Language:
1. Postdoctoral Fellow of the Postgraduate Program in Community Health – Department of Social Medicine – University of São Paulo. PhD and Master of Science from the Faculty of Medicine of Ribeirão Preto, University of São Paulo. Ribeirão Preto, SP, Brazil
2. PhD in Biomedical Engineering (2001) at the Federal University of Rio de Janeiro. Researcher at the Health Information Laboratory, Institute of Scientific and Technological Communication and Information in Health. – Oswaldo Cruz Foundation – ICICT/FIOCRUZ. Rio de Janeiro, RJ, Brazil 3. PhD in
Public Health from the Faculty of Medicine, University of Porto. Researcher – i3S – Institute for Innovation and Research in Health, University of Porto. Porto, Portugal
4. PhD in Medical Sciences (Obstetrics and Gynecology) from the State University of Campinas (UNICAMP). Campinas, SP, Brazil. Master of Statistics from the Federal University of São Carlos/SP, Brazil. Associate Professor at the University of São Paulo (USP/Ribeirão Preto). Ribeirão Preto, SP, Brazil
Edson Zangiacomi Martinez Department of Social Medicine, Ribeirão
Preto School of Medicine, University of São Paulo Zip Code: 14049-900 (
edson@fmrp.usp.br )
Descriptors: Adolescent, teenage pregnancy, ecological studies, spatial distribution of the population.
Abstract:
OBJECTIVE: To study the relationship between teenage pregnancy rates and the Minas Gerais Social Responsibility Index.
METHODS: An ecological study was conducted based on records of live births of teenage mothers, obtained from the Live Birth Information System (SINASC). The Minas Gerais Social Responsibility Index (IMRS) and its dimensions were used as covariates. Statistical analysis was based on a Bayesian regression model with a spatial-temporal structure.
RESULTS: A reduction in teenage pregnancy rates was observed over the years. For the entire period studied, there were profound differences between the North and South regions of the state. Both in the most developed regions of the state and in those marked by significant structural and socioeconomic deficiencies, there is an inverse relationship between social responsibility in public management, measured by the IMRS, and the reduction in teenage pregnancy rates.
CONCLUSION: This finding shows that teenage pregnancy cannot be studied without considering the geographic and socioeconomic context in which these young women are inserted.
Palabra Clave: Adolescent, teenage pregnancy, ecological studies, spatial distribution of the population.
Abstract:
OBJECTIVE: Study the relationship between adolescent pregnancy rates and the State Social Responsibility Index in Minas Gerais State.
METHODS: Was conducted an ecological study based on records of live births from adolescent mothers, which were obtained from the Live Births Information System (SINASC). The Minas Gerais State Social Responsibility Index (MGSSRI) and its dimensions were used as co-variables. The statistical analysis was based on a Bayesian space-time regression model.
RESULTS: There was a reduction in the adolescent pregnancy rates over the years. With regard to the study period, profound differences were observed between the northern and southern regions of the state. In the most developed regions of the State as well as in with significant structural and socio-economic deficiencies, there is an inverse relationship between public administration social responsibility, measured by MGSSRI, and reduction in the adolescent pregnancy rates.
CONCLUSION: This finding is an evidence that pregnancy during adolescence cannot be studied without considering the geographical and socio-economic contexts in which these adolescents are inserted.
Abstract:
OBJECTIVE: To study the relationship between the tasks of embarazo in adolescence and the Mineral Index of Social Responsibility in Minas Gerais.
METHODS:An ecological study was carried out based on records of live births of teenage mothers, obtained by the Live Birth Information System (SINASC). The Mineral Social Responsibility Index (IMRS) and its dimensions were used as a co-variable. The statistical analysis is based on a Bayesian regression model with a spatio-temporal structure.
RESULTS: A reduction in the costs of embarrassment in adolescence was observed over the years. For the entire period studied, there are profound differences between the North and South regions of the State. Both in the most developed regions of the State and in those marked by significant structural and socio-economic deficiencies, there is an inverse relationship between social responsibility in public management, measured by the IMRS, and the reduction of embarrassment levels in adolescence.
CONCLUSION: This study shows that teenage embarrassment cannot be studied without considering the geographic and socioeconomic context in which these young women are inserted.
INTRODUCTION
Adolescence is a phase marked by the transition between childhood and adulthood. Much more than a biological process, adolescence encompasses psychosocial aspects, and is a period in which cognitive development and personality structuring are accelerated, marked by bodily changes resulting from hormonal action, in both boys and girls. 1 At this stage, individuals tend not to perceive their vulnerability, do not recognize behaviors that involve personal risks, and thus fail to use means that could protect them. It is in this context that unexpected and sometimes unwanted pregnancies become a relevant problem, both from a social and health perspective. 2,3 Several studies highlight the causes and consequences of early pregnancy, including social, economic, educational and behavioral aspects. 4,5 Children of adolescent mothers are more likely to be born prematurely, with low weight for gestational age, 6,7 iron deficiency anemia and cephalopelvic disproportion, in addition to presenting nutritional complications and infectious diseases more frequently. 8 Studies indicate that obstetric complications result mainly from incomplete development of the pelvic and uterine bones. 9 In addition, adolescent pregnancy has been associated with school dropout, 10 with recurrence being more frequent in young people who are out of school or in a school year that is inappropriate for their age. 11 Excluding pregnant adolescents from the school environment reinforces the cycle of perpetuating poverty, since these young women will have difficulty accessing qualified job opportunities and improving their social status. 12
According to the Demographic Census conducted by the Brazilian Institute of Geography and Statistics (IBGE), in 1991, 32.5% of births to first-time mothers were concentrated in pregnant women aged between 10 and 19 years. In 2000, this percentage was over 38%. According to data from the Live Birth Information System (SINASC), between 2010 and 2013, an increase of approximately 3.5% in live births to teenage mothers between the ages of 10 and 14 was recorded in Brazil. Some studies describe the vulnerability of these younger adolescents due to their early sexual initiation and the fact that they come from poor families and are victims of physical and sexual abuse. 13,14
Teenage pregnancy has proven to be a serious public health problem; however, it is not due to a lack of information on the part of adolescents; some studies 15conclude that adolescents show a high level of knowledge regarding the existence of contraceptive methods. Thus, it is believed that the environmental conditions to which they are exposed are one of the factors that lead them to early pregnancy. Therefore, this study aims to measure the geographic distribution of teenage pregnancy in a Brazilian Federation Unit, using a Bayesian model of space-time analysis that allows describing the possible associations between the phenomenon and indicators of social responsibility of the areas that make up this space.
METHODS
An ecological study was conducted, a model that considers people in the context of various environments or ecological systems in which they live: family, relationships, neighborhood, community and institutions such as school and workplace. This model is based on the premise that individuals cannot be studied without considering the various ecological systems in which they live. 16
The state of Minas Gerais has approximately 20 million inhabitants, being the second most populous in Brazil and the fourth largest in territorial extension. It is composed of 853 municipalities, 66 microregions and 12 mesoregions, as shown in figure 1. It is the Federative Unit of Brazil with the largest number of municipalities, where it presents striking characteristics of social and economic inequalities when comparing different regions, with the Central and North regions being deeply marked by quite adverse natural and structural conditions. 17
To characterize the state of Minas Gerais according to the number of live births to teenage mothers, data from the Live Birth Information System (SINASC) of the Ministry of Health (DATASUS) were used, considering the total number of live births in each microregion of the state of Minas Gerais in the period from 2000 to 2010, and the number of live births whose mothers were between 10 and 19 years old, in the respective microregions. The percentage of teenage pregnancies was calculated by dividing the live births of mothers aged between 10 and 19 years old by the total number of live births.
As an independent variable, the Minas Gerais Social Responsibility Index (IMRS) for the year 2000 was used. The IMRS was created with the objective of obtaining a quantitative indicator of social responsibility in the public management of the 853 municipalities of Minas Gerais, with the João Pinheiro Foundation responsible for its construction. The index is based on the policies, plans, programs, projects and actions implemented by the municipal administration that ensure the population’s access to education, health, social assistance, public safety, income and employment, sanitation and housing, environment, culture and sports. Together with an indicator focused on public finances, these items make up the ten dimensions of the IMRS. Each of these dimensions is transformed into indices that range from 0 to 1, with the “general” IMRS being given by a weighted average of the indices of the ten dimensions. The data are available on the João Pinheiro Foundation website. To obtain indicators related to each microregion, since the IMRS is available for each municipality, an average weighted by the respective population sizes was used.
This study was submitted to and approved by the local Research Ethics Committee (Process No. HCRP-10157/2011).
Statistical analysis
A Bayesian space-time conditional autoregressive (CAR) model 18 was used , in which: Y ij denotes the count of births to adolescent mothers, N ij denotes the total number of live births and θ ij denotes the rate of adolescent pregnancies, with i representing each microregion and j each year of the series analyzed. The statistical model considers that Yij is a random variable that follows a binomial distribution with probability of “success” θ ij in Nij independent trials ( Nij known), where i = 1,…,66 microregions and j = 1,…,11 years ( j = 1 denotes the year 2000, j = 1 denotes the year 2000, j = 1 denotes the year 2000, and j = 1 denotes the year 2000 ).=2 denotes the year 2001, and so on). A logit link function was assumed between the teenage pregnancy rates θij and an observation xi of the independent variable X (the “general” IMRS or each of its dimensions), written in the form
logitθij = α 0j + d i + w ij + b j ( x i – m )
where: m is the arithmetic mean of the observations of X , α 0j and b j are fixed effects, d i are spatial effects associated with the i -th microregion and w ij are the respective temporal effects. In the Bayesian analysis, it is considered that each d i assumes a spatial prior distribution with a CAR structure, which allows the correlations between nearby areas in space to be greater. The estimation of the model parameters was based on stochastic simulation based on MCMC ( Markov Chain Monte Carlo ) methods, using the GeoBUGS 19 module of the Win-BUGS program. It was assumed that the effects w ij follow a priori a multivariate normal distribution with a vector of means equal to zero and a variance matrix described by Branscum et al., 20 which attributes larger covariances between successive times, which ensures a longitudinal structure for the data. It was considered that the other a priori distributions are non-informative and, among themselves, independent. Using the MCMC method, 30,000 samples were generated for each parameter of interest, with the first 1,000 samples being discarded to avoid any effect of the initial values ( burn-in samples ). The DIC criterion 21 was used to compare models , such that models that present lower DIC values are those with the best fit to the data.
RESULTS
Among the pregnancies with live births that occurred in the state of Minas Gerais in the years 2000 and 2010, respectively 20.49% and 16.82% were adolescents. Figure 2 shows that the rates tend to decrease over the years in most microregions. The maps presented in Figure 2 were the result of the adjustment of the Bayesian space-time model, considering the IMRS as a covariate. Throughout the period, the profound differences between the rates observed in the North and South regions of the state prevailed, with the lowest rates tending to be concentrated in the South region.
The lowest rates coincidentally refer to the regions with the largest populations and greatest development, such as Belo Horizonte, Divinópolis, Itaguara and Conselheiro Lafaiete. The highest rates of teenage pregnancy are concentrated in the north of the state, where the microregions of Pirapora, Frutal, Grão Mogol, Unaí and Paracatu stand out. The microregion of Pirapora stands out for presenting the highest rates in all years of the series.
The greatest reductions in the percentages of teenage pregnancy for the period were observed in the microregions of Uberlândia, which fell from 24% in 2000 to 16% in 2010, and Mantena, which decreased from 27% to 19%. On the other hand, the microregions of Conceição do Mato, Guanhães and Andrelândia had their teenage pregnancy rates increase from 17% to 21%, from 18% to 21% and from 18% to 19%, respectively. The microregions of Januária, Grão Mogol, Diamantina, Pedra Azul, Curvelo, Itabira, Peçanha and Santa Rita do Sapucaí had no changes in their percentages.
Alternatively, a model was also adjusted without the inclusion of spatial effects (DIC=6703), but the DIC value obtained from the model including these effects was lower (DIC=6513). This shows that teenage pregnancy rates are not randomly distributed among the different microregions of the state, but there is some significant spatial effect.
Table 1 describes the mean teenage pregnancy rates according to the “general” IMRS classes (0.3 to 0.5; 0.5 to 0.6 and 0.6 to 0.8), and the ratios between the rates, with their respective Bayesian credibility intervals (95% CI). Intervals that do not include the value 1 indicate significant associations with teenage pregnancy rates (indicated with asterisks “*” in Table 1). For all years of the study period, an inverse relationship was observed between teenage pregnancy rates and IMRS values.
Spatiotemporal models were adjusted considering each of the ten dimensions of the IMRS. Table 2 shows the association between the dimensions of education, income and employment, and health, and the rates of adolescent pregnancy in the years 2000, 2005 and 2010. Although these models consider all years of the series, only the results related to these three years were described in the table, for parsimony. Table 2 showed higher rates of adolescent pregnancy in municipalities with lower indices of social responsibility related to these dimensions. Other models, considering the other dimensions of social responsibility, showed significant associations between the respective indices and the rates of adolescent pregnancy (results not shown).
DISCUSSION
Given that the North and Northeast regions of Minas Gerais have lower levels of development than those found in the South of the state (Figure 2), it is observed that these profound asymmetries have some effect on the phenomenon of teenage pregnancy. The northeast of the state includes the mesoregion of the Jequitinhonha Valley, described in the literature as the poorest and least developed in the state 17 .
The microregion of Pirapora, located in the north of the state, had the highest percentages of teenage pregnancy throughout the period. Other notable microregions with the highest percentages are Frutal and Ituiutaba, both belonging to the mesoregion of the Triângulo Mineiro/Alto do Paranaíba and to the microregions of Jequitinhonha and Vale do Mucuri. The microregions where the highest percentages of teenage pregnancy were observed were those that had the lowest values of the “general” IMRS (Table 1) and its dimensions (Table 2). In the adjustment of the space-time model considering the “general” IMRS as a covariate, the microregions of Montes Claros and Salinas showed a marked improvement at the end of the series, in 2010. In all dimensions of IMRS, the lowest percentages were observed in the Metropolitan mesoregion of Belo Horizonte, a region characterized by municipalities with larger populations, greater supply and opportunities for employment, better educational opportunities, and more leisure and sports options.
The present study was based on an ecological model, in which the social phenomenon of teenage pregnancy is treated collectively, while most studies published in the literature 6,22,23 used women living in specific regions as sampling units. The greatest limitation of ecological studies is the fact that they are susceptible to ecological bias or fallacy, such that an association observed between groups of individuals does not necessarily mean that the same association occurs at the individual level. In the present study, another important limitation is that microregions with larger population sizes, such as the Belo Horizonte microregion, possibly have a very heterogeneous spatial distribution of their social indicators, which is not characterized in the model used, which describes each microregion uniformly.
However, the results found were similar to those observed in studies that considered pregnant women as sampling units 6,22,23 and in similar ecological studies, such as the study conducted by Nogueira et al. 5 on the analysis of the spatial distribution of adolescent pregnancy in the city of Belo Horizonte. These authors demonstrated the presence of clusters with high proportions of adolescent mothers associated with the worst socioeconomic conditions. Martinez et al. 24 conducted an ecological study with spatial analysis on teenage pregnancy and socioeconomic characteristics of municipalities in the state of São Paulo, showing that the occurrence of early pregnancy is higher in municipalities with lower per capita gross domestic product, lower human development index and higher poverty rates. Martins et al. 9 conducted an ecological study with spatial analysis for the health microregions of the state of Mato Grosso do Sul and found that fertility among adolescent women is higher in microregions with worse indicators of education and socioeconomic development.
The Bayesian model used here proved to be efficient in estimating the adjusted rates, with the spatial structure and neighborhood matrix adopted being adequate for the data, since, when incorporated into the models, the DIC values showed a large reduction. The results obtained show a strong spatial dependence for teenage pregnancy rates among municipalities in Minas Gerais, which suggests that neighborhood structure plays a fundamental role in understanding them in relation to economic and social indicators.
Another potential limitation of this study is that the completeness of SINASC information may not be uniform across the entire state of Minas Gerais, as is the case in other Brazilian states. 25 A study 26 that evaluated SINASC in 132 municipalities in Minas Gerais in 2010 showed that the system was not properly implemented in most of the municipalities evaluated, highlighting the lack of qualified professionals, unsatisfactory collection and completion of the Declaration of Live Births, underutilization of data, and poor dissemination of information.
On the other hand, although these limitations may have significant effects on the findings of this study, it can be concluded that teenage pregnancy goes beyond the biological realm. For example, evidence was found that low levels of public safety, one of the dimensions of the IMRS, are associated with higher rates of teenage pregnancy. A global study27 conducted with adolescents aged 15 to 19 years living in disadvantaged urban areas of five different cities (Baltimore-USA, Johannesburg-South Africa, Ibadan-Nigeria, New Delhi-India and Shanghai-China) found a similar pattern. This study showed that the chances of a teenager becoming pregnant are higher in violent neighborhoods and where the fear of being robbed or assaulted is high. Furthermore, the study conducted by Copping et al.28 supports the argument that the perception of violence, sexual precocity and teenage pregnancy are related to environmental conditions.
CONCLUSION
This study demonstrated a relationship between social responsibility in public management, measured by the IMRS, and the reduction in teenage pregnancy rates. This finding contributes to the consensus that teenage pregnancy cannot be studied without considering the geographic and socioeconomic context in which these young women live, so that, on a larger scale, public investments that allow the population access to multiple sectors, such as education, environment, culture, and leisure, should always be understood as essential for health promotion.
1. Osório LC. What is adolescence, after all? In: Osório LC. Adolescente hoje. Porto Alegre: Artes Médicas; 1989. p.10-3.
2. Correia LL, Mcauliffe JF, Rouquayrol MZ, Almeida-Filho N. Maternal and child health. In: Rouquayrol MZ, Almeida-Filho N (org.). Epidemiol e Saúde. 5th ed. Rio de Janeiro: MEDSI; 1999. p.375-403.
3. Guimarães EM. Teenage pregnancy: a multidisciplinary view. Ped Mod. 2001;37(Special Edition).
4. Chalem E, Mitsuhiro SS, Ferri CP, Barros MC, Guinsburg R, Laranjeira R. Teenage pregnancy: behavioral and socio-demographic profile of an urban Brazilian population. Cad Saude Publ 2007;23:177-86.
5. Nogueira MJ, Silva BF, Barcelos SM, Schall VT. Analysis of the spatial distribution of adolescent pregnancy in the city of Belo Horizonte. Rev Bras Epidemiol 2009;12:297-312.
6. Gama SG, Szwarcwald CL, Leal MC. Pregnancy in adolescence, associated factors, and perinatal results among low-income post-partum women. Cad Saúde Publ 2002;18:153-61.
7. Costa MC, Santos CA, Nascimento Sobrinho C, Moura MS, Souza KE, Assis DR. Pregnancy in adolescence: sociodemographic and biomedical maternal variables association and neonatal outcomes. Rev Baiana Saúde Publ 2005;29:300-12.
8. Henriques-Mueller MH, YUNES J. Adolescence: mistakes and hopes. In: PAHO/WHO Gender, Women and Health in the Americas. Washington, DC: Scientific Publication. 1993;541:46-67.
9. Martins PC, Pontes ER, Paranhos Filho AC, Ribeiro AA. Adolescent pregnancy: an ecological study in the health micro-regions of the State of Mato Grosso do Sul, Brazil – 2008. Epidemiol Serv Saúde 2014;23:91-100.
10. Silveira RE, Santos AS. Pregnancy in adolescence and school dropout: an integrative literature review. Rev Enferm Atenção Saúde 2013;2:89-98.
11. Nery IS, Mendonça RC, Gomes IS, Fernandes AC, Oliveira DC. Relapse in to preganancy in adolescents from Teresina, PI, Brazil. Rev Bras Enferm 2011;64:31-7.
12. Almeida MC, Aquino EM. The role of education level in the intergenerational pattern of adolescent pregnancy in Brazil. Int Perspect Sex Reprod Health 2009;35:139-46.
13. Teixeira SA, Taquette SR. Violence and unsafe sexual practices in adolescents under 15 years of age. Rev Ass Med Bras 2010;56:440-6.
14. Taquette SR, Vilhena MM, Paula MC. Sexually transmitted diseases in adolescence study of risk factors. Rev Soc Bras Med Trop. 2004;37:210-4.
15. Belo MA, Pinto JL. Knowledge, attitudes, and practices on previous use of contraceptive methods among pregnant teenagers. Rev Saude Publ 2004;38:479-87.
16. Brooks-Gunn J, Duncan GJ, Klebanov PK, Sealand N. Do neighborhoods influence child and adolescent development? Am J Sociol 1993;99:353-95.
17. Stefani J, Nunes MA, Matos R. IMRS and its Dynamics in the Planning Region Jequitinhonha – Mucuri. Cad Geogr 2014;24:17-33.
18. Besag J, Kooperberg CL. On conditional and intrinsic autoregressions. Biometrics. 1995;82:733-46.
19. Thomas A, Best N, Lunn D, Arnold R, Spiegelhalter D. Geo BUGS user manual. Cambridge: Med Res Counc Biostat Unit; 2004.
20. Branscum AJ, Perez AM, Johnson WO, Thurmond MC. Bayesian spatiotemporal analysis of foot-and-mouth disease data from the Republic of Turkey. Epidemiol Infect 2008;136:833-42.
21. Spiegelhalter DJ, Best NG, Carlin BP, Van Der Linde A. Bayesian measures of model complexity and fit (with discussion). JR Stat Soc Series B 2002;64:583-616.
22. Cesar JA, Mendoza-Sassi RA, González-Chica DA, Mano PDS, Goulart-Filha SDM. Socio-demographic characteristics and prenatal and childbirth care in southern Brazil. Cad Saúde Pub. 2011;27:985-94.
23. Gama SGND, Szwarcwald CL, Leal MDC, Theme Filha MM. The pregnancy during adolescence as a risk factor for low birth weight, Brazil. Rev Saúde Publ 2001;35:74-80.
24. Martinez EZ, Roza DL, Caccia-Bava MC, Achcar JA, Dal-Fabbro AL. Teenage pregnancy rates and socioeconomic characteristics of municipalities in São Paulo State, Southeast Brazil: a spatial analysis. Cad Saúde Publ 2011;27:855-67.
25. Silva LP, Moreira CM, Amorim MH, Castro DS, Zandonade E. Evaluation of the quality of data in the Live Birth Information System and the Information System on Mortality during the neonatal period on the state of Espírito Santo, Brazil, between 2007 and 2009. Cienc Saude Colet 2014;19:2011-20.
26. Guimarães EA, Hartz ZM, Loyola Filho AI, Meira AJ, Luz ZM. Evaluating the implementation of Information System on Live Births in municipalities of Minas Gerais, Brazil. Cad Saúde Publ 2013;29:2105-18.
27. Brahmbhatt H, Kågesten A, Emerson M, Decker MR, Olumide AO, Ojengbede O, et al. Prevalence and determinants of adolescent pregnancy in urban disadvantaged settings across five cities. J Adolescent Health 2014; 55:S48-57.
28. Copping LT, Campbell A, Muncer S. Violence, teenage pregnancy, and life history: ecological factors and their impact on strategy-driven behavior. Hum Nat 2013;24:137-57