O processo de regionalização da saúde no Brasil: acesso geográfico e eficiência

Abstract

This dissertation analyzes the geographic access and relative efficiency of Brazilian health regions. Three empirical exercises on the regionalization of health in Brazil are presented, investigating i) its effects on geographic access; ii) its results over health indicators; and iii) the relative efficiency of the regions. The first article analyzes the consistency of the health regionalization process in Brazil, based on quantitative indicators of adequacy and geographic access. Regional dropout rates, average distances covered, and average travel times are estimated considering all regions of the country. and disaggregated by state, for hospitalizations of medium complexity in SUS. Microsimulation exercises identify the effects of changes in the design of health regions in relation to changes in the use of health services in the years 2002, 2008, 2011 and 2015. The main results indicate that the spatiality of patient flows is primarily internal to the very health region, since that is where most of the consultations take place. Cross border movements to other health regions imply distances of over 100 km. It is concluded that the normative regionalization process without a correspondence with the optimization and improvement of the regional distribution of the provision can generate advances in some aspects of the regionalization, but it has little effect on the geographic access, measured in this work by the indicators of average distance traveled and the evasion rate. The second article analyzes the direct quantitative effects of the regionalization proposed in the Pact for Health for Brazil on health outcomes, supply of beds, geographic access and financing. For this, models of differences-in-differences with fixed municipal effects and controls for socioeconomic and political conditions are estimated. The identification strategy is based on the differences in the municipalities' adherence dates (time of adherence) to the Pact. A robustness analysis is also carried out, estimating the model considering the years before joining the Pact (2003 to 2006). For most outcome indicators, the effect of treatment (years of adherence to the Pact) shows the expected direction: longer adherence is related to lower mortality rates (infant and neonatal) and hospitalizations, but this relationship is not statistically significant. In the analysis of heterogeneity, it is observed that the North region benefited from this process with a faster and more significant convergence from the adhesion to the Pact with regard to the average distance covered. The results allow to reinforce the need to discuss regionalization beyond a normative view and to rethink the investment mechanisms considering the limitations imposed by the federative model. Finally, the third article analyzes the relative efficiency of the health regions from the input-oriented data envelopment analysis. Two models are estimated: the first measures inputs using the expense variable and the second considers quantities from various categories of inputs (doctors, nurses, beds and equipment). The variables that measure production are the same for both models. The estimated average efficiency scores show significant differences depending on the input used, being much lower for the model that uses the expense. There is also a significant proportion of inefficient regions, more related to technical inefficiencies than inefficiencies of scale. These results show the importance of deepening the discussion on the regionalization of health in Brazil, considering its centrality to guarantee comprehensive health care.