Impact of palliative care on hospital length of stay and charges in hospitalized patients with cancer at end of life.

Tien-Chan Hsieh Division of Hematology-Oncology, Department of Medicine, UMass Chan Medical School, Worcester, MA info_outline Tien-Chan Hsieh, Guangchen Zou

Authors Tien-Chan Hsieh Division of Hematology-Oncology, Department of Medicine, UMass Chan Medical School, Worcester, MA info_outline Tien-Chan Hsieh, Guangchen Zou Organizations Division of Hematology-Oncology, Department of Medicine, UMass Chan Medical School, Worcester, MA, Johns Hopkins Medicine, Baltimore, MD Abstract Disclosures Research Funding No funding sources reported Background: Early palliative care has been shown to enhance the quality of life and survival for cancer patients. However, the economic implications of using palliative care in hospitalized cancer patients remain inconclusive. While palliative care can prevent aggressive interventions in end-of-life patients, it may also prolong hospitalization. Methods: This retrospective study used four years of National Inpatient Sample data (2016-2019) to explore the association between palliative care and hospital length of stay and charges among cancer patients aged 18 and above, hospitalized for at least 7 days and died. Hospital charges were divided by length of stay to derive the average daily charges. Continuous variables, length of stay and average daily charges, were standardized by subtracting means and dividing by standard deviations. We top-coded length of stay and average daily charges at the 99th percentile. Length of stay was also modeled into categories: 7-13, 14-20, 21-27, and 28+ days. A mixed-effects model with individual hospitals as a random effect was employed to account for potential variations in clinical practices and billing at the institutional level. Results: The study included 59,355 distinct cancer hospitalizations, totaling 296,775 weighted hospitalizations. Palliative care utilization was observed in 54.0%, and Do-not-resuscitate (DNR) status was noted in 57.9%. The observed palliative care use and DNR status were most prevalent in the 7-13 days group (56.6% and 61.3%) and least in the 28+ days group (46.9% and 48.8%). 85.2% didn’t receive cardiopulmonary resuscitation (CPR) and that rate was similar across all length of stay groups. In linear mixed-effects models using standardized length of stay, palliative use was associated with shorter length of stay (adjusted Odds Ratio {aOR} 0.89; confidence interval {CI} 0.88-0.90). After adjustment for covariates, DNR was also associated with shorter hospitalization (aOR: 0.81; CI: 0.80-0.82). Receiving CPR had lower adjusted risk for longer hospitalization (aOR: 0.94; CI: 0.93-0.95). In multivariate model, daily charges were inversely correlated to the length of stay (aOR: 0.92; CI: 0.92-0.92). When using average charges per day as dependent variable, palliative care use was associated with less charges (aOR: 0.83; CI: 0.83-0.84). DNR also showed a lower adjusted odds for hospital charges (aOR: 0.89; CI: 0.88-0.89). Receiving CPR had higher adjusted risk of higher charges (aOR: 1.23; CI: 1.22-1.24). Conclusions: The use of palliative care was associated with a shorter length of stay and lower hospital charges per day among cancer patients who died in the hospital after adjusting for covariates. Many of the 46% who died without palliative care could have benefitted from it. Interventions aimed at promoting palliative care services among inpatient cancer patients nearing the end of life might yield economic benefits.