Modeling changes in distress in a psychotherapy training setting from a middle-income country: A three-level analysis

This study explored change in distress during psychotherapy in a training setting from a middle-income country, examining client demographics and between-therapist variability.

Participants were 220 adults attending ≥5 sessions. Most were women (66.82%); aged 18-62 (Mean = 23.48 [22.61, 24.42], SD = 6.90). Distress was assessed with the CORE-10. Two- and three-level models with linear, quadratic, and logarithmic time functions were compared. Three-level models were fit with and without adjustment for first score, age and gender.

The quadratic three-level model including first score, age and gender fit best (R² = 0.72). Higher initial distress was associated with higher symptom levels across sessions but did not moderate the rate of improvement over time. Therapist effects were small but statistically significant (longitudinal: 2.9%; pre-post: 1.4%). Pre-therapy gender differences were not statistically significant; however, men improved faster initially. Age showed no statistically significant association with pre-therapy distress.

Clients showed early improvements that decelerated and between-therapist differences were small after accounting for baseline distress. Findings suggest that strengthening early engagement and attending to possible gender-related differences in early response may help inform service delivery in middle-income countries, although these strategies were not directly tested in the present study.