Background: Increasing involvement of palliative care generalists may improve access to palliative care. It is unknown, however, if their involvement with and without palliative care specialists are associated with different outcomes. Aim: To describe physician-based models of palliative care and their association with healthcare utilization outcomes including: emergency department visits, acute hospitalizations and intensive care unit (ICU) admissions in last 30 days of life; and, place of death. Design: Population-based retrospective cohort study using linked health administrative data. We used descriptive statistics to compare outcomes across three models (generalist-only palliative care; consultation palliative care, comprising of both generalist and specialist care; and specialist-only palliative care) and conducted a logistic regression for community death. Setting/participants: All adults aged 18–105 who died in Ontario, Canada between April 1, 2012 and March 31, 2017. Results: Of the 231,047 decedents who received palliative services, 40.3% received generalist, 32.3% consultation and 27.4% specialist palliative care. Across models, we noted minimal to modest variation for decedents with at least one emergency department visit (50%–59%), acute hospitalization (64%–69%) or ICU admission (7%–17%), as well as community death (36%–40%). In our adjusted analysis, receipt of a physician home visit was a stronger predictor for increased likelihood of community death (odds ratio 9.6, 95% confidence interval 9.4–9.8) than palliative care model (generalist vs consultation palliative care 2.0, 1.9–2.0). Conclusion: The generalist palliative care model achieved similar healthcare utilization outcomes as consultation and specialist models. Including a physician home visit component in each model may promote community death.
Predictive ability of an expert-defined population segmentation framework for healthcare utilization and mortality - a retrospective cohort study
