(SSCP1-10) SSCP Submission: Who Will Do the Work? Predictors of Client Engagement in DBT

There is substantial evidence that retention and homework compliance are significantly related to positive treatment outcomes in cognitive behavioral therapy for depression (Fernandez, Salem, Swift, & Ramtahal, 2015; Kazantzis, Whittington, & Dattilio, 2010), with recent evidence suggesting these findings extend to dialectical behavioral therapy (DBT) for borderline personality disorder (BPD; Yin, Stern, Kleiman, & Rizvi, 2023). We are interested in examining predictors of treatment engagement (i.e., retention, group attendance, homework completion) to identify individuals who may be most at risk for insufficient DBT treatment dosage. With a sample of 83 individuals enrolled in a DBT program at a graduate training clinic, we investigated predictors of dropout and, for a subset of 71 individuals who completed post-session surveys, estimated homework completion and group attendance. We conducted three LASSO regressions (a machine learning approach) to detect relevant predictors of dropout, percent of skills groups attended, and self-reported percent of weekly skills group homework completed. As possible predictors each model included: demographic variables (i.e., age, sex, marital status, race, education, income), therapist assignment, number of comorbidities, client-reported emotion dysregulation and BPD severity at baseline, and client-reported working alliance (i.e., goal, task, and bond) with their therapist after session four. Here, we only report retained predictor variables for each measure of client engagement. Lower likelihood of dropout (pseudo R² = 0.19) was associated with female sex, less severe emotion dysregulation and BPD features at baseline, more comorbidities, younger age, White race, more years of education, and a stronger affective therapeutic bond and agreement on goals. Higher percent homework completion (pseudo R² = 0.20) was associated with female sex, more years of education, older age, being partnered, therapist assignment, and a stronger affective therapeutic bond and agreement on therapeutic task. Lastly, higher percent group attendance (pseudo R² = 0.11) was associated with less severe emotion dysregulation and BPD features at baseline, therapist assignment, and a stronger therapeutic bond. These results highlight that individuals with more years of education and lower severity may be at an advantage during DBT treatment due to an increased likelihood of engagement with treatment, which may warrant future studies assessing tailored treatment for populations with fewer years of education and greater severity. We also found support that a strong working alliance is associated with higher engagement throughout treatment, emphasizing the importance of building rapport and common goals with clients early in treatment. In the future, research testing whether client engagement variables mediate the relationship between baseline or early treatment variables with treatment outcomes could determine if client engagement is a mechanism through which early treatment predictors impact treatment response.