Hospice Utilization in the United States: A Prospective Cohort Study Comparing Cancer and Noncancer Deaths

John G. Cagle, Joonyup Lee, Katherine A. Ornstein, Jack M. Guralnik

Research output: Contribution to journalArticlepeer-review

20 Scopus citations


OBJECTIVES: Reliable national estimates of hospice use and underuse are needed. Additionally, drivers of hospice use in the United States are poorly understood, especially among noncancer populations. Thus the objectives of this study were to (1) provide reliable estimates of hospice use among adults in the United States; and (2) identify factors predicting use among decedents and within subsamples of cancer and noncancer deaths. DESIGN: We conducted a prospective cohort study using the Health and Retirement Study survey. Excluding sudden deaths, we used data from the 2012 survey wave to predict hospice use in general, and then separately for cancer and non-cancer deaths. SETTING: Study data were provided by a population-based sample of older adults from the U.S. PARTICIPANTS: We constructed a sample of 1,209 participants who died between the 2012 and 2014 survey waves. MEASUREMENTS: Hospice utilization was reported by proxy. Exposure variables included demographics, functionality (activities of daily living [ADLs]), health, depression, dementia, advance directives, nursing home residency, and cause of death. RESULTS: Hospice utilization rate was 52.4% for the sample with 70.8% for cancer deaths and 45.4% for noncancer deaths. Fully adjusted model results showed being older (odds ratio [OR] = 1.54), less healthy (OR =.79), having dementia (OR = 1.52), and having cancer (OR = 5.47) were linked to greater odds of receiving hospice. Among cancer deaths, being older (OR = 1.64) and female (OR = 2.54) were the only predictors of hospice use. Among noncancer deaths, increased age (OR = 1.58), more education (OR = 1.56), being widowed (OR = 1.55), needing help with ADLs (OR = 1.13), and poor health (OR =.77) were associated with hospice utilization. CONCLUSION: Findings suggest hospice remains underutilized, especially among individuals with noncancer illness. Extrapolating results to the US population, we estimate that annually nearly a million individuals who are likely eligible for hospice die without its services. Most (84%) of these decedents have a noncancer condition. Interventions are needed to increase appropriate hospice utilization, particularly in noncancer care settings. J Am Geriatr Soc 68:783–793, 2020.

Original languageEnglish
Pages (from-to)783-793
Number of pages11
JournalJournal of the American Geriatrics Society
Issue number4
StatePublished - 1 Apr 2020


  • Health and Retirement Study
  • end-of-life care
  • healthcare utilization
  • hospice


Dive into the research topics of 'Hospice Utilization in the United States: A Prospective Cohort Study Comparing Cancer and Noncancer Deaths'. Together they form a unique fingerprint.

Cite this