Abstract
Dropout rates for treatments for adult posttraumatic stress disorder (PTSD) are high. To date, most investigations have focused primarily on patient factors to understand dropout, but it is also important to investigate therapist-patient factors, such as the therapeutic alliance. The alliance is commonly measured as three components: the therapist-patient bond, agreement on tasks, and agreement on goals. The contributions of each of these components to dropout have not been examined but could highlight important points of intervention to increase engagement. An observational coding system, the Working Alliance Inventory-Observer Version-Short Form (WAI-O-S), was used to code these alliance components in audio-recorded early sessions of cognitive processing therapy (CPT), a gold-standard treatment for PTSD. The therapist-patient Bond, Tasks, and Goals subscales were examined as predictors of dropout in CPT. Participants were 63 treatment-seeking adults with PTSD who received CPT as part of a randomized noninferiority trial. Those who completed the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) at pretreatment and had at least one session available to code (n = 52) were included in analyses. Of those, 15 (28.8%) dropped out of CPT early (all dropped in ≤9 sessions) and 37 (71.2%) completed treatment. Scores on each of the alliance subscales were averaged across the first seven available session recordings. Logistic regression analyses showed that more disagreement on tasks predicted higher risk of dropout (OR = 0.26, 95% CI [0.08, 0.90], p = .034), whereas the bond and agreement on goals did not predict dropout. These findings highlight the importance of examining the specific components of the alliance and suggest that addressing patient reactions to therapeutic tasks may enhance collaboration and decrease the risk of dropout.