AI anticipates dementia
Researchers from the University of Exeter, UK, using data from more than 15,300 patients in the US, have found that machine learning can accurately identify patients at high risk of developing dementia. The method works by identifying hidden patterns in the data and studying who is most at risk. The study also showed that the algorithm can reduce the number of people who are misdiagnosed with dementia.
The researchers analyzed data from patients who visited one of the 30 clinics of the National Alzheimer's Coordinating Center memory clinics in the United States. At the beginning of the study, the participants did not have dementia, although many experienced problems with memory or other brain functions. From 2005 to 2015, every tenth participant (1,568 people) was diagnosed with dementia for the first time within two years after visiting the clinic. In the study, a machine learning model predicted these new cases of dementia with up to 92% accuracy – much more accurately than existing alternative methods of BDSI and CAIDE – dementia risk prediction models that assign patients a score representing their risk of developing dementia over a wider time frame.
The researchers also estimated that about 8% (130) of dementia diagnoses were probably made by mistake, as they were subsequently canceled. Artificial intelligence has accurately identified more than 80% of such contradictory diagnoses. In other words, he was able not only to accurately predict who would be diagnosed with dementia, but also has the potential to improve the accuracy of these diagnoses.
Machine learning used patient information commonly available at the clinic, including assessments of memory and brain function, cognitive test results, and certain lifestyle factors.
Artificial intelligence, the authors hope, can improve the diagnostic process for people who have the first symptoms of dementia. This method will give doctors the basis for recommendations on lifestyle changes and identify people who need an in-depth examination.
Currently, the group plans to refine the method for practical use in medical institutions.
Article by C.James et al. Performance of Machine Learning Algorithms for Predicting Progression to Dementia in Memory Clinic Patients is published in JAMA Network Open.
Aminat Adzhieva, portal "Eternal Youth" http://vechnayamolodost.ru based on the materials of the University of Exeter: Artificial intelligence accurately predicts who will develop dementia in two years.