Highlight
– In a SEER–Medicare cohort of 315,089 patients aged ≥66 who died 2015–2020, 7.6% received systemic anticancer therapy (SACT) within 30 days of death.
– Any SACT at end of life (EOL) was associated with markedly increased odds of emergency department visits (OR 3.05), hospital admission (OR 2.64), ICU admission (OR 1.78), and inpatient death (OR 2.02), and with lower hospice utilization (OR 0.51).
Background
Quality of end-of-life (EOL) cancer care is commonly measured by markers such as hospital and emergency department (ED) use, intensive care unit (ICU) admissions, inpatient death, and hospice enrollment. Historically, administration of cytotoxic chemotherapy very late in the disease trajectory has been linked to worse quality of life, higher health care utilization, and lower hospice use. Over the last decade the therapeutic landscape shifted: cytotoxic chemotherapy use at EOL has decreased in many settings while the use of targeted agents and immune checkpoint inhibitors has risen. These novel agents have distinct toxicity profiles, varying time to benefit, and different goals of care implications. Whether late-life use of these newer agents is associated with the same markers of aggressive EOL care as cytotoxic chemotherapy has been less well studied at the population level.
Study design
The study by Canavan and colleagues used SEER–Medicare linked data to examine associations between systemic anticancer therapy (SACT) administered within 30 days of death and health care utilization in older adults. Inclusion criteria were beneficiaries aged ≥66 with Part D coverage (excluding Medicare Advantage), diagnosed with a broad set of cancers (breast, colorectal, lung, prostate, bladder, cervical, kidney, liver, ovarian, pancreatic, melanoma, uterine) between 2005 and 2019, who died between 2015 and 2020. The primary exposure was receipt of any SACT within 30 days of death, with SACT further categorized as cytotoxic chemotherapy, targeted therapy, immunotherapy, or combination regimens. Outcomes were health care utilization in the last 30 days of life—ED visits, hospital admissions, ICU admissions, inpatient death—and hospice utilization. Multivariable regression adjusted for sociodemographic and cancer covariates to estimate adjusted odds ratios (ORs) for outcomes associated with SACT use.
Key findings
Overall prevalence and types of SACT at EOL
Among 315,089 beneficiaries, 23,970 (7.6%) received SACT within 30 days of death. Distribution by therapy type among those receiving SACT was: cytotoxic therapy 50.6%, immunotherapy 20.8%, targeted therapy 18.0%, and combination therapies 10.6%. These data indicate that although cytotoxic chemotherapy remained the single largest category, a meaningful proportion of late-life treatment consisted of targeted agents and immune checkpoint inhibitors.
Associations with acute-care use and hospice
After multivariable adjustment, receipt of any SACT within 30 days of death was associated with substantially higher odds of multiple markers of acute, high-intensity care:
- Emergency department visits: OR 3.05 (95% CI, 2.95–3.15)
- Hospital admissions: OR 2.64 (95% CI, 2.56–2.72)
- ICU admissions: OR 1.78 (95% CI, 1.72–1.83)
- Inpatient death: OR 2.02 (95% CI, 1.96–2.08)
- Lower hospice use: OR 0.51 (95% CI, 0.50–0.53)
Importantly, each SACT subtype—cytotoxic chemotherapy, targeted therapy, immunotherapy, and combinations—was individually associated with higher acute-care utilization and lower hospice enrollment (P < .001 for subtype comparisons). Thus, the association was not limited to conventional chemotherapy; newer agents shared similar population-level associations with markers of aggressive or lower-quality EOL care.
Interpretation and effect sizes
The magnitude of association is notable. An adjusted OR of ~3 for ED visits and ~2.6 for hospital admissions suggests that, even after accounting for patient and disease covariates, SACT administered within 30 days of death is a strong marker—if not a direct driver—of increased acute-care needs. The approximately 50% reduction in odds of hospice use among those receiving late SACT highlights a major care trade-off: ongoing disease-directed therapy at the very end of life appears to be accompanied by substantially lower hospice uptake.
Expert commentary
Clinical implications
These findings have immediate relevance for clinical decision-making, quality measurement, and system-level policies. Clinicians should explicitly discuss likely benefits, harms, and alternatives (including palliative care and hospice) when considering systemic therapy for patients with limited life expectancy. Shared decision-making should incorporate realistic timelines for benefit and the probability of adverse events that may precipitate ED visits, hospitalizations, or ICU care.
Biological plausibility and mechanism
Multiple mechanisms could explain higher acute-care use after late SACT. Treatment-related toxicities (eg, neutropenic fever from chemotherapy, immune-related adverse events from checkpoint inhibitors, or organ-specific toxicities from targeted agents) can cause emergency presentations and hospitalization. In addition, receipt of SACT often reflects a clinical trajectory in which disease-directed care is prioritized over palliation—potentially delaying hospice referral and contributing to increased inpatient deaths. Finally, frequent clinic visits for infusions or monitoring may create more opportunities to escalate care when clinical deterioration occurs.
Limitations and alternative explanations
The observational design precludes definitive causal inference. Residual confounding is possible: patients selected for late SACT may have been younger, had fewer comorbidities, or had expectations/preferences for aggressive care not captured in claims data. Conversely, some patients may have had acute complications from progressing disease that prompted both SACT and subsequent hospitalization. The cohort excluded Medicare Advantage enrollees and required Part D coverage; findings may not generalize to those groups or to younger patients. Claims data lack granular information on performance status, goals-of-care conversations, lines of therapy, dosing intensity, or clinician rationale, all of which would clarify appropriateness of late SACT. Finally, the analysis focuses on associations in the last 30 days of life; different thresholds (eg, 14 days, 60 days) may produce different associations and decision points.
Consistency with prior literature
These results extend prior observations linking late chemotherapy to aggressive EOL care by showing that targeted therapies and immunotherapies share similar population-level associations. The work aligns conceptually with randomized data demonstrating the benefits of early palliative care for symptom control and reduced aggressive care near death (eg, Temel et al., 2010), and with earlier population studies tying late-life systemic therapy to lower hospice use and higher acute-care utilization.
Implications for practice, policy, and research
For clinicians: Foster early, periodic goals-of-care conversations and document prognostic expectations and care preferences. When the probability of meaningful benefit from additional systemic therapy within the relevant time horizon is low, clinicians should discuss de-escalation and timely hospice referral.
For health systems and payers: Quality metrics that benchmark late SACT use and encourage timely hospice enrollment may reduce low-value, high-burden care. Education and decision support tools that help clinicians estimate near-term prognosis and treatment benefit can facilitate appropriate care planning.
For researchers: Prospective studies should examine drivers of late SACT use, including patient preferences, clinician incentives, and the role of molecular testing or expanded indications. Studies that combine claims with clinical data (performance status, symptom burden, molecular markers) and patient-reported outcomes would better delineate when late SACT is appropriate versus potentially harmful. Interventional trials testing decision aids, automatic palliative care triggers, or oncology–palliative care co-management for patients starting new lines of therapy could evaluate whether these approaches reduce acute-care use and improve alignment with patient preferences.
Conclusion
In a large SEER–Medicare cohort of older adults, administration of systemic anticancer therapy within 30 days of death—regardless of modality—was associated with significantly higher acute-care utilization and substantially lower hospice use. These associations support the need for careful prognostic communication, routine incorporation of palliative care, and system-level interventions to align late-stage cancer care with patient-centered goals. Decision-making about systemic therapy near the EOL should explicitly weigh the probability of near-term benefit against the risks of treatment-related complications and the potential for reduced access to hospice.
Funding and clinicaltrials.gov
The primary article reports its funding and conflicts of interest; readers should consult the original publication for details. This SEER–Medicare analysis used existing registry and administrative data and was not a registered interventional trial.
References
1. Canavan ME, Cheng L, Xiang JJ, et al. Association Between Systemic Anticancer Therapy Administration Near the End of Life With Health Care and Hospice Utilization in Older Adults: A SEER Medicare Analysis of End-of-Life Care Quality. J Clin Oncol. 2025 Nov;43(31):3391-3402. doi: 10.1200/JCO-25-00530 IF: 41.9 Q1 .
2. Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non–small-cell lung cancer. N Engl J Med. 2010 Aug 19;363(8):733–742.
3. World Health Organization. Palliative Care. WHO Fact Sheet. 2002. https://www.who.int/health-topics/palliative-care