Telehealth use at the end of life among Medicare decedents with cancers.

e23120 Background: Telehealth use has surged since the COVID-19 pandemic. For individuals with cancer approaching end-of-life (EOL) facing high symptom burden, declining mobility, and increasing reliance on caregivers, telehealth may enhance access to specialty and palliative care services. National evidence comparing EOL telehealth use between cancer and non-cancer populations remains limited. Methods: We conducted a retrospective, population-based cohort study of telehealth use using 100% Traditional Medicare (TM) and Medicare Advantage (MA) claims. We included Medicare beneficiaries died in 2019-2023 with continuous TM or MA enrollment in the 12 months preceding death. Patients with cancer were identified by 1 inpatient or 2 outpatient visits with cancer diagnosis codes in the last year of life. Telehealth use in the last year of life was identified using procedure codes and modifiers and classified into synchronous patient–provider (audio-only and audio–video), asynchronous, and remote physiologic or therapeutic monitoring (RPM/RTM) encounters. Telehealth use, overall and by service type, was compared between Medicare decedents with and without cancer; multivariable logistic regression examined patient characteristics associated with telehealth use after COVID-19 onset. Results: Among a total of 10,767,407 Medicare decedents, average age was 81.56, with 51.8% female, 9.7% Non-Hispanic Black, 2.3% Hispanic, and 33.1% with cancer diagnosis. Telehealth use among decedents with cancer increased from 2.7% to 47.8% in 2019-2023, peaking in 2021 (70.0%), which was persistently higher than those without cancer (2.4% to 34.4% in 2019-2023). Synchronous video encounters (1.2% to 41.2%) were most common, followed by synchronous audio only encounters (1.6% to 20.9%) in 2019-2023. RTM/RPM was limited but increased from 0.1% to 2.0% in 2019-2023. Decedents with cancer were consistently more likely to use all telehealth modalities than those without cancer. In adjusted analysis, older (85+ vs. 66-74 years: -6.10 percentage points [ppts]), Non-Hispanic Black (vs. Non-Hispanic White: -3.27 ppts), and beneficiaries in counties with a higher social vulnerability index (Q4 vs. Q1: -0.92 ppts, p-values<.001) were less likely to use telehealth. More comorbidities (14.70 ppts), multiple hospitalizations in the prior year (5.18 ppts), and counties with better broadband internet access (80%+ households with 25+ Mbps internet connection vs. ≤40%: 13.98 ppts, p-values<.001) were associated with higher likelihood of telehealth use. Conclusions: EOL telehealth use differs meaningfully between cancer and non-cancer decedents, which may reflect greater care intensity and more frequent clinical contact. Telehealth use was lower among older, racial and ethnic minority beneficiaries, and those living in more socially vulnerable areas, and higher among patients with greater medical complexity and better broadband access.