A Permutation-Based ANCOVA to Evaluate Psychotherapy Effects on Depression Outcomes Adjusted for Baseline Severity

Author: Amira Mandour
Biostatistician | Clinical Trials & Statistical Modeling

2023-05-19

Research Question:

Does the type of psychotherapy lead to different post-treatment depression scores after adjusting for baseline depression severity?

Objective:

To compare depression outcomes between a new psychotherapy and standard therapy while adjusting for baseline severity using a permutation-based ANCOVA.

Statistical Analysis:

Continuous variables were summarized using means and standard deviations. The primary outcome was the post-treatment depression score, and the primary exposure was psychotherapy group (new psychotherapy vs. standard therapy). Baseline depression severity was included as a covariate to adjust for pre-treatment differences between participants.

Group differences in post-treatment depression scores were evaluated using a permutation-based ANCOVA adjusting for baseline depression severity. The analysis was performed using the adonis2 function in the vegan package in R. A distance matrix was computed using the Bray–Curtis dissimilarity metric, and statistical significance was assessed using 999 permutations under a free permutation scheme. The model included psychotherapy group and baseline severity as explanatory variables, entered sequentially.

Statistical significance was evaluated at a two-sided α level of 0.05. All analyses were conducted using R.

Results:

The permutation-based ANCOVA analysis suggested that baseline depression severity is a strong and statistically significant predictor of post-treatment depression scores. The type of psychotherapy, however, does not appear to have a significant effect on depression outcomes after adjusting for baseline severity.

Baseline severity explains 27% of the variance in depression scores, indicating that individuals with more severe depression at baseline are more likely to have higher post-treatment depression scores, regardless of the therapy type.

The therapy type (New Psychotherapy vs. Standard Therapy) does not significantly influence post-treatment depression scores after adjusting for baseline severity, suggesting that the differences between the two therapies may be minimal when baseline severity is controlled for.

The p-values for the group variable (0.383) and baseline severity (0.001) support the conclusion that baseline severity is a significant covariate in explaining post-treatment depression scores.