During the past decade, many hospitals experienced difficulty integrating primary care practices into their health systems. We hypothesized that this difficulty may be, in part, a result of limited understanding of practice organizational designs. The structure and function of practices have not been well studied. In this article, we answer the following questions: Are practices all the same, or do variations in their organizational design exist? Do hospital designs predict the designs of affiliated practices? If variation exists, what are the management implications? Eighteen family practices, including nine affiliated with five separate hospital systems, were studied using an in-depth comparative case study design. A content analysis of the rich descriptive data from these cases indicates that a great variety exists in the organizational design of primary care practices, and this variety appears to be influenced by the initial conditions under which the practice was organized. Hospital system design in and of itself did not predict the design of affiliated practices. In fact, both affiliated and independent practices exhibited a range of design characteristics, some of which did not fit traditional models. Hospital systems that allowed greater flexibility of practice organizational designs were more effective at integrating and managing practices. Practices response to environmental change was greater when practice autonomy was highest. These findings suggest that a science of practice organizational design separate from that of hospitals is needed to help explain the success and failure of practices within health systems and to provide information for planning practice change.