Get access

Spillover Effects of Restrictive Drug Formularies on Physician Prescribing Behavior: Evidence from Medicaid


  • This study was funded by AstraZeneca Pharmaceuticals. We thank Y. Aileen Lin and Zhihua Qiao for help with statistical analysis, John Cohen for help with data extraction, and Lisa Croll for administrative assistance. We also thank a coeditor and two anonymous referees at this journal for helpful comments that substantially improved this paper.


Restrictive drug formularies may influence physician prescribing behavior for other patients with more generous drug benefit, so-called “spillover effects.” We focus on Protonix, a proton pump inhibitor, introduced at a discount price, and study subsequent physician prescribing decision for open-formulary Medicaid patients. Using two national databases on physician prescribing patterns and health plan drug formularies, we find consistent evidence of significant, positive spillover effects from 1-PPI and 2-PPI formularies that include Protonix onto Medicaid, with a larger effect from the most restrictive, 1-PPI formularies. Physicians who prescribe a higher proportion of Protonix to their non-Medicaid patients because of restrictive formularies, also prescribe a higher proportion of Protonix to their Medicaid patients with an open formulary. Each 10% increase in Protonix's predicted non-Medicaid share results in an 8% increase in its share of Medicaid prescriptions (p < 0.001). Similar spillover effects seem to exist for older PPI products as well. A restrictive drug formulary therefore influences physician prescribing behavior for other patients not directly subject to its restrictions.