTs) can affect the generalizability of findings.
• The Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) enrolled patients with bipolar disorder, allowing for observational participation and optional RCT enrollment during depressive episodes.

Purpose of the Study:

• To investigate whether patients participating in RCTs for acute bipolar depression differ from those in the observational arm of the STEP-BD study.
• To identify demographic, clinical, and site-related factors associated with RCT participation.

Main Methods:

• Logistic regression models were used to estimate the odds of RCT enrollment among 2,222 STEP-BD participants with acute bipolar depression.
• Predictor variables included patient demographics, clinical severity (Clinical Global Impression scale), comorbidities, and study site.
• Area under the receiver operating characteristic curve (AUC) was calculated to assess the predictive power of study site on RCT participation.

Main Results:

• RCT participation was associated with having no insurance (OR=1.58) and greater symptom severity (OR=1.52).
• Study site emerged as the most significant predictor of RCT enrollment, with predicted probabilities ranging from 8% to 31%.
• The model including site had a higher AUC (.70) compared to the model excluding site (.61), indicating site's strong influence.

Conclusions:

• Few demographic or clinical differences were observed between RCT participants and the observational group in the STEP-BD study.
• Study site was the strongest determinant of RCT participation, suggesting potential biases in generalizability.
• Further research is needed to explore site characteristics influencing participation and their impact on patient outcomes in real-world settings.