Impact of a personalized antidepressant prescription using genetics, socio-demographic and clinical data in major depressive disorder patients: A clinical study

Abstract

Background: Major depressive disorder (MDD) is common and disabling, and antidepressant selection often follows a trial-and-error process. Predictix is an artificial intelligence-based clinical decision support tool integrating genetic, socio-demographic, and clinical data to assist personalized prescribing.

Objective: To evaluate feasibility, preliminary clinical outcomes, tolerability, and acceptability of Predictix-guided antidepressant selection.

Methods: This single-center, open-label pilot study was conducted at Pitié-Salpêtrière Hospital (Paris) between August 2020 and July 2021. Thirty adults meeting DSM-5 criteria for MDD were enrolled. Treatment was initiated according to Predictix recommendations and assessed at weeks 4 and 8. The primary outcome was response (≥50% reduction in QIDS-SR-16). Secondary outcomes included PHQ-9 and CGI-I response, tolerability, and acceptability. Analyses were performed on study completers.

Results: Twenty patients completed the study and 18 were included in the primary analysis. The primary response rate was 55.6% (95% CI: 33.7-75.4%). Secondary response rates were 61.1% (PHQ-9) and 72.2% (CGII). Clinical improvement (CGI-I ≤ 3) occurred in 94.4%. Mean treatment duration was 9.4 ± 2.3 weeks. Tolerability was favorable, with 66.7% showing FIBSER burden ≤2. Acceptability was generally positive, although the two-week genetic analysis delay was noted as a practical limitation.

Conclusions: Predictix-guided prescribing was feasible and showed response rates within reported clinical ranges. Larger randomized controlled trials are required to determine clinical added value over standard prescribing.

Keywords: Artificial intelligence (AI) in psychiatry; Clinical decision support systems (CDSS); Major depressive disorder (MDD); Personalized antidepressant prescription; Pharmacogenomics.