SOCIOECONOMIC DETERMINATION AND SPATIAL DISTRIBUTION OF TUBERCULOSIS/HIV CO-INFECTION IN A STATE OF BRAZIL

Objective: To investigate the influence of social, economic and housing variables due to Tuberculosis/HIV in the state of Alagoas, Brazil. Methods: This is an ecological and cross-sectional study that evaluates the independent effect of variables from the Brazilian Institute of Geography and Statistics and United Nations Development Programme in the Annual Average Incidence Rate of Tuberculosis/HIV coinfection in the state of Alagoas from 2001 to 2016. We used Generalized Estimating Equation models implemented in SPSS software. Results: Among indicators related to cases of tuberculosis in people living with HIV in the municipalities of Alagoas, such as, the variables of household income per capita and percentage of households without piped water. A significant association in tests and measuring the quantification of important effects in the relationship among the strength association of each risk factors and the development of co-infection. Conclusion: The results can be consistent, allowing the identification of factors associated with the socioeconomic and socio-housing determination of TB/HIV co-infection in the state of Alagoas.


INTRODUÇÃO
With the emergence of the Human Immunodeficiency Virus (HIV), in lately decades, a genuine resurgence of Mycobacterium tuberculosis infection has been observed (WHO, 2018).
In an attempt to understand the context of tuberculosis (TB) and HIV and to try to justify why certain groups become more ill than others, several studies have incorporated in their analysis the framework of the determinants of the health-disease process (MUNAYCO et al., 2015;MARTINS, 2018). At the same time, research that outlines socioepidemiological profiles related to tuberculosis and HIV/AIDS mention features that are often associated with these diseases, with an emphasis on the conditions and life situations that increase health risks (PARKER; CAMARGO JUNIOR, 2000;ROSSETTO et al., 2018).
Studies had investigated associations of TB/HIV co-infection with environmental, social, behavioral and cultural characteristics, relating these findings as indicators of vulnerability (MAFFACCIOLLI; OLIVEIRA; BRAND, 2017). The set of economic, social, regional, ethnic inequalities and the fragility of public services offer have also been reported as factors influencing the epidemiological situation of these diseases CECCON et al., 2017;ROCHA et al., 2018).
Alagoas is one of the Brazilian federative units with the highest incidence of tuberculosis, ranking 4 th in relation to the coefficient of new people affected by TB in the country and 5 th in relation to the incidence of TB cases in people living with HIV/AIDS (BRASIL, 2017;SANTOS JÚNIOR et al., 2019). The state is still one of the poorest in the country (BRASIL, 2010).
With the intention of deepening the analysis of the scenarios in which people most affected by TB/HIV co-infection are inserted, a survey was carried out in 2018, in the state of Alagoas. This study analyzes indicators and understand its relationship with the occurrence of those affected by TB/HIV co-infection.
This study aimed to investigate the social, economic and housing influences in TB/HIV, in one of the poorest regions of Brazil, with high rates of infection with TB and HIV.
We believed that when we understand the sociodemographic dynamics of the areas in which the greatest records of the occurrence of these coinfection are concentrated we could have a better control and effectiveness of the treatment. Beside it, the association among characteristics of the locations constitute important measures for the health surveillance sector, since they can assist the improvement of public policies favorable.

Design of the study
An ecological study was designed with a focus on the possible determinants of socioeconomic and housing components in the Annual Average Incidence Rate (AAIR) of TB/HIV incidence in the municipalities of Alagoas, from January 2001 to December 2016. Lima-Costa & Barreto (2003) define that ecological studies compare the occurrence of a health-related condition (a disease -for example) and the exposure of interest among households of individuals (populations of countries, regions or municipalities -for example) to verify the possible existence of association among them. In this type of study, there is no information about the disease and exposure of the individual, but it is about the entire population group.  Simple and multiple linear regression analysis were implemented in order to obtain the regression coefficients that relate each variable in the model to AAIR through generalized estimating equation. From the univariate analysis, those indicators for which the AAIR regression had a significance >0.01 were introduced in the multiple regression model. Due to the asymmetric distribution of the data, gamma regression with log connection was used and, as the aggregation of municipalities was not taken into account, the independent correlation structure was chosen.

Location of the study
In order to fulfill the proposed objective, that is, to verify which variables would have the greatest explanatory power on the incidence of TB/HIV, the premise was that all the analyzed information presented homogeneity within the aggregation units studied in this work, such a reasonable consideration, since the municipalities are homogeneous units by IBGE to carry out their census.
The association analysis were estimated by Odds Ratio (OR) and used the SPSS The data were presented using graphs, tables and maps, and the results were discussed in the light of the specialized literature on the subject.

Ethical aspects
As this is a study using data from secondary sources, approval by the Research Ethics Committee was not required. The database employed was made available by the  (Table 2). In the analysis of the average incidence per municipality, 42%, it means, 43 of the 102 municipalities in the state had AAIR above the state average ( Figure 2).
The cities with the largest AAIR were located mostly in the eastern part of the state and distributed unevenly in the space of the Federative Unit.
There was no formation of "clusters" and / or spatial clusters of high incidences for AAIR in the period under analysis ( Figure 2).

DISCUSSION
The results obtained through the method of generalized estimating equation allowed the identification of the factors associated with the socioeconomic and socio-housing determination of TB/HIV co-infection in Alagoas, as well as the exploratory analysis allowed the visualization of the spatial distribution of the related variables in the territory.
The bibliography presents a series of determinants and factors related to the occurrence of tuberculosis and HIV (PEDRO; PELISSARI et al., 2018). In this study it was found that the population living in areas with the worst vulnerability indicators get ill more due to TB/HIV coinfection compared to population that reside in the state's territories with the best categorizations in socioeconomic and socio-housing indicators.
Among the indicators associated with tuberculosis cases in people living with HIV in the municipalities of Alagoas, two variables are associated with multivariate analyses. The results of those who lived in regions with higher family income per capita were less likely to have TB/HIV co-infection. This finding is in accordance with publications, which emphasizes the higher TB morbidity in families with the main financial conditions (DYE et al., 2009;XIMENES et al., 2009;PEREIRA et al., 2015;PELISSARI et al., 2018;PEREIRA et al., 2018).
However, an observed odds ratio for those who live in regions with a higher percentage of households without water supply was meaningless contrary to what was expected. This finding contradicts publications, which emphasized greater TB morbidity in houses without water supply (CANTWELL et al., 1998;DYE et al., 2009;PEREIRA et al., 2015;PEDRO et al., 2017;SILVA et al., 2018) and may be linked to Luis David Castiel called an ecological fallacy (cross-level bias),a phenomenon of imbalance where, in an analysis of aggregated data -as adopted in this study, a result is generated that seems paradoxical but, in a more detailed analysis, it appears plausible at the individual level (CASTIEL, 1998).
On a socioeconomic determination of TB/HIV co-infection, Parker and Camargo Jr. 21 to analyze anthropological aspects to HIV related to international documented literature as factors that facilitate the transmission of AIDS viruses and their influence in geographic and population. In their study they conclude that economic underdevelopment and poverty are the main factors related to HIV (PARKER; CAMARGO JUNIOR, 2000). They emphasized that inequalities put women and men in situations of vulnerability and that the process of underdevelopment creates social displacement that produces actions and practices that increase the risk of HIV infection (PARKER; CAMARGO JUNIOR, 2000). Other advanced studies in the African continent and in Brazil have also confirmed this relationship (PARKHURST, 2000;BRUNELLO et al., 2011;FOX, 2012). In the same vein, Pereira et al. (2015) reported a higher occurrence of TB in the neighborhoods of Rio de Janeiro and presented the main socioeconomic conditions. The authors, when carrying out a spatial analysis of TB cases, identified the formation of spatial clusters and major cases of disease in slum regions -areas that shelter housing in precarious conditions, devoid of regularization by public authorities and, often, services. such as treated water, sewage, among others (PEREIRA et al. 2015).
Information on the percentage of illiteracy, poor and/or extremely poor, unemployed, households without garbage collection and the number of residents per household, which revealed significance in the univariate analysis, has also been cited in the specialized literature as conditions associated with infection by TB. Studies (ALBUQUERQUE et al., 2007;XIMENES et al., 2009;BELO et al., 2011;PEREIRA et al. 2015, PEDRO et al., 2017MAGALHÃES et al., 2017)  Finally, it is worth emphasizing that this study has limitations common to those using secondary databases, in which the accuracy and completeness of the information can limit the findings by not allowing the researcher to control possible errors resulting from typing, registration, omission in the filling in fields and possible underreporting. There may also be a limitation linked to the design adopted, of data analysis at the aggregate level, which may not explain exposure and event at the individual level. However, even with these considerations and respecting the limits of the type of study carried out, there is a possibility that the results obtained can be used to enable a more incisive action by the public authorities in situations associated with a greater risk of the occurrence of TB/HIV coinfection, optimizing the allocation of resources.