Access to health care, health outcomes and hospital performance in Brazilian cities

Background

In September 2015, economists from 44 countries signed the Economists Declaration on Universal Health Coverage, calling upon global policymakers to prioritize it as a pillar of sustainable development. Brazil is exemplary among developing countries in having instituted universal health coverage. However the system is struggling under a growing burden of chronic disease, alongside continued high levels of infectious disease. In 2003, the government responded by introducing a National Policy on Emergency Care (PNAU). One of the major initiatives under the PNAU was the introduction of Urgent Care Units (UPA 24hs) in larger municipalities. These walk-in facilities were intended to act as both a primary care unit and a pre-hospital medical facility to absorb part of the rising demand for ambulatory care at hospitals. There are now around 300 UPAs active in the country, and nearly 1,000 more planned over the next few years.

This is a time when analysis of health system responses to management of crowding has wide relevance. In the United States for instance, emergency department visits increased 40% in the last two decades but, in this time, the number of emergency departments declined 11%, increasing pressure on the available departments. In the United Kingdom, waiting times at emergency units and for appointments with a GP have been increasing and this constitutes an increasingly important concern among voters.

Project aims

This project assesses the impacts of innovations within this new scheme on mortality and hospitalization rates and hence on the efficiency of the health system. It will also assess whether UPAs have reduced inequalities in access to health care and explore heterogeneous effects along the distribution of pre-intervention characteristics, such as local transport, health infrastructure, and health system integration across the different levels of government.

Methods

The research will be carried out by combining administrative records from the Ministry of Health with geocoded information containing the location and date of creation of the UPAs and with data from the Brazilian Censuses of 2000 and 2010 at the census tract level. The administrative data provide information since the mid-1990s, at the individual level, on patient admissions to hospitals, including main diagnosis, date of admission, mortality outcomes, and patient demographic characteristics, including zip code of residence. These data can be collapsed into (i) a monthly panel of data at the census tract level containing health indicators and distance to UPAs and other health facilities; and (ii) an analogous monthly panel of data at the hospital level containing hospital indicators. Both data sets allow the use of a difference-in-differences econometric approach to identify the effects of UPAs on the outcomes of interest given that the exact location of UPAs is largely influenced by fixed geographical features, such as the pre-existing transport infrastructure and other health facilities (that determined access to health care).