A new paper by researchers at the University of Birmingham shows that a ‘one size fits all’ approach may not be addressing the needs of all young people in the early stages of psychosis. Instead, in the paper, published in Translational Psychiatry, the researchers argue that machine learning techniques could be employed to deliver treatment plans that are specifically targeted at groups or even individuals.
This approach could mean greater precision in designing treatment plans and also a better success rate in identifying patients who are on the wrong treatment pathway.
Early Intervention in Psychosis Services were first established in the 1990s and became recognised as offering the best chance of recovery for young people with a first episode of psychosis. Available treatments currently include antipsychotic medication, intensive community-based care, and social and psychological interventions.
Lead researcher, Dr Lowri Griffiths, who was invited by the Editor-In-Chief of Translational Psychiatry to contribute this review, said: “It is well known that early intervention leads to better outcomes, particularly among young people. However, despite receiving gold standard treatment, a significant number of people are not benefitting from these interventions.
“We need to consider a range of factors from psychological, biological, and social circumstances to find the right treatments, for the right people, at the right time, to maximise a young person’s life chances. But in the first instance, this requires doing more to reach out to diverse and representative groups to ensure care is equitable for all”.
The technology needed to devise treatment plans for individual patients, or groups of patients, already exists. What we need is a framework that will enable us to investigate and test these technologies so we can harness them to improve outcomes for patients with psychosis.Paris Lalousis, Centre for Human Brain Health
A machine learning approach, the researchers argue, could serve as a ‘guide’ for clinical decision-making, identifying with increasing accuracy the key markers in patient data that would indicate the likely success or failure of any particular pathway.
This approach would also help to ensure that more patients were able to access the treatments most likely to benefit them, regardless of environmental and social circumstances which might otherwise lead to inequality in healthcare.
Co-lead author Dr Paris Lalousis said: “The technology needed to devise treatment plans for individual patients, or groups of patients, already exists. We see machine learning already in use in a number of clinical areas, such as predicting responses to cancer treatment, or identifying individuals at risk of needing intensive care. What we need is a framework that will enable us to investigate and test these technologies so we can harness them to improve outcomes for patients with psychosis.”