Machine learning could play a significant role in the future treatment of schizophrenia patients, according to new research published in JAMA Network Open.
Antipsychotic medication can help reduce the symptoms of schizophrenia and limit the chances of a relapse, but finding the right medication for patients with first-episode schizophrenia remains a critical challenge. Developing an individualized treatment rule (ITR) for such individuals could help patients receive the medication they need as early as possible—and the study’s authors thought machine learning could potentially assist with that development.
The researchers explored data from more than 32,000 Taiwanese patients with first-episode schizophrenia, using an ensemble machine learning technique to develop a preliminary ITR. The data came from Taiwan’s National Health Insurance Research Database, which covers all reimbursed insurance claims from the region’s mandatory government-sponsored insurance program. The team’s preliminary ITR was trained on 70% of the dataset, resulting in an estimated 51.7% success rate when applied to the remaining 30%. If that remaining 30% were treated without the ITR, their success rate would have been 44.5%, showing the considerable boost in care that came from the researchers’ efforts.
“To our knowledge, this work is the first machine learning study to develop an ITR for first-episode schizophrenia,” wrote first author Chi-Shin Wu, MD, PhD, National Taiwan University, and colleagues. “We found that a preliminary ITR based on relatively simple patient characteristics had a significantly higher cross-validated estimated treatment success rate compared with the observed population rate.”
The authors did discuss their study’s limitations, including its observational design and the limited measured predictors available for the development of the ITR. Wu et al. were clear that their findings were positive—but there is still more work to be done before implementation.
“Results of this study suggest that developing a clinically useful ITR to increase 12-month treatment success among patients with first-episode schizophrenia may be possible,” the authors concluded. “Rigorous evaluation in a clinical trial of such an ITR is needed before clinical implementation.”
Michael has more than 16 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.