Identifying Patients at Risk for Psychotherapy Dropout in Community Settings

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Importance Psychotherapy dropout poses a significant public health problem and is associated with depression relapse and illness persistence. Identifying who is at risk of psychotherapy dropout can guide personalized and scalable strategies to mitigate risk and improve outcomes. Objective To estimate the prevalence and predictors of psychotherapy dropout in a large community setting. Design, Setting, and Participants In this prognostic study, electronic health record data collected between 2008 and 2022 were analyzed from 2 large academic medical centers, 6 community hospitals, and their affiliated outpatient networks in Massachusetts. Logistic regression (LR) and random forests (RF) machine learning were applied to identify predictors of dropout. Participants were individuals aged 18 to 80 years with depression who received at least 1 individual psychotherapy session. Data analysis was performed between January 2023 and March 2026. Main Outcomes and Measures Predictors included age, self-identified race and ethnicity, medical conditions, and health care encounters. Dropout rate was assessed after 3 or fewer psychotherapy sessions. LR and RF models were trained to maximize accuracy (area under the curve [AUC]) and extracted top predictors of dropout. Results The final cohort included 40 732 individuals (mean [SD] age, 44.2 [14.2] years; 69.3% female; 2% Asian, 8% Black, 71% White, and 18% other races; 9% Hispanic). The psychotherapy dropout rate was 28.4% (n = 11 571), with 60.9% dropping out after the first session. AUC values were 0.64 (95% CI, 0.63-0.65) for LR and 0.66 (95% CI, 0.65-0.67) for RF. Prior mental health encounters (eg, group psychotherapy [odds ratio (OR), 0.91; 95% CI, 0.90-0.91]), psychiatric evaluation (OR, 0.86; 95% CI, 0.85-0.87), and White race (OR, 0.90; 95% CI, 0.89-0.90) and/or non-Hispanic ethnicity (OR, 0.95; 95% CI, 0.95-0.95) predicted decreased dropout likelihood. A mental disorder due to a physiological condition (OR, 1.05; 95% CI, 1.04-1.05), and prior medical admission (OR, 1.03; 95% CI, 1.03-1.04) predicted increased likelihood of dropout. Conclusions and Relevance In this prognostic study, more than one-quarter of patients dropped out of psychotherapy, most after a single session, highlighting the need for early risk detection. Brief mental health encounters early in the health care pathway were associated with lower dropout risk, suggesting that exposure to mental health within medical settings may facilitate engagement. Electronic health record data can inform risk detection and targeted interventions in the community.