Prognostic and prescriptive predictors of PTSD response to prolonged exposure and sertraline

In developing personalized care for PTSD, baseline predictors that inform clinicians' and patients' decision making across treatment options are limited. In 200 individuals with PTSD in a multi-site, doubly randomized preference trial, baseline psychopathology, personality/mood regulation, health/social functioning, treatment credibility and history, and demographic factors predicted PTSD severity from baseline, ten sessions, to post-treatment. Using a stepwise prognostic and prescriptive modeling approach, rates of change and post-treatment outcomes between prolonged exposure [PE] vs sertraline and preferred vs non-preferred treatments were examined. Stronger credibility of PE predicted faster improvement (b = -0.28, SE = 0.10) and lower post-treatment scores (b = -2.40, SE = 1.08). Higher avoidance predicted poorer outcomes for those who did not receive their preferred treatment (b = -5.33, SE = 2.42). Similarly, being psychiatric medication naïve predicted poorer response for those who did not receive their preferred treatment (b = -13.03, SE = 5.71). Across a range of baseline predictors, stronger credibility of PE, potentially reflecting the patient's perceived need to actively discuss and process the trauma memory, emerged as one of the best predictors of PTSD treatment outcome. Particular attention should be paid to matching more avoidant and medication-naïve patients to their preferred treatment.

All data, program code, and other methods developed by others have been cited appropriately in text. The dataset for this study comes from a previously published clinical trial [NCT00127673, 29] and is available by request by competent researchers. Computer syntax used for this study is available upon request. Research materials, including treatment rationales and manuals, are available upon request.