Background Guidelines suggest that the preferred place of care (PPC) for patients at the end of life is in their own home. Existing literature is largely from younger cancer patients. Increasing numbers are living into very old age who may have different care needs and challenges. Studies on PPC and prognostication outside the ‘acute’ dying phase in older people are limited. We aim to investigate the casemix (cancer vs non-cancer), the PPC and accuracy of prognostication of referrals to the Palliative Discharge Team (PDT).

Methods Analysis of observational database data collected as part of routine clinical practice of the PDT - referred inpatients thought to be in the last 3 months of life. The data were analysed using Stata 14.

Results n=987. Mean age at referral 78 years. 60.2% had palliative cancer diagnoses. The odds ratio of cancer diagnosis decreased with increasing age (OR 0.957, 95% CI 0.944–0.971; p≤0.001).

Home was the PPC in 34% patients. Logistic regression analysis found an increased likelihood of change in PPC with age (OR 1.03, 95% CI 1.02–1.04; p≤0.

Mean time from hospital discharge to death was 47 days. 90% of deaths occurred <109 days. No statistically significant difference in time from discharge to death with age or cancer vs non-cancer diagnosis (p=0.1684).

Discussion Home was not the PPC for the majority of patients and the association of changing PPC with older age and non-cancer diagnosis suggests this group may have different wishes from previous study participants. This is likely to be multifactorial, with different barriers to dying at home in an older population. Cancer dominance of referrals was less prominent in the oldest old. Prognostication was not significantly affected by cancer status and the accuracy suggests underuse of the service. Additional research is required into PPC in older, multi-morbid populations and what factors affect it.