IntroductionIntensive Short-Term Dynamic Psychotherapy has shown promising effects for treatment-resistant depression, but it remains unclear whether its hypothesized mechanisms — reducing emotional repression, negative affect, and psychological distress — actually mediate treatment outcomes.MethodsWe reanalyzed publicly available data from a randomized controlled trial (N = 86) comparing 20 sessions of Intensive Short-Term Dynamic Psychotherapy to waitlist control for treatment-resistant depression. Depression and process measures were assessed at baseline, post-treatment, and 3-month follow-up. Linear mixed-effects models analyzed trajectories; bootstrap mediation and cross-lagged panel analyses provided an exploratory examination of proposed process variables.ResultsTreatment produced large effects on depression at post-treatment (Cohen’s d = 1.68) that continued to increase through 3-month follow-up (d = 2.50, 95% CI [1.88, 3.11]). All proposed process measures also showed very large effects (d = 1.96–2.95). However, neither emotional repression nor negative affect significantly mediated depression improvement. Distress showed apparent mediation, but a sensitivity analysis removing the overlapping depression subscale eliminated this effect entirely, confirming it reflected construct overlap rather than a genuine indirect effect. Cross-lagged analyses revealed no temporal precedence for any process measure, indicating concurrent rather than sequential change.DiscussionThese findings confirm that this psychotherapy produces large, durable effects on treatment-resistant depression. However, the theorized sequential process — whereby reducing defensive functioning leads to improved affect regulation, which in turn alleviates depression — was not supported in the present data. Instead, the treatment appears to produce broad, simultaneous therapeutic change across multiple psychological domains. These exploratory findings suggest that understanding how psychotherapy works may require finer temporal measurement and observational methods that capture in-session processes.
Radically different