The association between profitability, clinical benefit, and physicians’ selection of cancer drugs.

11074 Background: Compensation for provider-administered cancer drugs (such as intravenous chemotherapy and immunotherapy) is volume-based and proportional to drug price. Providers receive greater compensation for more expensive drugs. If providers respond to this incentive by selecting more-profitable but clinically less-beneficial treatments, cancer care spending may be increased unnecessarily and care quality may be impacted. The goal of this study was to estimate the association between the billing (profit) margin of cancer treatments and use by oncologists. Methods: This was a population-based cohort study using fee-for-service Medicare claims. We included beneficiaries with an incident cancer diagnosis (a new occurrence of a cancer diagnosis code after a >= 1 year washout period) from 2014-2020. The primary outcome was which cancer treatment, among available options, each patient received. The treatment-level characteristics of interest were provider billing margin (using Medicare reimbursement rates) and clinical benefit (using the National Comprehensive Cancer Network Evidence Blocks scores), both measured coincident with each patient’s diagnosis date. We included cancer treatment “indications” (e.g., metastatic melanoma, adjuvant therapy for stage III colon) that had variation in both the clinical benefit and billing margin of available treatment options. We modeled the association between treatment received, billing margin, and clinical benefit, including inverse probability-of-treatment weights to control for patient (age, comorbidity, frailty, low income subsidy, regional income and poverty prevalence, rurality) and provider (years in practice, academic setting, patient volume) characteristics. Models were estimated within individual cancer indications, and results were then aggregated via meta-analysis. Results: We included 12 cancer indications comprising 19,397 individual patients. Across all treatments for the 12 cancer indications, provider billing margin ranged from $0-$12,692 per course of treatment. There was no association between a $100 increase in provider billing margin and likelihood of treatment use (OR 0.97, 95%CI: 0.91-1.03). Higher clinical benefit was associated with greater treatment use (OR 1.62, 95%CI: 1.15-2.29). These findings were unchanged in sensitivity analyses applying different methods for billing margin calculation, measurement of clinical benefit, and weighting. Conclusions: In this observational study of Medicare beneficiaries, selection of cancer treatments was associated with treatment clinical benefit but not billing margin. Restated, oncologists preferred more beneficial treatments but not more profitable ones. These results suggest that changes in the billing margin of cancer treatments may be unlikely to shift utilization patterns.