OBJECTIVES: To evaluate the effectiveness of an automated prior authorization system (SmartPA) in 1) reducing utilization/expenditures of Cox-2 inhibitors; 2) increasing utilization/expenditures of Cox-2 substitutes (traditional nonsteroidal anti-inflammatory drugs (NSAIDs), other products for pain or musculoskeletal conditions, and GI protective agents) to no more than the decrease seen for Cox-2's; and 3) decreasing Cox-2 utilization/expenditures more in patients at low risk for GI complications than in patients at high risk. METHODS: The study used a before and after with control group design. Changes in utilization/expenditures of Cox-2 inhibitors and Cox-2 substitute products were compared after implementation of SmartPA in Missouri's Medicaid plan. The Medicaid plan of an eastern state that had no PA system for Cox-2's was used as the control. Subjects were patients with a claim for a Cox-2 inhibitor in the 12-month baseline period that were continuously enrolled for the 24-month study period. Analyses consisted of comparisons of means and linear regression. Regressions controlled for differences in age, gender, risk for GI complications, and severity of illness. RESULTS: Regression results indicated that increases in expenditures on Cox-2 inhibitors and GI-protective products were $178 per patient per year (PPPY) higher and $191 PPPY higher, respectively, in the control state. Increases in expenditures on traditional NSAIDs and other pain and musculoskeletal products were $34 PPPY higher in the treatment state. Patients at low risk for GI complications in the treatment state experienced greater reductions in expenditures than did those in the high-risk group. Results were similar for utilization for all of the above analyses. CONCLUSION: SmartPA was successful in reducing expenditures/utilization on Cox-2 inhibitors for Missouri Medicaid compared to the control, while keeping the increase in expenditures/utilization for Cox-2 substitutes substantially lower than this decrease. These effects were concentrated among patients at low risk for GI complications.