Abstract: OBJECTIVE: Despite effective treatment options for posttraumatic stress disorder (PTSD), many patients do not complete therapy. This includes U.S. active duty service members, yet factors linked to attendance in this population remain understudied and dropout remains difficult to predict. Additionally, most studies have not examined samples with PTSD and co-occurring major depressive disorder (MDD) despite high rates of comorbidity. METHOD: The current study explored predictors of dropout among service members with comorbid PTSD and MDD (N = 94) randomized to cognitive processing therapy enhanced with behavioral activation (BA + CPT) or CPT as part of a clinical trial. RESULTS: Using the Fournier approach, only two predictors were associated with lower dropout risk among over 20 examined: shorter duration between pretreatment assessment and Session 1 (p = .041) and past 3-month PTSD treatment engagement (p = .036). CONCLUSION: Results suggest the possible utility of early momentum in starting therapy and leveraging recent treatment to improve attendance. However, this study also highlights the possible limitations of commonly assessed pretreatment factors in predicting attendance and current challenges in measuring dropout risk. Strategies to improve prediction, such as shifting focus to assess modifiable factors and processes more proximal to dropout during treatment, may be needed.