Artist's rendering of a brain cell damaged by amyloid beta plaque in Alzheimer's disease

A new tool using electronic health records data could help clinicians avoid missed or late dementia diagnoses in older adults, according to a National Institute of Aging-funded study.

To create the eRADAR tool, investigators examined health records from seniors who participate in an ongoing Kaiser Permanente Washington study. Among 4,330 patients across more than 16,000 clinical visits related to the study, 1,015 visits resulted in a dementia diagnosis. Yet of these diagnoses, nearly half had not been coded in the patients’ EHR records, the study team reported.

The researchers created a model based on the key markers that best predicted undiagnosed dementia among these patients. The final model included markers of dementia-related symptoms such as psychosis diagnoses and antidepressant fills; healthcare utilization patterns such as emergency room visits; and dementia risk factors such as cerebrovascular disease and diabetes. Participants with eRADAR scores in the top 5% were more than five times as likely to have unrecognized dementia than the study’s patient population overall.

The findings suggest that a tool using readily available EHR records could be used in primary care to detect more at-risk patients for referral to needed screening, the investigators concluded. 

The researchers hailed from the University of California, San Francisco; the University of Washington, Seattle; and the Kaiser Permanente Washington Health Research Institute.

Full findings were published in the Journal of the American Geriatrics Society.