OBJECTIVES: General practitioners (GPs) play a critical role in the early recognition and management of dementia. Timely diagnosis is important in light of potential disease-modifying therapies and to improve patient outcomes. We aimed to establish a real-world data cohort utilizing GP data of individuals with dementia as a starting point, with the goal of gaining insights into trajectories and management of dementia and patient outcomes. Here we describe the rationale and design.

METHODS: We selected individuals with dementia using Dutch GP data from the PHARMO Data Network, which includes diagnoses, symptoms, examinations, prescriptions, and communication between GPs and specialists. Diagnosis of dementia was defined as a diagnosis or prescription of anti-dementia drugs between 2011 and 2020. We described the cohort in terms of demographics and screening tests for cognitive impairment.

RESULTS: 52,911 individuals with dementia were selected from a population of 4.7 million persons. Mean age was 81 years (standard deviation [SD] = 8.65) and 59% were female. On average, patients have 8.6 years (SD = 4.09) of data available prior to diagnosis and can be followed-up for 2.8 years (SD = 2.19). Reason for end of follow-up was death for 32%, end of data availability for 40%, and 28% reached December 2020 (i.e., active). We found the Mini-Mental State Examination (MMSE) in GP records of 60% of all persons, and the Montreal Cognitive Assessment (MoCA) and Rowland Universal Dementia Assessment Scale (RUDAS) for only 3%, and <0.5%, respectively.

CONCLUSIONS: We created a cohort of 52,991 individuals with dementia, providing a starting point for further research on trajectories and management of dementia and patient outcomes. Next steps include matching the cohort with dementia-free controls, enriching the cohort by established linkages to other data sources (e.g. hospital data), and examining healthcare resource utilization, indicators of cognitive decline, treatment, and young-onset dementia.