Patterns of emergency department attendance among older people in the last three months of life and factors associated with frequent attendance: a mortality follow-back survey

Background: frequent emergency department (ED) attendance at the end of life disrupts care continuity and contradicts most patients' preference for home-based care. Objective: to examine factors associated with frequent (>= 3) end of life ED attendances among older people to identify opportunities to improve care. Methods: pooled data from two mortality follow-back surveys in England. Respondents were family members of people aged >= 65 who died four to ten months previously. We used multivariable modified Poisson regression to examine illness, service and sociodemographic factors associated with >= 3 ED attendances, and directed content analysis to explore free-text responses. Results: 688 respondents (responses from 42.0%); most were sons/daughters (60.5%). Mean age at death was 85 years. 36.5% had a primary diagnosis of cancer and 16.3% respiratory disease. 80/661 (12.1%) attended ED >= 3 times, accounting for 43% of all end of life attendances. From the multivariable model, respiratory disease (reference cancer) and >= 2 comorbidities (reference 0) were associated with frequent ED attendance (adjusted prevalence ratio 2.12, 95% CI 1.21-3.71 and 1.81, 1.07-3.06). Those with >= 7 community nursing contacts (reference 0 contacts) were more likely to frequently attend ED (2.65, 1.49-4.72), whereas those identifying a key health professional were less likely (0.58, 0.37-0.88). Analysis of free-text found inadequate community support, lack of coordinated care and untimely hospital discharge were key issues. Conclusions: assigning a key health professional to older people at increased risk of frequent end of life ED attendance, e.g. those with respiratory disease and/or multiple comorbidities, may reduce ED attendances by improving care coordination.