Researchers from University College London (UCL) in the UK have developed an algorithm that predicts the risk of patients developing dementia based on their health data collected from routine GP visits.
During the study, the UCL researchers used anonymised patient data collected from 377 medical practices in the UK to predict an individual’s future risk of developing the disease within the next five years.
According to research published in the open access journal BMC Medicine, the Dementia Risk Score has an up to 85% accuracy rate.
“The score could be especially useful for identifying people at a very low risk of dementia (as recorded by their GP),” said lead researcher and primary care and population health researcher at UCL, Kate Walters. “This could help general practitioners working with people who are anxious about developing dementia.”
To develop the test, the researchers collected the records of 930,395 anonymous patients between the ages of 60 and 95 who showed no signs of dementia, cognitive impairment or memory problems. The data was collected from general practices in The Health Improvement Network (THIN) database, which consists of anonymised electronic medical records (EMRs) collected at Primary Care clinics throughout the UK.
The researchers then analysed the data, examining four variables as possible predictors of dementia risk. These included socio-demographic measures, health and lifestyle measurements, medical diagnoses and the use of prescription medication.
Over five years of follow-up visits by the patients to their GPs, the researchers observed these variables to monitor their association with newly recorded dementia diagnoses.
Putting what they learned into the algorithm, the researchers tested their Dementia Risk Score evaluator against a new group of 264,224 patients, which they assessed in two separate subgroups based on age – those aged 60 to 79 and those aged 80 to 95.
The researchers found that for people aged 60 to 79, the algorithm performed well, while it was less accurate for people between 80 and 95 due to the risk of dementia increasing sharply, skewing the algorithm.
“Our algorithm can be added to clinical software systems and a practice could, for example, run this risk model on all eligible people and offer those at risk more detailed testing or specific preventive management,” said Walters in an interview with The Telegraph.
“Before this score is widely used we would recommend that it is independently tested in further populations of people, and that the ethical implications of using it in practice are considered,” concluded Walters.