Associations Between Reward Probability and PTSD Symptom Clusters after Controlling for Depression: A Structural Equation Model

Abstract

The Reward Probability Inventory (RPI) is emerging as a novel self-report measure for assessing posttraumatic stress disorder (PTSD) severity. However, prior studies have not examined RPI associations with specific PTSD symptom clusters nor how depression accounts for variance in the RPI-PTSD relationship. In a cross-sectional design, 428 university student participants (n = 95 males; n = 318 non-Hispanic White) completed the DSM-5 PTSD Checklist, the Depression, Anxiety, and Stress Scale, and the RPI subscales: Reward Probability (RPI-RP) and Environmental Suppression (RPI-ES). Structural equation modeling (SEM) tested associations between RPI subscales and four PTSD symptom clusters (Intrusion, Avoidance, negative cognitions/mood, and hyperarousal), with and without depression as a latent covariate. Sex was included as a covariate, whereas age and ethnicity were not significantly associated with study measures and were thus excluded. RPI-ES was negatively associated with all four PTSD symptom clusters across both models. RPI-RP showed variable and counterintuitive positive associations with Intrusion and Avoidance in joint models, contrary to zero-order correlations, suggesting a suppressor effect. Two exploratory models confirmed that when modeled in isolation, RPI-RP demonstrated the expected negative associations with all clusters, indicating that the sign reversal reflects a statistical artifact driven by the substantial intercorrelation between RPI-RP and RPI-ES (r = 0.58). Depression accounted for additional variance in PTSD symptom severity beyond the RPI subscales. Findings support the utility of the RPI, particularly RPI-ES, in understanding PTS symptom severity beyond depression, and bear relevance to treatments targeting reward-related deficits in comorbid PTSD and depression.