Learning Outcomes that Maximally Differentiate Psychiatric Treatments

Eric V Strobl,Semmie Kim

DOI: https://doi.org/10.1101/2024.12.03.24318424

2024-12-05

Abstract:Background: Matching each patient to the most effective treatment option(s) remains a challenging problem in psychiatry. Clinical rating scales often fail to differentiate between treatments because most treatments improve the scores of all individual items at only slightly varying degrees. Methods: We introduce a new exploratory analysis technique called Supervised Varimax (SV). The algorithm combines the individual items that only slightly differ between treatments into a few scores that greatly differ between treatments. SV further enforces uncorrelatedness between the scores, so that they represent distinctly interpretable biopsychosocial factors. We applied SV to multi-center, double-blind, randomized and large-scale clinical trials called CATIE and STAR*D which were long thought to identify few to no differential treatment effects. Outcomes: SV identified differential treatment effects in Phase I of CATIE (n=1444, absolute sum = 1.279, p = 0.0007). Post-hoc analyses revealed that olanzapine is more effective than quetiapine and ziprasidone for hostility in chronic schizophrenia (range = 0.284, p_{FDR}=0.046; range = 0.283, p_{FDR}=0.046), and perphenazine is more effective than ziprasidone for emotional dysregulation (range = 0.313, p_{FDR}=0.046). SV also discovered that buproprion augmentation is more effective than buspirone augmentation for treatment-resistant depression with increased appetite from Level 2 of STAR*D (n=520, range = 0.280, p_{FDR}=0.004). Interpretation: SV represents a powerful methodology that enables precision psychiatry from clinical trials by optimizing the outcome measures to differentiate between treatments.

Psychiatry and Clinical Psychology