Inequalities in health: living conditions and infant mortality in Northeastern Brazil

Abstract

OBJECTIVE

To analyze the variation of infant mortality as per condition of life in the urban setting.

METHODS

Ecological study performed with data regarding registered deaths of children under the age of one who resided in Aracaju, SE, Northeastern Brazil, from 2001 to 2010. Infant mortality inequalities were assessed based on the spatial distribution of the Living Conditions Index for each neighborhood, classified into four strata. The average mortality rates of 2001-2005 and 2006-2010 were compared using the Student’s t-test.

RESULTS

Average infant mortality rates decreased from 25.3 during 2001-2005 to 17.7 deaths per 1,000 live births in 2006-2010. Despite the decrease in the rates in all the strata during that decade, inequality of infant mortality risks increased in neighborhoods with worse living conditions compared with that in areas with better living conditions.

CONCLUSIONS

Infant mortality rates in Aracaju showed a decline, but with important differences among neighborhoods. The assessment based on a living condition perspective can explain the differences in the risks of infant mortality rates in urban areas, highlighting health inequalities in infant mortality as a multidimensional issue.

INTRODUCTION

Infant mortality rates aid in the assessment of the living conditions of a population. Infant mortality is commonly used to verify the level of life and health of a location due to the fact of it being sensitive to improvements in the access to health care and quality of life. ²² The condition of infant health, considered as being a classic indicator of public health, is associated with, among other factors, access to health services, sanitation, education level of the mother, housing, and nutrition. Reducing the mortality rate in children under the age of five is the fourth millennium goal of the United Nations Development Program (UNDP). Brazil reduced this rate by 5.2% from 1990 to 2008, ²¹ being one of the countries that have achieved the goal. According to the Pan American Health Organization, ^a infant mortality rate (IMR) estimates of deaths per thousand live births (LB) for children under the age of one for Mexico, Brazil, and Chile in 2011 was 14.1, 16.2, and 7.4, respectively. In 2010, the IMR for the United States, Spain, and Sweden was 6.1, 3.2, and 2.5 deaths per thousand LB, respectively. ^b

Infant mortality is uneven throughout Brazilian regions. ²¹ The average IMR in Brazil in 2010 was 16.0/1,000 LB and the highest rates were found in the North (21.0/1,000 LB), Northeast (19.1 per thousand LB), and in the Midwest (15.9 per thousand LB) region. Sergipe (SE), Northeastern Brazil, presented an IMR of 18.2 deaths per thousand LB. ^c The urban environment wherein the child lives is considered to be a social determinant of the child’s health. The risk of death to which the child will be submitted during the first year of life is associated with the demographic and socioeconomic indicators of where he/she lives, such as urbanization, access to sanitation and housing, income level, and mother’s level of education. ²⁰ Therefore, indicators composed of social variables are created to estimate the living conditions of the population ^{5 , 20} and to contribute to creating monitoring areas of social injustices. ¹⁷

The objective of this study was to analyze the variation in infant mortality per living conditions.

METHODS

Ecological population study based on the analysis of secondary data regarding deaths of children under the age of one (without differentiation between neonatal and post-neonatal components) living in Aracaju, SE, from 2001 to 2010. The period was divided into two five-year periods (2001-2005 and 2006-2010) to analyze the variation of infant deaths per indicator of living condition.

Aracaju, the capital of Sergipe, has 39 neighborhoods and an area of 181.8 km². In 2010, the city presented a Human Development Index (HDI) of 0.770, an urbanization rate of 100%, a population of 571,149 inhabitants, ^d and a total of 9,311 LB. ^c

Data from the Sistema de Informação de Mortalidade (SIM – Mortality Information System) and from the Sistema de Informação de Nascidos Vivos (SINASC – Live Births Information System) from Aracaju were used for the calculation of IMR. The IMR was calculated in a direct way, given that Aracaju had an overall mortality rate greater than 5.3 and a percentage of ill-defined causes below 2.0% over the study period (97.5% for 2001-2005 and 97.8% for 2006-2010). ^{6 , 19}

The general IMR was calculated by dividing the number of deaths of children under the age of one by the total number of LB for every thousand LB. ¹⁹ The specific IMR was calculated for each basic cause of death. The determination of the basic cause of death followed the recommendations of the International Statistical Classification of Diseases and Related Health Problems (ICD-10). The causes were grouped in accordance with a reduced list of compiled causes of infant death: perinatal problems (ICD-10: P00-P96), acute respiratory infections (ICD-10: J00-J39), diarrheal diseases (ICD-10: A00- A09), other infectious and parasitic diseases (ICD-10: A15-B99), congenital malformations (ICD-10: Q00-Q99), neoplasms (ICD-10: C00-C97), external causes (ICD-10: V01- Y98), and other causes (other codes ICD-10).

The elaboration of the Living Conditions Index (LCI) for each neighborhood was performed based on the demographic data from the 2010 census of the Brazilian Institute of Geography and Statistics (IBGE), ^d by means of adapting the methodology described by Costa et al ⁵ (2001), who used five indicators related to the proxy variables of living conditions. Two of the indicators were excluded for not having statistically significant differences between the neighborhoods: average number of residents per household and percentage of households connected to the water supply network. Three indicators were used representing education, income, and housing: proportion of illiterate people who were heads of households; proportion of households earning less than 1/4 of the minimum wage per capita; and proportion of households in subnormal clusters.

The LCI was established for each neighborhood, based on the aforementioned indicators. The organization of scores enabled the grouping of each district into four strata (quartiles), comprising relatively homogeneous areas of living conditions according to the final score: “high” (3-24), “intermediate” (25-44), “low” (45-60), and “very low” (62-90). This stratification was chosen to maintain the balance among the divisions as each stratum was composed of 25.0% of the observations. The first three strata were composed of ten neighborhoods and the last one was composed of nine. The LCI was spatially distributed on the thematic map by means of the ArcGIS software (Figure 1), which makes the creation and the modeling of maps possible.

Figure 1. Spatial distribution of the Living Conditions Index according to strata in neighborhoods. Aracaju, SE, Northeastern Brazil, 2001 to 2010.

The Student’s t-test with p < 0.05 significance level was used to compare the average rates of infant mortality and specific mortality during both periods.

This study was approved by the Ethics and Research Committee of Universidade Tiradentes (Protocol 060 711) and complied with all the recommendations of Resolution 466/2012 of the National Commission on Research Ethics.

RESULTS

The IMR decreased from 29.5 in 2001 to 17.7 deaths per thousand LB in 2010 (Figure 2). The average IMR decreased in all the living condition strata between the two five-year periods. The strata “very low” and “high” were the ones with the highest reduction percentages (34.7% and 33.8%, respectively). Average infant mortality rates differed between the two periods in all the living condition strata, except for the “intermediate” strata (Table 1).

Figure 2. Infant mortality per thousand live births. Aracaju, SE, Northeastern Brazil, 2001 to 2010.

Table 1. Average infant mortality rate per thousand live births. Aracaju, SE, Northeastern Brazil, 2001 to 2005 and 2006 to 2010.

Living Condition Index	Average infant mortality rate		p*
	2001-2005	2006-2010
High	20.7	13.7	0.020
Intermediate	23.1	18.4	0.100
Low	25.4	18.0	0.009
Very low	30	19.6	< 0.001
Aracaju, SE	25.3	17.7	0.002

* Student t test

The main causes of death in 2001-2005 were perinatal problems, with an average IMR of 18.1 deaths per thousand LB, which significantly reduced during 2006-2010 to 12.3. Deaths from diarrheal causes and from other infectious and parasitic diseases have ceased to exist in the “high” strata. This did not occur in the other strata, despite the significant reduction (Table 3).

Table 3. Specific average of mortality rate per 1,000 live births for groups of causes as per the strata of living conditions. Aracaju, SE, Northeastern Brazil, 2001 to 2005 and 2006 to 2010.

Causes of death	2001-2005					2006-2010
	High	Intermediate	Low	Very Low	Aracaju	High	Intermediate	Low	Very Low	Aracaju
Perinatal problems (ICD-10: P00-P96)	15.7	16.4	18.5	20.4	18.1	10.4	12.3	12.7	13.2	12.3
Congenital malformations (ICD-10: Q00-Q99)	2.4	3.3	2.7	2.7	2.8	1.8	4.3	2.5	3.0	2.8
Acute respiratory infections (ICD-10: J00-J39)	0.4	0.7	0.8	1.0	0.8	0.3	0.2	0.8	0.6	0.5
Diarrheal diseases (ICD-10: A00- A09)	0.2	0.7	1.1	1.6	1.0	–	0.2	0.2	0.1	0.1
Other infectious and parasitic diseases (ICD-10: A15-B99)	0.4	0.3	0.5	0.5	0.5	–	0.2	0.3	0.4	0.3
Neoplasms (ICD-10: C00-C97)	–	–	0.1	–	–	–	–	–	0.1	–
External causes (ICD-10: V01- Y98)	0.2	0.3	0.2	0.2	0.2	0.1	0.1	0.5	0.4	0.3
Other causes (other codes ICD-10)	1.0	1.0	1.1	2.3	1.4	1.0	0.8	0.8	1.0	0.9

ICD-10: tenth revision of the International Classification of Diseases

Deaths from diarrheal diseases showed the greatest reduction (-84.8%). The causes of infant mortality due to congenital malformations and external causes presented an increase over the study period (Table 2).

Table 2. Specific average of mortality rate per 1,000 live births. Aracaju, SE, Northeastern Brazil, 2001 to 2005 and 2006 to 2010.

Causes of death	2001-2005	2006-2010	Difference (%)
Perinatal problems (ICD-10: P00-P96)	18.11	12.30	-32.1
Acute respiratory infections (ICD-10: J00-J39)	0.77	0.53	-31.2
Diarrheal diseases (ICD-10: A00- A09)	0.99	0.15	-84.8
Other infectious and parasitic diseases (ICD-10: A15-B99)	0.45	0.37	-17.8
Congenital malformations (ICD-10: Q00-Q99)	2.77	2.85	+2.9
Neoplasms (ICD-10: C00-C97)	0.02	0.02	0
External causes (ICD-10: V01- Y98)	0.26	0.32	+23.1
Other causes (other codes ICD-10)	1.42	0.89	-37.3

ICD-10: tenth revision of the International Classification of Diseases

DISCUSSION

IMR decreased significantly between the two five-year periods in Aracaju. This decrease has been associated to improvements in access to prenatal care, ²¹ water supply, ¹⁸ sanitation, and an increase of public expenditure on health in Brazil. ²²

The expansion of the Primary Health Care (PHC) in Brazil ^{10 , 21} enables greater access to basic health services by the population, which is important for prenatal, delivery and postpartum care, as well as child care after discharge from the hospital. The Estratégia de Saúde da Família (Family Health Strategy) had an important role in reducing child mortality in Brazilian municipalities from 1996 to 2004. ⁴ The program’s coverage in Aracaju went from 61.8% in 2002 to 84.8% in 2010. ^e The same occurred in Sao Paulo, where the decrease in infant mortality was influenced by the growth of GDP per capita and by the coverage provided by the Equipes de Saúde da Família (Family Health Teams) between 1998 and 2008. ¹⁰ The reorganization of the PHC in Colombia was associated to a reduction of 19.0% of IMR between 2003 and 2007, primarily of deaths associated to socioeconomic inequalities and housing. ¹³

The “high” strata showed lower average rates in comparison to other strata in the two periods, in the comparisons of the average and specific IMR. This indicates inequalities in the occurrences of infant death between the strata, accentuated from 2006 to 2010, when the risk increased in the ”intermediate,” “low,” and “very low” strata in comparison to the previous period. Despite the slowing down of IMR in all the strata during that decade, inequality with regards to infant mortality increased in the areas with the worst living conditions in comparison to areas with better conditions. Maia et al ¹¹ (2012) pointed to extreme inequalities in the profile of infant mortality in five Brazilian municipalities of different regions. The inequalities were strongly associated with socioeconomic conditions and access to health services, becoming an indicator of lack of equality. The chances of death in children under the age of one, were also higher among children from disadvantaged social classes in the United Kingdom from 1994 to 2011. ²³

In Belgium, from 1999 to 2008, 85.0% of deaths in children under the age of one occurred during the first month of life (neonatal deaths), especially during the first week, mainly due to perinatal problems related to prematurity. ¹⁵ In Aracaju, from 2001 to 2010, nearly 77.0% of deaths occurred in the first month of life, and perinatal problems accounted for 70% of all infant deaths, rendering it the main cause of death. The reduction in 32.0% of deaths due to perinatal causes, from 2006 to 2010, suggests an improvement in prenatal access and hospital care, especially in neonatal intensive care units. ^{3 , 11 , 22}

Congenital malformation was the second cause of death, accounting for about 13.0% of all deaths and the specific rate remained stable, with variations among strata. The average IMR for this cause in the “intermediate” stratum went from 3.3 deaths per thousand LB in 2001-2005 to 4.3 in 2006-2010. Other Latin American studies showed percentage variations in the occurrence of congenital malformations between 2.0% and 5.0%. ^{1 , 3 , 7} Despite these deaths being difficult to prevent, improvements in family planning and in prenatal care can contribute towards reducing them, since many anomalies are linked to environmental and nutritional factors related to the mother, such as the central nervous system. ³ Other studies to determine the causes would be required.

The average IMR due to diarrheal diseases from 2001 to 2005 is a strong indicator of inequality in health, greater in the “low” and “very low” strata. Moreover, the occurrence of this cause of death was sporadic from 2006 to 2010. The decrease of the rate per diarrheal causes was the greatest among all causes and may be associated to improvements in sanitation and greater access to health services. ¹⁴ Furthermore, the provision of the rotavirus vaccine as of 2006 proved to be efficient in reducing child mortality from acute diarrhea. ^{8 , 9}

The decrease suggests that most neighborhoods experienced improvements in sanitation and access to health services from 2006 to 2010, especially those classified as having “low” and “very low” LCI. Nevertheless, despite the ample reduction, especially in the “high” stratum, the other strata showed deaths due to diarrhea, indicating that they possess important differences in health conditions in comparison to those with better living conditions.

There was a reduction in deaths due to respiratory causes, represented mainly by pneumonia, which like the diarrhea, is treatable through primary care. ¹⁶ The decrease in IMR from respiratory causes probably occurred due to improved access to health services, proper management of cases, and the use, within the public health system, of vaccine against Haemophilus influenzae, an important causative agent of pneumonia in this age range. ²

The increase of deaths in children under the age of one due to external causes, ² especially in the “low” and “very low” LCI neighborhoods, indicates the issue of accidents and violence in this stratum of living conditions, showing the social and health inequalities to which the population is subject. ¹⁷ In the municipality of Cuiabá, upon studying infant deaths due to external causes, it was noted that the more children in a family, the lower was the educational level, and that the lower the mother’s age, the greater were the difficulties in preventing accidents within this age group. ¹²

Given that this study was based on averages of social and demographic variables and on IMR itself, some limitations are inherent to this type of research. Local differences within the same neighborhood could become masked in the average, thus rendering them impossible to be made evident. However, despite it being a study using aggregated data, it is nonetheless extremely important for the assessment and planning of health care. The “low” and “very low” strata had the worst average rates, except for the analysis of congenital malformations, confirming that, despite progress, health inequalities persist in neighborhoods with the worst living conditions.

Infant mortality rates in Aracaju are decreasing, but they show important differences between neighborhoods. The research from the perspective of living conditions shows the differences in infant mortality within urban areas. We are not attempting to list definite associations, but are in fact attempting to ratify health inequalities in infant mortality as a multidimensional phenomenon that cannot be analyzed from a single perspective.

Despite the significant decrease in infant mortality in Aracaju, it still constitutes a serious problem in certain neighborhoods of the city. The actions of child health policies should consider the urban location, the characteristics of the locales and, social indicators to attain a more accurate estimate of the possible applicability of the resources available.

Funding Statement

Research supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES – Financial Aid – AUX-PE-PROSUP 2203/2010, Process 23038009763/201).