AI could be used to predict outcomes for people at risk
Artificial intelligence (AI), or more precisely in this case – machine learning, has scored another point against its human counterparts in the field of medicine by outperforming clinicians in predicting how well people at risk of psychosis and those recently diagnosed with depression are likely to function socially.
Noting the results of brain imaging and clinical measures from client interviews, the research team, led by Professor Stephen Wood from Orygen, the National Centre of Excellence in Youth Mental Health (Australia) compared machine learning algorithms trained on functional, neuroimaging, and combined baseline data.
The study followed 116 people (aged 15 to 40) at high risk of developing psychosis, 120 people who have been recently diagnosed with depression, and 176 healthy controls for 18 months in 7 academic early-recognition services in 5 European countries.
According to Professor Wood, while all tested approaches did impressively well, the combined approach came out on top, accurately predicting social outcomes one year later in up to 83 percent of patients at high risk of psychosis and 70 percent of patients with recent-onset depression.
“Adding neuroimaging machine learning to clinical machine learning provided a 1.9-fold increase of prognostic certainty in uncertain cases of patients in CHR [clinical high-risk] states, and a 10.5-fold increase of prognostic certainty for patients with ROD [recent-onset depression],” wrote the authors in the paper.
The ability to foresee the likely life path of people at high risk of psychosis and clinical depression with increased accuracy could allow clinicians to develop individualised treatments for each patient, thereby potentially improving their ability to function in social situations and remain employed.
“Predicting social outcomes is important as among young people and emerging adults in OECD countries the top causes of ‘disability’ – and poor social functioning is included in that – are mostly disorders of mental health, including those that typically present with first episode of psychosis,” said Wood.