Equity in the Accessibility to Specialized Care in Oral Health in Brazil: Analysis of PMAQ-CEO External Evaluation Data According to the User's Perspective

Objective: To analyze the distribution of Centers of Dental Specialties (CEO) implemented in Brazil until 2014 and identify the contextual and individual variables associated with the geographical accessibility, considering the user's perspective. Material and Methods: The study was conducted with data from the Improving Access and Quality of CEOs (AVE/PMAQ CEO) External Evaluation and contextual characteristics of states and municipalities. The AVE/PMAQ CEO was an evaluative research with cross-sectional observational character, performed in all CEOs in Brazil, with 932 services evaluated and 8,897 users interviewed. Data analysis was conducted in two stages, considering the study dimensions. Analyses were conducted on the availability of units by region, state, population size of the municipality and contextual variables. Then, the relationship between explanatory variables and the user's displacement time and with the intention to change the service location was evaluated using generalized linear regression analysis. Results: The 932 CEOs evaluated were located in 780 of the 5,570 municipalities of the country and the majority was located in the northeastern (38.3%) and southeastern regions (36.2%), with the northern and midwestern regions presenting the lowest absolute number of units. The average displacement time to the CEO was 28.4 minutes, while the intention to change CEO location due to the distance from home was reported by 7.8% of users. Lower geographical accessibility was reported in the northern region and for individuals who reported living in the rural area and in cities with larger populations. Conclusion: The availability of CEOs in the country is still low and not equitably distributed among states and regions.


Introduction
The National Oral Health Policy (PNSB), launched in 2004 and nicknamed "Brasil Sorridente" aims at universal access and comprehensive oral health care. To that end, it seeks to consolidate a care model guided by the principles and guidelines of the Unified Health System (SUS), where the Family Health Strategy (PSF) reorients this model in primary care with the expansion of oral health teams to ensure the integrality of dental care, with the provision of specialized assistance through the Centers of Dental Specialties (CEO) [1].
In this sense, a good coverage of primary care that allows the coordination between levels of care and the adequate use of oral health services are requirements to achieve the desired quality. It is also necessary to take into account that the use of services can be influenced by their accessibility.
Accessibility is defined as a characteristic of the relationship between need, demand and supply to health actions and services, being composed of two dimensions: socio-organizational and geographic. The first is related to the characteristics of the service offering, that is, resources that make it easier or difficult for people to reach care. The second refers to displacement, such as distance, time and cost to reach the service [2,3].
The barriers in accessibility affect different social groups in different ways, especially those of greater social vulnerability, being relevant for research on equity in health systems. Barriers such as cost and transportation difficulties, structural problems and human resources of health units, time waiting for care, among others, tend to be higher for lower income groups, who generally live in areas where the availability of services is lower, making access difficult [4]. Groups with better socioeconomic conditions have an easier way to overcome the imposed access difficulties [5].
In addition, there is evidence that contextual characteristics, such as population size and the Human Development Index (HDI) of regions served by CEOs, are related to their performance [6][7][8]. However, few studies have addressed the distribution of supply and its influence on accessibility and equity. Thus, the aim of this study was to analyze the distribution of CEOs implemented in Brazil up to year 2014, as well as to identify the contextual and individual variables associated with geographical accessibility, considering the user's perspective.

Improving Access and Quality for CEOs Program
The study was developed with data from the 1st cycle "Improving Access and Quality of The external evaluation instrument was organized into three modules: module I -Direct observation of the infrastructure conditions, equipment, instruments, materials, inputs and medicines. The quality assessor was followed up by a team professional evaluated during the on-site visit at the health facility to be evaluated; Module II -Interview with professionals (manager and dentist surgeon of CEOs) to obtain information about the team's work process and about the organization of the care with users; Module III -Interview with users to verify their satisfaction and perception regarding health services, in relation to their access and use.
The quality evaluators of PMAQ-CEO, selected and trained by the Institutions of Education and Research, have applied tools to verify access and quality standards achieved by professionals and management. Observations of CEO's infrastructure were made, interviews with CEO's Managers, one dental surgeons and 10 users (over 18 years of age) and document verification.

Distribution of CEOs in the Country
Initially, a description was given of the CEO number by region, unit of the federation and population size of municipalities in which CEOs were present. Then, the number of CEOs per Data analysis was conducted in two stages, considering the study dimensions: 1) Analysis of the distribution of CEOs; and 2) Analysis of user accessibility. In the first step, after descriptive analysis of data by region and population size of the municipality, analysis was carried out using scatter plots between health indicators and contextual variables.
In the second stage, the relationship between explanatory variables and outcomes was evaluated through generalized linear regression analyses. As the displacement time variable presented exponential distribution, the association between the outcome and the independent variables was evaluated considering the Gamma distribution, with model adjustment by the maximum likelihood method. The logarithmic function was used and coefficients were expressed as averages for categorical variables or variation in the displacement time according to the coefficient for continuous variables. For the outcome related to the option of change the CEO location, the Poisson distribution was used, also using a logarithmic link function, expressing the rate ratio effect (proportions) for the categorical variables or rate variation according to the coefficient for continuous variables. Variables with p<0.10 in the crude analyses were included in the adjusted analysis. The number of individuals with missing data was low (<5%), except for variable CEO absenteeism percentage, which presented a high percentage of loss (>30%), which required the imputation of data through the multiple imputation method. All analyses were performed using the Stata SE Software, version 13. The significance level was set at 0.01.

Ethical Aspects
The AVE / PMAQ-CEO project was approved by the Ethics Research Committee of the Federal University of Pernambuco under CAAE No. 23458213.0.0000.5208.

Results
The 932 CEOs surveyed were located in 780 of the country's 5,570 municipalities. While almost 70% of Brazilian municipalities had up to 20,000 inhabitants, only 11.7% (n = 109) of CEOs were located in one of these municipalities. On the other hand, almost 40% (n = 371) of CEOs were in municipalities with more than 100,000 inhabitants, 27.7% (n = 257) in municipalities with more than 200,000 inhabitants. Little more than 5% of Brazilian municipalities had population of more than 100,000 inhabitants in 2014.
Most of the evaluated CEOs were located in the northeastern (38.3%) and southeastern regions (36.2%). The northern and mid-western regions had the lowest absolute number of CEOs.
When evaluated by the number of inhabitants, the northern region was highlighted by the lowest number of CEOs (0.348 per 100,000 inhabitants) and the northeastern region by the highest number (0.635 per 100,000 inhabitants). The state of Roraima presented the most unfavorable population indicators, both in relation to the number of CEOs and the number of dental surgeons in the minimum specialties. The best indicators were found in Paraíba. Variables and health indicators are described in Table 1.    In the users' analyses, the longer average displacement times were reported in the northern region. In this region, the highest values were also found, which probably express displacement time longer than one day. Longer times were also found for users of municipalities with larger population sizes and higher HDI-M, and CEOs with higher absenteeism percentage and whose access was only of the referenced type. Users living in states with higher number of CEOs per inhabitant and municipalities with greater BSE coverage reported less displacement time. Longer times were also found for individuals who reported not being white; living in the rural area and households with high number of residents, lower income and schooling and those participating of the family allowance program.  The measures of effect and their respective confidence intervals in the crude and adjusted analyses are described in Tables 3 and 4   The northern region presented higher proportion of users who would change the CEO location due to the 'far from home' motivation, compared to the southeastern and northeastern regions; the larger the population size, the greater this proportion. In addition, greater proportion of users whose form of access is exclusively given by spontaneous demand was found; and living in the rural area, having lower income and some family member receiving family allowance are associated with the intention to change CEO location.  accessing the service, reproducing the reverse care law [10], already demonstrated in other collective oral health actions in the country, such as fluoridation of public water supply [11]. This incongruence between need and specialized oral health care has already been previously identified, when 339 CEOs were implemented in the country in the year 2006 [12]. In addition, the greater geographic and social distances, cultural and economic barriers faced by the most vulnerable populations undermine regionalization policies, reaffirming existing inequalities [7,13,14].

Discussion
The pioneering and active participation of several actors in the northeastern region in the development and implementation of the current National Oral Health Policy has made the scenario in the region to be distinct [8,12], with a general attitude towards equity, with high availability of services in agreement with social and health indicators. As can be seen, this characteristic is capable of influencing the general panorama of the country, underestimating inequities in the federative units of the other regions.
In addition, it should be considered that studies evaluating CEOs suggest that there is an inadequacy regarding procedures performed in these units, with the service being used as a substitute for primary care rather than complementary in the care network [7,15]. The inadequacy of the offered care can act as an aggravating factor to the low availability of the service and does not contribute to integrality in oral health [16].
Individually, the low accessibility to secondary care reported by users of CEOs in the northern region, of municipalities of greater population and residents of the rural zone stands out.
These variables presented high effect measures for both variables used as outcome in the analyses. In addition, the HDI-M and the CEO / inhabitant indicator showed significant effect on the time required to reach the CEO, and the fact that access to the CEO did not take place solely by spontaneous demand showed a great protective effect in the intention to change CEO location. It is important to remember at the moment, that the outcomes were collected from the users, representing, therefore, the barriers faced by those who, somehow, already had access to the service.
Thus, the results were interpreted in addition to previous analyses in order not to consider them as representative of the general population's access to secondary care in oral health. For example, the reports of users from larger municipalities show greater difficulties to access CEOs cannot be translated as better access to the service by the population of small municipalities, which, as observed, presented less availability of units.
Thus, the greater the population size of municipalities with CEOs, and, on average, higher HDI-M, greater distances and difficulties of urban locomotion can be faced by users. In addition, a smaller expansion of primary care in these municipalities [17] may compromise the care network.
Although there was no effect of the percentage of BSE coverage in municipalities, perhaps due to the relative homogeneity of municipalities that presented secondary attention, the association of the lower intention of changing CEO location with the form of access to the CEO, which not only by spontaneous demand, can indicate that a better organization of the assistance network would favor the user's accessibility, in addition to the effectiveness of the principle of integrality [16]. On the other hand, another study demonstrated the impact of distance to the municipality with CEO on the use of services, suggesting that there is a relationship between primary care coverage and secondary care organization [14].
The results suggest the presence of iniquities also among individuals who use the service, since users with worse socioeconomic conditions, living in the rural area and in the northern region reported worse geographical accessibility to CEOs. Considering that conditions of social vulnerability also lead to poorer oral health conditions [18][19][20], individuals with greater needs would be facing more barriers to access specialized services. Thus, difficulties in accessing the service can be considered as predictors of oral health conditions [21]. In addition, alternatives in the private network are also more difficult for individuals with greater social vulnerability [22]. Thus, the need to expand specialized services should take into account the needs of the population and the organization of the health care network, with the aim of reducing inequities in the use and access to secondary care in oral health in the country. In addition, standardization in the CEOs coverage is necessary in order to guarantee minimum availability and access to services.

Conclusion
The availability of CEOs in the country is still low and unevenly distributed across states and regions, concentrating in larger municipalities. In addition, users from the northern region living in municipalities with more than 200,000 inhabitants and from the rural area have reported barriers in the geographical accessibility to these health units.