Health and social care costs at the end of life: a cross-sectional matched analysis of linked patient records in East London

2019 
Abstract Background After a decade of austerity, controlling costs in the health system while addressing the needs of a growing population is one of the most pressing health policy objectives in England. Care in the final year of life accounts for 10% of inpatient hospital costs in England, but there has been scarce analysis of end-of-life costs in other care settings. We aimed to investigate the publicly funded costs associated with end-of-life care across different health and social care settings. Methods We did a cross-sectional analysis of linked electronic health records of residents aged over 50 years in a borough in East London, between 2011 and 2017. Individuals who died during the study period were matched to survivors on age group, sex, deprivation status by area of residence, number of long-term conditions, and time period. We calculated costs in the final year of life as the mean difference in costs between paired survivors and decedents, with bootstrapped CIs and significance testing via the Wilcoxon signed-rank test. We further disaggregated costs by care setting, age, and months to death. Study approval was provided by the Barking and Dagenham, Havering, and Redbridge Information Governance Steering Committee. Findings Across 8720 matched patients, the final year of life was associated with an increase in mean costs of £7450 (95% CI 7086–7842, p Interpretation There is a large increase in costs in the final year of life, most of which relate to unplanned hospital care, suggesting that better end-of-life planning might lead to cost savings. Social care costs, unlike health-care costs, increase with age after accounting for time-to-death. As such, an ageing population places more pressure on the social care system. Funding None.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []