Objectives.This article describes Medicaid participation among office-based primary care physicians in cities and examines its determinants.Methods.Data used in this study were collected through the 1993 and 1994 American Medical Association Socioeconomic Monitoring System telephone surveys. The sample includes 1,300 primary care physicians. Our multivariate model includes a variety of personal, practice, community, and policy factors thought to influence participation. Logistic regression was used to examine determinants of accepting any Medicaid patients and ordinary least square regression was used to examine determinants of the extent of participation among participants.Results.The authors found that 19% of respondents did not participate in Medicaid and 62% had practices with 9% or fewer Medicaid patients. Multivariate analyses indicated that Medicaid payment levels were not associated with observed patterns of Medicaid participation. Community sociode-mographic characteristics and demand from Medicaid-eligibles, by contrast, play a significant role in influencing observed levels of participation.Conclusions.Strategies other than raising Medicaid payment levels will be needed to achieve equitable access to office-based primary care for the poor residing in cities.