I. INTRODUCTION This article examines how racial differences in medical care utilization vary with respect to intensity of demand. There are many studies of health-care consumption among nonwhite minorities, most of which compare racial differences in utilization between blacks and whites. Results vary depending on the measure of utilization and population of interest, but the general consensus is that blacks consume fewer health-care services than whites (e.g., Currie and Thomas, 1995; Gilbert et al., 2002; Johnson-Lans and Bellemore, 1997; LaVeist et al., 2002). These findings reveal an important social problem in which nonwhite minorities face disadvantages in receiving health care. The issue is particularly urgent if racial differences in health-care utilization have detrimental effects on overall health among minority groups. While empirical evidence of racial differences in utilization is convincing and not usually disputed, it is not clear whether racial differences are pervasive among all members of a given racial group. If differences in utilization apply to one type of individual more than another then policies aimed at reducing racial differences that are targeted toward all members of a particular minority group are an inefficient allocation of resources. Deb and Trivedi (1997, 2002) provide evidence that health-care consumers can be characterized into two groups consisting of heavy and light users. Furthermore, they show that heavy users tend to be relatively unhealthy. If racial differences in utilization apply equally to all members of a minority group and do not depend on whether individuals are healthy or unhealthy then differences are a general race-related concern that should be addressed. However, if differences depend, in part, on health status then policy implications depend significantly on who is most affected. For example, if the difference between white and black utilization is largest among relatively unhealthy individuals but not among healthy individuals then this would indicate an urgent social concern in which minorities most in need of care have difficulty obtaining it. In this case, health-care policy should focus on reducing racial differences among unhealthy minorities. On the other hand, if racial differences mainly apply to healthy individuals then minorities most in need of care are able to obtain it, but policy should focus on potential reasons that healthy minorities use fewer services, such as lack of knowledge of routine treatments and physicals, distrust of physicians and medical personnel, or potential overuse by white users. This article offers several contributions to the debate on racial differences in utilization. First, health-care utilization is estimated for different subpopulations with respect to various measures of health status, which allows for the classification of health-care consumers into subgroups with different intensities of demand. This specification allows us to determine whether racial differences vary between subgroups of consumers. In addition to blacks, we also consider utilization by Hispanics. Second, the model is estimated separately for five measures of utilization recorded on an annual basis: office-based physician visits, office-based nonphysician visits, outpatient department visits, emergency room (ER) visits, and hospital discharges. (1) Thus, one can determine whether racial differences are more related to office-based services or hospital-related emergency care. Finally, the models of health-care utilization explicitly control for access-related factors such as insurance status, which allows us to examine whether racial differences in utilization are still present after controlling for such factors. These remaining differences are more of a policy concern as they reflect racial "disparities" in health-care utilization. (2) As Medical Expenditure Panel Survey (MEPS) tabulations show below, African Americans and Hispanics have lower rates of private insurance coverage and higher rates of public insurance coverage than other Americans. …