Prevalence and outcomes of recorded dementia vary by data source: a population cohort study of 133 407 older adults

Dementia diagnoses are captured across multiple routine data sources, but discrepancies between these may affect both care and research. This study determined the prevalence and overlap of recorded dementia across primary care, hospital and community prescribing data in a UK regional cohort, and examined whether outcomes differed by the setting in which dementia was first recorded.

A retrospective cohort study of adults aged ≥65 years (n = 133 407) in a large Scottish health board. Dementia diagnoses recorded from 1 April 2016 to 1 April 2020 were identified across linked primary care, hospital discharge and prescribing records. Associations between source of first recorded dementia diagnosis and subsequent mortality and emergency hospitalisation were estimated using Cox proportional hazards and Fine-Gray competing-risks models.

At baseline (1 April 2016), 7544/133407 individuals (5.7%) had recorded dementia: 95.1% in primary care, 73.3% in hospital and 54.3% in prescribing records. Over four years, 7359 of the remaining 125,863 individuals (5.8%) had newly recorded dementia: 70.2% in primary care, 22.2% in hospital and 7.6% in prescribing records. Only 35.9% of hospital-recorded diagnoses were coded in primary care within a year or at the time of death, if earlier. People first diagnosed in hospital were older, frailer, more socioeconomically deprived and had higher mortality than those first diagnosed in primary care [<30 days: adjusted hazard ratio (aHR) 8.96, 95% CI 6.94-13.52; >365 days: aHR 1.29, 95% CI 1.19-1.41].

Dementia is variably recorded across routine datasets, and the setting in which dementia is first recorded identifies groups with different prognoses. Improved data integration and scrutiny of hospital-based diagnostic pathways are needed to ensure diagnoses are reliably transferred and people with dementia receive timely, equitable postdiagnostic care.