The ABOARD-RWD infrastructure: Using linked real-world data to study the diagnosis and management of dementia

BackgroundResearch into Alzheimer's disease has primarily focused on clinical cohorts, yet many patients are diagnosed and followed in primary care. An infrastructure comprising routinely collected and linked electronic health records from primary care, hospitals, and pharmacies, i.e., real-world data (RWD), can provide insight into disease trajectories.ObjectiveWe aim to describe the ABOARD-RWD infrastructure, focusing on its design, data sources, representativeness, and potential applications.MethodsThe ABOARD-RWD infrastructure includes 79,114 Dutch adults with incident dementia (2011-2022), identified through dementia diagnoses, free-text mentions, or use of dementia-related drugs. Linked data includes general practitioner (GP), hospital care, outpatient pharmacy, laboratories, and specialized registries. Descriptive analyses assessed demographics, prevalence, and clinical characteristics.ResultsThe ABOARD-RWD Infrastructure (age 81 ± 9, 59% female) has a median 8.6-year look-back period (IQR 4.7-12.3) and 2.2-year follow-up (IQR 0.9-4.2) from first dementia diagnosis. Dementia was recorded in GP for 86%, hospitals for 34%, and 22% received dementia-related drugs. Dementia prevalence in 2022 was 9.3 per 1,000 (10.9 women, 7.6 men), aligning with national estimates. Mini-Mental State Examinations were recorded for 62%: 35% scored ≥24, 26% scored 18-23, 13% scored <18, and 25% had an unknown score. Cardiovascular disease was recorded in 48%, hypertension in 51%, diabetes in 23%, hypercholesterolemia in 19%, cerebrovascular disease in 20%, asthma/chronic obstructive pulmonary disease in 14%, psychiatric conditions in 17%, and sleep disturbances in 13%.ConclusionsThe ABOARD-RWD Infrastructure enables real-world, longitudinal dementia research across multiple healthcare settings. It captures representative patient journeys, supporting studies on risk factors, diagnosis, management, and outcomes.