[What do functionally defined populations contribute to the explanation of regional differences in medical care?].

Geographic variation in health care is increasingly subject to analysis and health policy aiming at the suitable allocation of resources and the reduction of unwarranted variation for the patient populations concerned. As in the case of area-level indicators, in most cases populations are geographically defined. The concept of geographically defined populations, however, may be self-limiting with respect to identifying the potential for improvement. As an alternative, we explored how a functional definition of populations would support defining the scope for reducing unwarranted geographical variations. Given that patients in Germany have virtually no limits in accessing physicians of their choice, we adapted a method that has been developed in the United States to create virtual networks of physicians based on commonly treated patients. Using the physician claims data under statutory insurance, which covers 90% of the population, we defined 43,006 populations-and networks-in 2010. We found that there is considerable variation between the population in terms of their risk structure and the share of the primary care practice in the total services provided. Moreover, there are marked differences in the size and structure of networks between cities, densely populated regions, and rural regions. We analyzed the variation for two area-level indicators: the proportion of diabetics with at least one HbA1c test per year for diabetics, and the proportion of patients with low back pain undergoing computed tomography and/or magnetic resonance imaging. Variation at the level of functionally defined populations proved to be larger than for geographically defined populations. The pattern of distribution gives evidence on the degree to which consensus targets could be reached and which networks need to be addressed in order to reduce unwarranted regional variation. The concept of functionally defined populations needs to be further developed before implementation.