Current tests used to predict the risk of dementia require collecting data from patients, so researchers have developed a model that only requires standard information from GP records
The team from University College London developed the model using data from more than 1 million patients from a UK general practice database. It takes into account variables including age and sex, smoking status, blood pressure, medical diagnoses such as diabetes, and prescription medications such as antihypertensives.
The researchers found their model performed well at predicting dementia risk in the next five years for patients aged 60–79 years, but not for older patients.
However, the team suggests that the model may be best suited to ruling out people who are unlikely to develop dementia, sparing them from further investigation, because it had a higher negative predictive value than positive.