A randomized controlled trial of a novel home-based palliative care program: A report of a trial that could not be completed

Background In response to a growing need for accessible, efficient, and effective palliative care services, we designed, implemented, and evaluated a novel palliative care at home (PC@H) model for people with serious illness that is centered around a community health worker, a registered nurse, and a social worker, with an advanced practice nurse and a physician for support. Our objectives were to measure the impact of receipt of PC@H on patient symptoms, quality of life, and healthcare utilization and costs. Methods We enrolled 136 patients with serious illness in this parallel, randomized controlled trial. Our primary outcome was change in symptom burden at 6 weeks. Secondary outcomes included change in symptom burden at 3 months, change in quality of life at 6 weeks and 3 months, estimated using a group t-test. In an exploratory aim, we examined the impact of PC@H on healthcare utilization and cost using a generalized linear model. Results PC@H resulted in a greater improvement in patient symptoms at 6 weeks (1.30 score improvement, n = 37) and 3 months (3.14 score improvement, n = 21) compared with controls. There were no differences in healthcare utilization and costs between the two groups. Unfortunately, due to the COVID-19 pandemic and a loss of funding, the trial was not able to be completed as originally intended. Conclusions A palliative care at home model that leverages community health workers, registered nurses, and social workers as the primary deliverers of care may result in improved patient symptoms and quality of life compared with standard care. We did not demonstrate significant differences in healthcare utilization and cost associated with receipt of PC@H, likely due to inability to reach the intended sample size and insufficient statistical power, due to elements beyond the investigators' control such as the COVID-19 public health emergency and changes in grant funding.