Evaluating a Decision Support Tool for Antidepressant Selection

Key takeaways:

• At 8 weeks, treatment discontinuation for any cause was lower in the PETRUSHKA group than in the usual-care group.
• Discontinuation due to adverse events was also lower at 8 weeks with PETRUSHKA.
• At 24 weeks, depressive and anxiety symptom scores were lower with PETRUSHKA, although missing data and the lack of double blinding limit confidence.

In a JAMA randomized clinical trial, adults with major depressive disorder had lower 8-week treatment discontinuation with PETRUSHKA than with usual care, at 17% versus 27%.

PETRUSHKA is a web-based decision-support system for antidepressant selection that incorporates clinical and demographic predictors with patient preferences. Investigators compared the tool with usual care in a randomized multicenter study across 47 sites in Brazil, Canada, and the UK.

The trial enrolled adults aged 18 to 74 years with major depressive disorder at 47 sites in Brazil, Canada, and the UK. A total of 540 participants were randomized 1:1 to PETRUSHKA (n = 271) or usual care (n = 269); among 520 eligible participants, 493 were included in the primary analysis. Participants had a median age of 35 years, 58% were female, and baseline mean scores were 16.6 on the PHQ-9 and 11.5 on the GAD-7. PETRUSHKA combined clinical and demographic predictors with patient preferences to personalize antidepressant treatment, and the primary endpoint was any-cause discontinuation at 8 weeks.

At 8 weeks, 41 of 241 participants in the PETRUSHKA group discontinued treatment for any cause, compared with 69 of 252 in usual care. That corresponded to 17% versus 27%, with an adjusted relative risk of 0.62, a 95% CI of 0.44 to 0.88, and P = .007. Discontinuation due to adverse events was 22 of 241 participants, or 9%, with PETRUSHKA and 39 of 252, or 16%, with usual care. The adjusted relative risk for adverse-event discontinuation was 0.59, with a 95% CI of 0.36 to 0.97 and P = .04, leaving more PETRUSHKA participants on treatment at 8 weeks.

At 24 weeks, symptom outcomes also favored PETRUSHKA over usual care. Mean PHQ-9 scores were 7.1 with PETRUSHKA and 9.2 with usual care, with an adjusted between-group mean difference of −1.92. The 95% CI for that difference was −3.06 to −0.78, and the reported P value was less than .001. Mean GAD-7 scores were 4.6 with PETRUSHKA and 5.8 with usual care, producing an adjusted between-group mean difference of −1.39, with a 95% CI of −2.26 to −0.52 and P = .002.

The investigators concluded that PETRUSHKA was associated with more patients remaining on antidepressant treatment at 8 weeks and with lower depressive and anxiety symptoms at 24 weeks. Notably, the lack of a double-blind design and the large amount of missing data limit the validity of the results.