Hospital admissions and emergency department visits for people with dementia

**Background:** Previous research has shown that people with dementia (PwD) are more likely to be admitted to hospital, have prolonged hospital stays, and visit emergency departments (ED) compared to people without dementia. Common causes include urinary tract infections, falls, behavioral problems, and lower respiratory infections. The hospital environment can be distressing for PwD and their caregivers, and frequent admissions may increase mortality. This study aimed to assess rates, causes, and predictors of hospital admissions and ED visits in PwD, and to evaluate survival following such healthcare use.

**Methods:** This retrospective study used linked national datasets from Northern Ireland (2010-2016). PwD were identified through the Enhanced Prescribing Database (first dementia medication dispensed), hospital records (Patient Administration System and Symphony system), and death certificates. An initial age- and gender-matched control group was extracted from the Patient Administration System; 1,427 controls who had a dementia cause of death or hospital admission were transferred to the dementia group, yielding 26,875 PwD and 23,961 controls. Demographic data (age, gender, marital status, area of living, deprivation score), care home transitions, hospital admissions, ED visits, and deaths were linked via the Honest Broker Service. Deprivation was measured using the Northern Ireland Multiple Deprivation Measure (scale 1-10, higher = less deprived). Statistical analyses included independent-samples t-tests, chi-square tests, binomial regression models for predictors of admissions/ED visits, and Cox proportional hazard models for survival.

**Key Results:** PwD were significantly older (mean 77.65 vs. 76.85 years, p<0.001) and more likely female (66% vs. 64%, p=0.001). More PwD lived in urban areas (69% vs. 64%, p<0.001). 44% of PwD transitioned to a care home vs. 9% of controls (p<0.001). 50% of PwD died during the study period vs. 34% of controls (p<0.001). PwD had significantly higher odds of hospital admission (OR 1.79, 95% CI 1.62-1.79, p<0.001) and ED visit (OR 1.35, 95% CI 1.29-1.42, p<0.001), but lower odds of readmission (OR 0.60, 95% CI 0.57-0.63, p<0.001). PwD had a lower mean number of admissions (4.65 vs. 5.13, p=0.014) but longer mean hospital stay (18.15 vs. 11.54 days, p<0.001). Common causes for admissions in PwD included lower respiratory tract infection (28.4%), fracture (21.1%), urinary tract infection (20.8%), syncope/collapse (8.6%), and cataract (11.5%). For ED visits, common causes were fracture (13.5%), head injury (8.2%), urinary tract infection (6.3%), lower respiratory infection (5.4%), and soft tissue injury (5.1%). Predictors for hospital admissions in PwD: female gender (OR 0.68, p<0.001) and less deprivation (OR 0.96, p<0.001) reduced odds; urban living (OR 1.13, p=0.004), widowed (OR 1.37, p=0.002), and care home residence (OR 1.25, p<0.001) increased odds. For readmissions: female gender (OR 0.71, p<0.001) and less deprivation (OR 0.96, p<0.001) reduced odds; urban living (OR 1.19, p<0.001), widowed (OR 1.27, p<0.001), and care home (OR 1.30, p<0.001) increased odds. For ED visits: older age (OR 0.97, p<0.001) and less deprivation (OR 0.98, p<0.001) reduced odds; widowed (OR 1.27, p<0.001) and care home (OR 1.35, p<0.001) increased odds. Survival analysis: PwD with at least one hospital admission had higher mortality (HR 3.074, 95% CI 2.98-3.172, p<0.001) vs. controls. However, PwD with 1-20 admissions had lower mortality (HR 0.92, p<0.001) and those with 21-40 admissions had HR 0.76 (p=0.014) compared to PwD never hospitalized. PwD with 1-5 days hospital stay had lower mortality (HR 0.81, p<0.001), but stays ≥6 days increased mortality. ED visits were associated with lower mortality in PwD (e.g., 1-2 visits: HR 0.53, p<0.001). Within 6 months of admission, 30.04% of PwD died vs. 26.04% of controls (p<0.001).

**Clinical Implications:** This study confirms that PwD have higher rates of hospital and ED use, but surprisingly, moderate numbers of admissions (up to 40) and short stays (1-5 days) were associated with lower mortality in PwD, suggesting that appropriate hospital care may be beneficial. Key predictors for increased use include care home residence, widowhood, and urban living, while female gender and less deprivation reduce use. These findings highlight the need for policies that support hospital and ED access for PwD, especially those living alone or in rural areas, while also developing alternative care models (e.g., enhanced primary care, eHealth) to reduce unnecessary admissions and associated distress and costs.