A team of European scientists led by researchers from the Max Planck Institute recently developed the world’s first cybernetic system for predicting the outbreak of psychosis in high-risk patients.
According to the NIH, about three percent of the population (data are US specific) will have psychosis at some point in their lifetime. To put that in perspective, the chances of being stung by a bee are roughly six million to one.
Unfortunately, predicting psychosis in high-risk patients is a difficult task. The current paradigm requires intensive diagnosis by trained professionals in a specialized medical facility to which most people in the world do not have immediate access.
According to the scientists’ research report:
The clinical benefit of CHR [clinical high risk] The designation can be further restricted as it is tedious to identify and is limited to specialized, well-resourced health services that do not adequately cover the population at risk. Therefore, improved predictive accuracy and clinical scalability are required to accurately identify patients at real risk for psychosis.
In action, this means that healthcare workers are much better able to determine which patients will later develop psychosis. The current usefulness of the designation for clinical high risk (CHR) is questionable, since, according to the researchers, only about 22% of the identified people express psychoses.
The efforts of the European research team involved combining known human diagnostic methods into a cybernetic stack with myriad algorithmic components.
In this prognostic study, we identified generalizable risk assessment tools that can be arranged in a multimodal prognostic workflow for a clinically viable, individualized prediction of psychosis in patients with CHR conditions and ROD. To the best of our knowledge, our study has shown for the first time that expanding human forecasting capabilities by recognizing algorithmic patterns improves forecasting accuracy to margins that are likely to justify the clinical implementation of cybernetic decision-making aids.
Take quickly: Researchers identified several hundred CHR patients and trained ML models to determine risk for schizophrenia using “multimodal machine learning that optimally integrates clinical and neurocognitive data, structural magnetic resonance imaging (sMRI) and polygenic risk assessments (PRS). Assessment of the geographical generalizability of models; test and incorporate the clinician’s predictions; and to maximize clinical benefit by building a sequential prediction system. “
That’s a mouthful, but it means the researchers used the same sources of data that a healthcare professional would use for diagnostic purposes to predict psychosis, and then combined them with machine learning models that could draw more useful conclusions.
In fact, the system showed almost the same level of accuracy as humans in terms of detection and diagnosis. The reason for this is that, as mentioned earlier, there simply aren’t enough health care facilities in the world that can diagnose psychosis. This AI system could expand existing clinics and potentially enable advanced diagnostic capabilities in places where relative medical care is not available to humans.
Published on January 20, 2021 – 19:29 UTC