AI antidepressant choice tool improves patient adherence and outcomes, trial shows

An online AI-driven tool that helps guide choice of antidepressant can lead to better adherence and better long-term outcomes, a UK-led study has found.

The PETRUSHKA tool, developed by the University of Oxford uses AI to interrogate clinical and demographic information to help tailor choice of antidepressant.

It also takes into account patient preferences, especially around what side effects may be acceptable.

Now a large randomised controlled trial across the UK, Canada and Brazil has shown that those who selected which antidepressant to use with support from the tool were significantly more likely to continue their treatment.

The study included more than 500 adults with major depressive disorder. It showed that when the tool was used, participants were 40% less likely to discontinue their antidepressant within the first eight weeks of treatment

Writing in JAMA, the researchers said fewer people stopped treatment because of adverse effects, and by 24 weeks those in the PETRUSHKA group also reported greater improvements in depressive and anxiety symptoms.

Overall by eight weeks, 22 of 241 participants (9%) who had used the tool to guide treatment discontinued the prescribed antidepressant due to adverse events compared with of 252 (16%) in the usual care group.

But the researchers did note that the trial was not double blinded and there was missing data which needs to be taken into account when interpreting the results.

Designed to be used in everyday clinical practice, including by GPs, the tool only takes three minutes to complete, the researchers noted.

Study lead Professor Andrea Cipriani, professor of psychiatry at the University of Oxford and honorary consultant psychiatrist at Oxford Health NHS Foundation Trust, said: ‘Mental health is lagging behind other fields of medicine and for too long, antidepressant treatment has relied on trial and error.

‘PETRUSHKA shows that by combining the best available evidence with patients’ own preferences, we can personalise antidepressant treatment from the outset and help more people in the NHS stay on the medication that is right for them.’

Guidelines on treatment for depression, including from NICE, advocate personalised treatment decisions but implementation in practice remains limited.

Other areas of medicine such as oncology and cardiology already used prediction tools routinely but this had proven more difficult in psychiatry where trajectories can be complex.

He said the tool may be especially valuable in primary care and other non-specialist services, where most people with depression are treated and clinicians are pressed for time.

Mike Lewis, Scientific Director for Innovation at the National Institute for Health and Care Research, who provided funding for the study, said it showed the power of combining digital technology and personalised treatment. 

‘By harnessing data and embracing cutting-edge digital tools, we can tailor care more precisely to each patient — improving outcomes for individuals.’

More research is now needed to look at cost effectiveness and longer-term impact, the researchers said.