Addressing geographic access barriers to emergency care services: a national ecologic study of hospitals in Brazil

The present work can be classified as a cross-sectional ecological study. Were analyzed data about 5843 Brazilian hospitals in three phases: 1) evaluation of the distances among SH and nearest high complexity center (HCC) within administrative regions; 2) assessment of geographical accessibility to hospital care; 3) identification of clusters of municipalities with minimal-structured SH for surgical care located in regions with low HCC accessibility.

Brazil is an upper-middle income country with 206 million inhabitants (IBGE, 2016). The Brazilian health system is organized to provide universal access to entire population - the Sistema Único de Saúde - SUS [23]. The country is divided into five administrative regions (North, Northeast, Center-West, Southeast and South), 26 states, and the federal district (Fig. 1). Up to 2013, the Brazilian healthcare network was composed of 5843 hospitals, 2774 of which were SH and 3069 HCC. Considering the guidelines from Brazilian Ministry of Health, a facility is classified as SH if the number of beds available is equal or less to 50. HCC are defined as institutions with: presence of certification related to ECS and capacity to offer surgical procedures related to ECS, or realized more than 1000 deliveries by month. The HCC were stratified in 2559 exclusive high complexity centers and 486 facilities that were also acting as maternities. From the universe of HCC was made the option to exclude of analysis 24 facilities for been exclusively maternities. These maternities do not act like a complete ECS facility once they do not meet all the criteria defined by Brazilian Ministry of Health [23]. In Brazil, we have guidelines trying to organize the referrals of patients, but no rule or law defining which type of health facility a municipality should have. Thus, there are municipalities in our study in all possible conditions: with SH and HCC, only SH, only HCC, none hospital.

Data analysis was divided in three steps and was conducted using Python [26] and ARCMAP 10.2 softwares [27]. The first step is independent, while the second and third steps were related.

In total, 432 (7,76%) Brazilian municipalities had HCC (black squares – Fig. 2). The majority of municipalities (55,99%) had SH located less than 60 km from the closest municipality with an HCC. Yet, there was still a shortage of HCC in the Central-West, rural Northeast, and especially the North regions (Fig. 2). For these regions, there was a polarization of HCC in urban centers, particularly state capitals. From this analysis, it was noted 824 municipalities (Table 2) did not meet the criteria of maximum travel time of 2 h , recommended distance by Lancet commission on Global Surgery to assure timely access to essential surgery services [24].

Although there were states with good accessibility to SH, there was no pattern among regions (Fig. 3). The states of Paraná, Goiás, Minas Gerais, Bahia, Piauí, Rio Grande do Norte, Ceará e Pernambuco presented high levels of accessibility, of hospital beds by inhabitant located within the 60 km catchment area of the SH. For HCC accessibility, there was a concentration of high levels of accessibility in the South, Southeast, and Northeast coastal regions (Fig. 4). The North and Midwest regions experience a shortage of HCC. There were large portions of the country with nearly zero hospital beds by inhabitant considering the HCC accessibility index.

The ECS care gap analysis of SH and HCC accessibility index showed a group of 1595 municipalities (Table 3) that had high accessibility to SH and low to HCC (Fig. 5). Additionally, to indicate which areas could help overcoming the gap for ECS, the mean scores of SH MESR are depicted in Fig. 5. Considering all municipalities with below average access to HCC, 74% of them (1183) were covered by SH with at least 60% of MESR adherence proportion.

All municipalities selected after ECS care gap analysis were submitted to a spatial autocorrelation analysis considering its MESR values. Two significant clusters were identified (Fig. 6). The first spatial cluster spans through the Northern and Northeast regions and was comprised of 699 (12,56%) municipalities covered by SH with little ECS capability (blue cold spot). The second significant cluster group was located throughout the Southeast, South and a portion of the Central-West regions. This second cluster was composed of 589 (10,58%) municipalities comprising SH with good ECS capability (red hotspot). This result indicates hotspots areas with SH that meet the criteria to offer surgical care and are, simultaneously, located in regions with lack of access to HCC. The population potentially benefited in these areas comprises more than eight million inhabitants (Table 4). On the other hand, cold spots (blue areas) have SH which are not equipped to match ECS and Surgical care needs, but have enough presence to be a potential solution for the shortage of HCC in those regions. Non-significant areas (yellow) mark locations with dispersed characteristics of SH with and without ECS and Surgical capabilities, filling most of the Center-West region and Amazon area.