Life‐Prolonging Treatment Preferences and Their Association With Health Care Utilization and End‐Of‐Life Experiences in Older Adults

TL;DR

This study examines how older adults' preferences for life-prolonging treatments relate to their healthcare use and end-of-life experiences.

Contribution

The study reveals that hypothetical preferences for life-prolonging treatments do not significantly influence actual healthcare utilization or end-of-life outcomes.

Findings

Preferences for life-prolonging treatments in hypothetical scenarios were not associated with hospitalization, ICU visits, or costs.

Those accepting treatment for pain spent more time away from home, but preferences did not affect hospice use or EOL care quality.

Accepting life-prolonging treatments was linked to a higher chance of dying in a hospital.

Abstract

People markedly differ in their preferences for life‐prolonging treatments (LPT). We explored the association between LPT preferences and healthcare utilization and end‐of‐life (EOL) experiences.

We conducted a retrospective cohort study using data from 5373 older adults in the National Health and Aging Trends Study (NHATS) linked to Medicare/Medicaid. Among them, 1564 died, and 1124 had proxies who completed the Last Month of Life module. We categorized LPT preferences (accept or reject) in severe disability or pain scenarios, and we assessed healthcare utilization (hospitalization, ICU visits, receipt of LPT, hospice), total cost, and days out of home in the last year of life, place of death, and EOL experiences.

The average age of respondents was 76.7 years (SD 5.9), 55.1% female, and 7.8% Black people. Sixty‐nine percent would reject all LPT; 5.3% would accept LPT only in the severe disability scenario; 15.8% would accept LPT only in the pain scenario; and 9.9% in both. There was no association between LPT preferences and hospitalization, ICU visits, receipt of LPT, or costs. Respondents accepting LPT for pain spent more time away from home (3.4 days, 95% CI: 0.3, 6.6, p = 0.03). Among decedents, preferences were not linked to receipt of LPT, hospice use, or EOL care quality. Those accepting LPT for severe disability and pain had higher odds of dying in a hospital (OR = 1.8, 95% CI 1.1–3, p < 0.03) compared to those who would reject LPT.

LPT preferences established in hypothetical scenarios were not associated with health care utilization, cost, or perceived quality of end‐of‐life. As LPT preferences should influence care, subsequent work, such as understanding preference measurement, real‐world preference changes, or the limited opportunity for preferences to influence care, is warranted.