Geographic variations in end-of-life hospitalisations for patients with mental illness: a population-based observational study in England, UK.

Background: High rates of hospital admissions have been reported for patients with mental illness, but less is known about factors associated with multiple hospitalisations at the end-of-life in this group.

Aim: To describe the geographical variations in end-of-life hospitalisations and examine factors associated with multiple hospitalisations (≥ 2) in the last 90 days of life for people with mental illness.

Methods: A national population-based observational study in England UK using a linkage of Hospital Episode Statistics Admitted Patient Care (HES-APC) and the Office for National Statistics (ONS) death registry data. Our cohort comprised patients aged 18 and over, who died in England between 2018-04-01 and 2019-03-31 with HES-APC diagnoses of (1) Schizotypal, delusional disorders or schizophrenia, (2) schizoaffective or bipolar affective disorder, (3) substance use disorders; or (4) depressive episodes or recurrent depressive disorders. Geographic variations of end-of-life hospitalisations for each diagnostic group were described across National Health Services (NHS) regions. Modified Poisson regression models were used to estimate factors associated with multiple end-of-life hospitalisations in each diagnostic group.

Results: A total of 49,775 patients with mental illness died in the year 2018-2019, of whom 50.2% (n = 25,004) had multiple end-of-life hospitalisation in the last 90 days of life. Factors positively associated with multiple end-of-life hospitalisations included older age, being resident in an urban area, cancer related deaths, and, for patients with depressive disorders, higher socioeconomic deprivation.

Conclusion: Strengthening primary and community care services targeted at older adults with cancer could potentially reduce multiple end-of-life hospitalisations for patients with mental illness.