340b hospital participation and initial oncology treatment choice.

Abstract

29 Background: The 340b Drug Pricing Program, originally intended to support safety net providers, allows qualifying providers to purchase drugs at a discounted price. Eligibility criteria and participation have continued to expand; notably, hospitals can qualify if they have a “disproportionate share” of Medicaid and Supplementary Security Income patients. 340b participation may motivate increased prescribing of high-priced drugs to capture revenue from the discount. Such practices may be counter to care quality because more costly treatments are not always clinically preferred. We therefore studied the association between hospital 340b participation and use of guideline-preferred treatments, an important measure of cancer care quality. Methods: We analyzed a 100% Medicare fee-for-service sample from 2014 through 2020. Included beneficiaries had an inferred new cancer diagnosis (after at least 1 year of continuous enrollment with no cancer claims) and met eligibility criteria for 1 of 12 NCCN initial systemic therapy scenarios across 8 cancer sites. Chosen scenarios showed substantial variability in treatment quality and cost. Beneficiaries were attributed to a site of care using the plurality of oncology medical services claims, and sites were classified as 340b by linking to Health Resources and Services Administration data. Outcomes were a) receipt of any NCCN-recommended treatment within 180 days of inferred diagnosis date (yes or no), b) receipt of the most highly recommended treatment (yes or no), and c) relative clinical benefit of the received treatment as a continuous measure (ranging 0-100). Most highly recommended treatment and relative clinical benefit were calculated using NCCN “preferred” designation and Evidence Blocks scores. We accounted for patient and hospital covariates via augmented inverse propensity weighting with bootstrap confidence intervals. Results: Of 27,076 eligible beneficiaries, 26% could reliably be assigned a 340b hospital or comparable non-340b hospital. Of these, 5,648 (81%) received care at 340b hospitals, and 1,314 (19%) received care at non-340b hospitals (controls). After covariate adjustment, beneficiaries at 340b hospitals, when compared to those at non-340b hospitals, had a similar likelihood of receiving any treatment (risk difference of -1.5%, 95% CI (-5.8%, 2.7%)) and had, for those who received treatment, similar relative clinical benefit (difference of -0.4, 95% CI (-4.0, 3.3)). 340b beneficiaries were slightly less likely to receive the highest-recommended treatment, but this was not statistically significant (risk difference of -4.1%, 95% CI (-9.9%, 1.7%)). Conclusions: We did not find evidence of a difference in use of NCCN-preferred treatments between 340B and non-340B hospitals. Not considering potential impacts on patient mix or other hospital characteristics, 340b participation may not be a major driver of differences in initial cancer treatment decisions.