The Incidence of Hospital Drug Price Subsidies: 340B, Drug Utilization, and Subsidized Medical Care

Wednesday, June 26, 2019: 9:00 AM
Madison A (Marriott Wardman Park Hotel)

Presenter: Sayeh Nikpay

Co-Authors: Rena Conti; Melinda Buntin

Discussant: David Cutler

The Incidence of Hospital Drug Price Subsidies: 340B, Drug Utilization, and Subsidized Medical Care


The 340B program was created in 1992 to subsidize drug-based care for safety-net providers through mandated percentage discounts on outpatient drugs. Participants in the 340B program include public or non-profit hospitals with an adjusted disproportionate share (DSH) percentage of 11.75% or greater. The DSH patient percentage is a safety-net eligibility criterion that has been used by the Centers for Medicare and Medicaid Services (CMS) to qualify and allocate Medicare disproportionate share payments under the Medicare program. Congress allows hospitals to charge patients full price for discount drugs but revenue from drug administrations was intended to enhance care to low-income and uninsured patients. An important and unanswered question in the literature is whether hospitals are using program revenues to this end.


We link the universe of outpatient Medicare Part B claims from the revenue center file from 2013-2016 to 340B participation data, and data on subsidized medical care and hospital financial outcomes from Medicare hospital cost reports to estimate the incidence of 340B price subsidies on drug utilization, safety-net care, and financial performance (N=17,158).


We use difference-in-differences designs to compare hospitals that begin participating in the program between 2014-2016 (N=177) to two control groups: (1) a propensity-score re-weighted group of hospitals that never participate in 340B (N=940) and (2) a group of hospitals that are not participating in the program between 2013 and 2016 but will eventually participate in 2017 and 2018 (N=147). We then estimate the elasticity of subsidized medical care and the operating margin with respect to subsidized drug utilization and associated revenues.


Relative to never participants, participating in 340B increases Medicare revenue by 20.57% (SE=0.0576) and cost-sharing amounts billed to beneficiaries by 16.79% (SE=0.0534). Hospital drug administrations also increased among new participants relative to never participants (14.81% , SE=0.0487). The results were similar in magnitude and statistical significance when comparing to hospitals not-yet participating or comparing hospitals that gained clinics to those that did not. The results were larger for cancer drugs. Although administrations increased, there was no statistically significant increase in administrations per beneficiary between new participants and any comparison group. We estimate an elasticity of uncompensated care with respect to Medicare revenues of 0.1096 (SE=0.0555) before hospitals began participating, and this elasticity increased by 0.066 (SE=0.0216) after participation. This suggests that hospitals do provide more uncompensated care after participating in 340B, however we find no statistically significant relationship between Medicare bad debts and Medicare revenues before (-0.1468, SE=0.1058) or after 340B (-0.0036, SE=0.0628).


The 340B drug discount aims to increase unreimbursed care to uninsured and low-income populations by reducing the price of drug-based care. While the program does increase the administration of physician administered drugs, especially high-cost oncology drugs, patient out-of-pocket costs and uncompensated care provision increases after participation. On net, our results suggest 340B does improves safety net care provision but that the value of program participation is not fully passed onto patients.