Access to Specialty Cancer Care and Plan Disenrollment Among Medicare Beneficiaries

Beneficiaries diagnosed with cancer and enrolled in Medicare Advantage (MA) may face barriers to care due to MA plan network design. Given sizeable growth in MA enrollment, it is important to evaluate how well these plans serve patients with high health care needs.

To examine the association between MA network breadth and changes in coverage among beneficiaries with new cancer diagnoses.

This cohort used data from the 2019 to 2020 Surveillance, Epidemiology, and End Results (SEER)-Medicare and 2019 Ideon networks to examine MA plan network inclusion of American College of Surgeons Commission on Cancer-accredited facilities and National Cancer Institute-designated cancer centers. Beneficiaries enrolled in MA who were newly diagnosed with cancer in 2019 were included. Data were analyzed from October 2024 through September 2025.

The explanatory variable of interest was network breadth for these hospitals, measured as a continuous variable. Characteristics of patients in narrow networks (those with fewer than 25% of facilities in the network geographic service area that were in network) and nonnarrow networks were compared.

The main outcome was any plan switch between diagnosis and January 2020, along with the kind of switch (MA to traditional Medicare or new MA plan). Multivariable logistic regression models stratified by plan type (ie, beneficiaries in employer-sponsored and retiree MA plans, plans with premiums, and plans without premiums) were used given potential differences in plan choices for these beneficiary groups. Mean marginal effects were calculated, and coefficients were scaled by 10 percentage points.

Among 24 444 MA beneficiaries, 13 216 individuals had plans with narrow networks (mean [SD] age, 72.2 [7.7] years; 6465 male [48.9%]; 2503 Black [18.9%], 1963 Hispanic [14.9%], and 7563 White [57.2%]) and 11 228 individuals had plans with nonnarrow networks (mean [SD] age, 73.4 [7.7] years; 5638 male [50.2%]; 1613 Black [14.4%], 1346 Hispanic [12.0%], and 7405 White [66.0%]); those in narrow networks were younger and more likely to be Black. A 10-percentage point increase in network breadth was associated with a 4.5-percentage point (95% CI, 5.5 to 3.5 percentage points; P < .001) decrease in the probability of switching to traditional Medicare or a new MA plan for beneficiaries enrolled in a nonemployer plan that charged premiums. It was not associated with enrollment changes among those in employer or retiree plans.

In this study, increased network breadth was not associated with decreased Medicare plan switching for enrollees in employer plans. This finding suggests that employer subsidization of coverage may outweigh concerns about network breadth for patients newly diagnosed with cancer.