Reducing costs at the end of life through provider incentives for hospice care: A retrospective cohort study

Bradley Chen, Chin Chi Kuo, Nicole Huang, Victoria Y. Fan*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

10 Scopus citations


Background: Costs of medical care have been found to be highest at the end of life. Aim: To evaluate the effect of provider reimbursement for hospice care on end-of-life costs. Design: The policy expanded access to hospice care for end-stage renal disease patients, a policy previously limited to cancer patients only. This study employed a difference-in-differences analysis using a generalized linear model. The main outcome is inpatient expenditures in the last 30 days of life. Setting/participants: A cohort of 151,509 patients with chronic kidney disease or cancer, aged 65 years or older, who died between 2005 and 2012 in the National Health Insurance Research Database, which contains all enrollment and inpatient claims data for Taiwan. Results: Even as end-of-life costs for cancer are declining over time, expanding hospice care benefits to end-stage renal disease patients is associated with an additional reduction of 7.3% in end-of-life costs per decedent, holding constant patient and provider characteristics. On average, end-of-life costs are also high for end-stage renal disease (1.88 times higher than those for cancer). The cost savings were larger among older patients—among those who died at 80 years of age or higher, the cost reduction was 9.8%. Conclusion: By expanding hospice care benefits through a provider reimbursement policy, significant costs at the end of life were saved.

Original languageEnglish
Pages (from-to)1389-1400
Number of pages12
JournalPalliative Medicine
Issue number8
StatePublished - 1 Sep 2018


  • end-stage renal disease
  • Hospice care
  • palliative care
  • physician incentives


Dive into the research topics of 'Reducing costs at the end of life through provider incentives for hospice care: A retrospective cohort study'. Together they form a unique fingerprint.

Cite this