This study uses Medicare physician-claims data to examine patient and physician contributions to variations in cataract surgery rates across U.S. metropolitan areas. Utilization is modelled as having two phases: the decision to seek an ophthalmologist's care, which is made by patients, and the decision to perform surgery on patients who seek care, which is partially controlled by ophthalmologists. Under this model, the effect of physician practice style on cataract surgery rates occurs through the influence of practice style on the second phase of utilization. Variation in patient care-seeking behavior contributed to the variation in the rate of cataract surgery. Moreover, multivariate regression analyses found that cataract surgery rates were influenced by economic and sociodemographic variables in predictable ways. Using the regression results, a “purged” cataract surgery rate that was free of any possible influence of physician practice style was calculated. Variation in the purged surgery rate was only slightly lower than variation in the observed surgery rate, suggesting that eliminating practice style as a factor in physician decision making (e.g., through practice guidelines) would reduce variations in cataract surgery rates by only a small amount.