Background With increasing cost of healthcare in our aging society, a consistent pain point is that of end-of-life care. It is particularly difficult to prognosticate in non-cancer patients, leading to more healthcare utilisation without improving quality of life. Additionally, older adults do not age homogenously. Hence, we seek to characterise healthcare utilisation in young-old and old-old at the end-of-life. Methods We conducted a single-site retrospective review of decedents under department of Advanced Internal Medicine (AIM) over a year. Young-old is defined as 65–79 years; old-old as 80 years and above. Data collected was demographic characteristics; clinical data including Charlson Comorbidity Index (CCI), FRAIL-NH and advance care planning (ACP); healthcare utilisation including days spent in hospital, hospital admissions, length of stay of terminal admission and clinic visits; and quality of end-of-life care including investigations and symptomatic control. Documentation was individually reviewed for quality of communication. Results One hundred eighty-nine older adult decedents. Old-old decedents were mostly females (63% vs. 42%, p = 0.004), higher CCI scores (7.7 vs 6.6, p = 0.007), similarly frail with lower polypharmacy (62.9% vs 71.9%, p = 0.01). ACP uptake was low in both, old-old 15.9% vs. young-old 17.5%. Poor prognosis was conveyed to family, though conversation did not result in moderating extent of care. Old-old had less healthcare utilisation. Adjusting for sex, multimorbidity and frailty, old-old decedents had 7.3 ± 3.5 less hospital days in their final year. Further adjusting for cognition and residence, old-old had 0.5 ± 0.3 less hospital admissions. When accounted for home care services, old-old spent 2.7 ± 0.8 less hospital days in their last admission. Conclusion There was high healthcare utilisation in older adults, but especially young-old. Enhanced education and goal-setting are needed in the acute care setting. ACP needs to be reinforced in acute care with further research to evaluate if it reduces unnecessary utilisation at end-of-life.

No abstract

No abstract