JAMA Trial: A Web-Based Tool Personalises Antidepressant Choice — 540 Patients Across 47 Sites
- Multicentre RCT: 540 participants with MDD randomised to PETRUSHKA decision-support tool (n=271) vs usual care (n=269), across 47 sites in Brazil, Canada, and the UK
- PETRUSHKA is a web-based clinical decision-support system that uses patient-level data to recommend the most appropriate antidepressant — moving beyond trial-and-error prescribing
- Aimed at reducing premature antidepressant discontinuation by matching medication to individual patient profiles
- Published in JAMA — the first large-scale RCT of an evidence-based antidepressant personalisation tool
The "try this SSRI, and if it doesn't work in 6 weeks, try another" model is how most antidepressant prescribing works. It is also why 30-40% of patients discontinue prematurely — the wrong medication for the wrong patient produces side effects and no improvement, and the patient leaves treatment. This JAMA trial tests whether a decision-support tool can match patients to antidepressants more precisely.
How PETRUSHKA works
The tool collects patient-level data — demographics, symptom profile, comorbidities, prior medication history, side effect sensitivity — and generates an evidence-based recommendation for which antidepressant is most likely to work for that specific patient. It does not replace clinical judgement; it augments it with structured evidence synthesis that no clinician can do mentally across dozens of medication options.
Why this matters for prescribers
The 540-patient, 47-site, three-country design gives this trial external validity that most decision-support studies lack. If the tool demonstrates efficacy, it changes the prescribing conversation from "let's start with escitalopram because it's generally well-tolerated" to "your profile suggests sertraline has the highest probability of response with your specific comorbidity pattern."
For psychologists and therapists: this is the pharmacological equivalent of what you do when you match a patient to CBT vs EMDR vs DBT based on their presentation. Medication prescribing is catching up to the personalisation that psychotherapy selection has long practiced.
A 540-patient JAMA trial tests whether a web-based decision-support tool can match depressed patients to the right antidepressant from the start — replacing trial-and-error with evidence-based personalisation.
Full results pending publication of primary outcomes. Tool performance depends on quality of input data. Implementation requires clinician training and workflow integration. Three countries — may not generalise to all healthcare systems.