Prevalence and Outcomes of Recorded Dementia: a Population Cohort Study of 133,407 Older Adults using Linked Routine Data

2845
RS Penfold1,2, T Wilkinson3, TC Russ3, LE Stirland3,4, C MacRae1, SD Shenkin1,2, A Anand5, E Vardy6,7, B Guthrie1, EL Sampson8,9, AMJ MacLullich2
1. Advanced Care Research Centre, University of Edinburgh; 2. Ageing & Health, Usher Institute, University of Edinburgh; 3. Centre for Clinical Brain Sciences, University of Edinburgh; 4. Global Brain Health Institute, UCSF; 5. Centre for Cardiovascular S
Dementia
Primary and Community Care

Abstract

Introduction:

Recording dementia diagnoses is essential to ensure appropriate post-diagnostic support and care. We examined the prevalence of recorded dementia in different routine datasets and associations with emergency hospitalisation and mortality.

Methods:

This retrospective longitudinal cohort study included all adults ≥65 years registered with a Southeast Scotland GP on 1st April 2016. Dementia diagnoses were identified in primary care, hospital discharge and community prescribing records. New diagnoses were considered from 1st April 2016 to 1st April 2020. All individuals were followed up to 23rd October 2023. Cox proportional hazards and Fine-Gray models were used to estimate associations between recorded dementia and death and emergency hospitalisation, respectively. Diagnosis capture in other datasets was examined, accounting for mortality.

Results:

On 1st April 2016, 7544/133407 (5.7%) individuals had a recorded dementia diagnosis: 1254 (16.6%) in a single dataset, including 940 (12.5%) only in primary care and 279 (3.7%) in hospital data. Between 1st April 2016 to 1st April 2020, 7359/133,407 (5.8%) had a new diagnosis: 5165 (70.2%) first recorded in primary care, 1634 (22.2%)in hospital and 560 (7.6%)in community prescribing data. People with dementia had higher risks of death [adjusted hazard ratio (HR) 2.46 (95% Confidence Interval (CI) 2.39-2.54)] and emergency hospitalisation [adjusted subdistribution HR 1.58 (95%CI 1.56-1.60)] then those without dementia. People with diagnoses first recorded in hospital had higher mortality rates than those with community diagnoses [<30days: aHR 8.96 (95%CI 6.94-13.52); >365days: aHR 1.29 (95%CI 1.19-1.41)]. Only 562 (35.9%) of those with hospital diagnoses had recorded primary care diagnoses within a year.

Conclusions:

Dementia is often recorded in single datasets, sometimes only in hospital data. Dementia is associated with adverse prognosis, with highest mortality in those first diagnosed in hospital. Findings highlight the need for better recording, dataset integration and scrutiny of hospital-based diagnostic pathways to ensure appropriate post-diagnostic support and care.