Plan-level and provider-level variation in systemic therapy receipt among Medicare Advantage beneficiaries with advanced cancer.

Abstract

79 Background: Prompt initiation of systemic therapy is a critical determinant of survival in advanced cancer. Medicare Advantage (MA) plans commonly employ prior authorization, which may affect timely systemic therapy initiation. Moreover, healthcare providers may differ in their capacity to navigate these requirements. Yet little is known about how systemic therapy initiation varies across MA plans and provider organizations. Methods: Using the SEER–Medicare linked data, we identified Medicare beneficiaries aged ≥66 years diagnosed with distant-stage breast, colorectal, lung, pancreatic, or prostate cancer in 2016-2021 who survived ≥3 months and were continuously enrolled in Traditional Medicare (TM) or MA through death/end of follow-up. The primary outcome was systemic therapy initiation ≤3 months of diagnosis. We estimated adjusted probabilities of initiation across MA plan types (HMO, PPO, POS vs. TM) overall and by cancer prognosis, controlling for clinical and demographic characteristics. To quantify multilevel variation, we restricted to MA plans and treating organizations (Tax Identification Numbers, TINs) with ≥25 patients and used multilevel mixed-effects logistic regression with plan and TIN random effects to estimate median odds ratio (MOR) at each level after patient-level adjustment. MOR estimates how much the median odds of initiation would change due to provider or plan alone; a larger MOR indicates greater between-provider or between-plan variation. Results: Among 88,228 beneficiaries with advanced cancer, 40% were MA enrollees, covered by 470 plans and treated by 3,070 TINs. Adjusted probabilities of systemic therapy initiation were 76.2% (TM) and ranged from 71.3% to 76.2% across MA plan types; such differences were larger for poor-prognosis cancers (lung, pancreatic; MA-HMO vs. TM, 61.5% vs. 68.8%) than for better-prognosis cancers (breast, colorectal, prostate; MA-HMO vs. TM, 75.8% vs. 78.2%). Adjusted probabilities of initiation ranged from 46.5% to 88.1% across MA plans and more widely across provider organizations (42.3%–96.4%). TIN-level MOR was larger than plan-level MOR (1.35 vs. 1.23), meaning that for two otherwise-identical patients, the median odds of initiation differed by 35% in two randomly selected providers, vs 23% across randomly selected plans. Conclusions: Systemic therapy initiation was lower among MA than TM beneficiaries, particularly for those under MA-HMO plans with poor-prognosis cancers. The larger variation across providers than MA plans may reflect differential capacity in navigating prior authorization or other organization-specific factors. Identifying modifiable provider-level factors is needed to improve timely treatment access for MA beneficiaries with advanced cancer.