By feeding already collected data into a new machine learning model, researchers believe they can significantly cut down on the amount of people suffering from dementia that are undiagnosed.
Researchers from the University of Plymouth have developed a new model that routinely scans data collected by the National Health Service in the U.K. to help predict undiagnosed dementia in primary care.
“Machine learning is an application of artificial intelligence [AI] where systems automatically learn and improve from experience without being explicitly programmed,” professor Emmanuel Ifeachor, from the School of Computing Electronics and Mathematics at the University of Plymouth, said in a statement. “It’s already being used for many applications throughout healthcare such as medical imaging, but using it for patient data has not been done in quite this way before.
“The methodology is promising and, if successfully developed and deployed, may help to increase dementia diagnosis in primary care.”
In England, primary care practitioners are encouraged and given incentives to recognize and record dementia in an effort to improve diagnosis rates. However, dementia diagnosis rates in primary care are still low, and many remain undiagnosed or are diagnosed late, when opportunities for therapy and improving quality of life have passed.
According to the study, about half of those living with dementia in the U.K. do not receive a timely diagnosis.
The researchers collected Read-encoded data from 18 consenting general practitioner surgeries in the U.K. for 26,483 patients older than 65. The Read codes are a thesaurus of clinical terms used to summarize clinical and administrative data for U.K. practitioners. The codes were assessed on whether they may contribute to dementia risk, with factors that include weight and blood pressure.
The researchers used the codes to train a machine-learning classification model to identify patients that may have underlying dementia, and found that 84 percent of people who had dementia were detected as having the condition and 87 people without dementia had been correctly acknowledged as not having the condition.
The results show that the model detects those with underlying dementia at an 84 percent accuracy clip, suggesting that the model could in the future significantly reduce the number of those living with undiagnosed dementia from around 50 percent to about 8 percent.
“Dementia is a disease with so many different contributing factors, and it can be quite difficult to pinpoint or predict,” Camille Carroll, PhD, Consultant Neurologist at University Hospitals Plymouth NHS Trust and Researcher in the Institute of Translational and Stratified Medicine at the University of Plymouth, said in a statement. “There is strong epidemiological evidence that a number of cardiovascular and lifestyle factors such as hypertension; high cholesterol; diabetes; obesity; stroke; atrial fibrillation; smoking; and reduced cognitive, physical, or social activities can predict the risk of dementia in later life, but no studies have taken place that allow us to see this quickly.
“So having tools that can take a vast amount of data, and automatically identify patients with possible dementia, to facilitate targeted screening, could potentially be very useful and help improve diagnosis rates,” she added.
The study was published in BJGP Open.
