Market penetration of HMOs affect physician practice styles for non-HMO patients. To study the impact of a restrictive Medicaid drug formulary on prescribing patterns for other patients, ie, so-called spillover effects. A before-and-after, 3-state comparison study. On January 1, 2001, Maine's Medicaid program implemented a restrictive drug formulary for the proton pump inhibitor class, with pantoprazole as the only preferred drug. The Medicaid and non-Medicaid market shares of pantoprazole in Maine (vs New Hampshire and Vermont and among Maine physicians with different Medicaid share of practice. After 3 months, the market share of pantoprazole in Maine (vs 2 control states) increased 79% among Medicaid prescriptions (vs 1%-2%), 10% among cash prescriptions (vs 3%), and 7% among other third-party payer prescriptions (vs 1%). The market shares increased more among Maine physicians with a higher Medicaid share of practice (high vs middle vs low [market]: 16% vs 8% vs 5% [cash]; 11% vs 5% vs 4% [other third-party payers]). Linear regression results indicate that practicing medicine in Maine leads to a 72% increase in pantoprazole share among Medicaid prescriptions (P < .001). In addition, for each 10% Medicaid share of practice in Maine, the share of pantoprazole increases 1.8% among cash prescriptions (P = .01) and 1.4% among other third-party payer prescriptions (P < .001). Maine's Medicaid drug formulary generated spillover effects in cash and other third-party payer markets, with somewhat stronger effects in the cash market.