Using laws, common sense, and statistical approaches to design indicators for 'medical desertification'. An application on the Romanian case.

The study of spatial accessibility to healthcare services is key to health policy (Pförtner et al., 2019; Vergier et al., 2017). Terms such as 'rurality' or 'medical desertification' were employed to stress the relevance of the topic. Within the existing literature, there is little (if any) concern with the legitimacy of the existing ways to measure inequality. Expert systems were assumed to be in place, and discrepancies or consistency with existing regulations or to views of relevant stakeholders were not considered. This paper discusses spatial accessibility of medical services in three distinct approaches: normative, that is following what national regulations consider as standard; relative, that is what statistical approaches reveal; consensual, which adjusts the indexes based on representations of stakeholders. The three approaches are compared for the case of Romania, an EU country with low population density, ideal to inspect geographical discrepancies. For the relative and the consensual approach, population is adjusted according to different demand expressed by age groups, and also including population and providers from the nearby localities. The normative approach follows official regulations. The refinements in the consensual model are based on survey data from stakeholders and consider distances to neighbouring localities and adjustments according to the age structure of the population in the catchment area. The results reveal high consistency between the consensual and the relative approach. Both are more permissive with respect to detecting medical desertification as compared to the normative approach but prove to be more effective in directing policy when resources are scarce. The normative approach, however, is relevant in depicting the state of the system as contrasted to a desired standard. The relative approach also overlaps with the consensual one. Therefore, to fully characterize spatial accessibility to general practitioners (GPs) and pharmacies, one needs to consider at least the normative and the relative approaches.