dementia
Artificial intelligence can now predict whether someone will develop dementia within two years, a new study claims. Researchers hope the technology will cut down on false diagnoses and help doctors catch the condition earlier.

Researchers from the University of Exeter in the UK took data from more than 15,000 memory clinic patients in the US and used this information to train machine learning algorithms to spot patterns among them. After analyzing this data, the AI was able to detect who would develop dementia within the next two years with 92% accuracy, according to a study published by the scientists on Thursday.

Some 258 variables were analyzed to determine each patient's dementia risk, but the machine learning algorithm was able to reach an accuracy level of 91% with just six of these variables.

Furthermore, the researchers discovered that around 8% of patients had been misdiagnosed with dementia, with the AI system finding 80% of these erroneous diagnoses.

Currently, doctors can use a number of decision-making aids to predict a patient's risk of developing dementia on a longer timeline, such as the CAIDE Risk Score, which predicts the 20-year risk for dementia in middle-aged people; and the BDSI score, which identifies elderly patients at risk of dementia in the next six years. However, no decision-making aid has yet been developed to predict the onset of the condition in such a short period.

"This has the potential to reduce the guesswork in clinical practice and significantly improve the diagnostic pathway, helping families access the support they need as swiftly and as accurately as possible," study author Prof. David Llewellyn said.

Patients often turn up at memory clinics suffering from mild cognitive impairment, which then progresses to dementia. The researchers hope that by screening out those most likely to develop the condition, they can help doctors prioritize those needing follow-up care.

Dementia is a catch-all term for a range of progressive neurological disorders impacting memory, thinking, and behavior, of which Alzheimer's disease is the most common. There were over 55 million people worldwide living with dementia in 2020, according to Alzheimer's Disease International. This number is predicted to double, reaching 78 million in 2030 and 139 million in 2050, as the proportion of elderly people increases worldwide.

Reference: Charlotte James, PhD; Janice M. Ranson, PhD; Richard Everson, PhD; et al, Performance of Machine Learning Algorithms for Predicting Progression to Dementia in Memory Clinic Patients, JAMA Netw Open. 2021; doi:10.1001/jamanetworkopen.2021.36553