ERADAR detects primary care patients at risk of having undiagnosed dementia in two real‐world healthcare systems

Abstract

AbstractBackgroundFifty percent of people living with dementia are undiagnosed. The electronic health record (EHR) Risk of Alzheimer’s and Dementia Assessment Rule (eRADAR) was developed to identify older adults at risk of having undiagnosed dementia using routinely collected clinical data. We aimed to externally validate eRADAR in two real‐world healthcare systems, including examining performance over time and by race/ethnicity.MethodsExternal validation was conducted in Kaiser Permanente Washington (KPWA), an integrated health system providing insurance coverage and medical care, and three primary care practices at University of California San Francisco Health (UCSF), an academic medical system. This retrospective cohort study included 688,599 person‐years from 129,315 KPWA members and 42,371 person‐years from 12,567 UCSF patients aged 65 years or older without prior EHR documentation of dementia diagnosis or medication. Analyses validated performance of eRADAR scores, calculated annually from EHR data (including vital signs, diagnoses, medications, and utilization in the prior 2 years), for predicting EHR documentation of incident dementia diagnosis within 12 months.Results7,631 dementia diagnoses were observed at KPWA (11.1 per 1,000 person‐years) and 189 at UCSF (4.5 per 1,000 person‐years). AUC was 0.84 (95% confidence interval: 0.84‐0.85) at KPWA and 0.78 (0.74‐0.81) at UCSF. Using the 90th percentile as the cutpoint, sensitivity was 54% (53‐56%) at KPWA and 45% (38‐52%) at UCSF. All performance measures were similar over time, including across the transition from International Classification of Diseases, version 9 (ICD‐9) to ICD‐10 codes, and across racial/ethnic groups.ConclusionseRADAR showed strong external validity for detecting undiagnosed dementia in two health systems with different patient populations and differential availability of external health care data for risk calculations. In this study, eRADAR demonstrated generalizability from a research sample to real‐world clinical populations, transportability across health systems, robustness to temporal changes in health care, and similar performance across racial/ethnic groups.