Decomposition of socioeconomic inequalities in the uptake of intermittent preventive treatment of malaria in pregnancy in Nigeria: evidence from Demographic Health Survey

Background: Although malaria in pregnancy is preventable with the use of intermittent preventive treatment with sulfadoxine–pyrimethamine (IPTp-SP), it still causes maternal morbidity and mortality, in sub-Saharan Africa and Nigeria in particular. Socioeconomic inequality leads to limited uptake of IPTp-SP by pregnant women and is, therefore, a public health challenge in Nigeria. This study aimed to measure and identify factors explaining socioeconomic inequality in the uptake of IPTp-SP in Nigeria. Methods: The study re-analysed dataset of 12,294 women aged 15–49 years from 2018 Nigeria Demographic Health Survey (DHS). The normalized concentration index (Cn) and concentration curve were used to quantify and graphically present socioeconomic inequalities in the uptake of IPTp-SP among pregnant women in Nigeria. The Cn was decomposed to identify key factors contributing to the observed socioeconomic inequality in the uptake of adequate (≥ 3) IPTp-SP. Results: The study showed a higher concentration of the adequate uptake of IPTp-SP among socioeconomically advantaged women (Cn = 0.062; 95% confidence interval [CI] 0.048 to 0.076) in Nigeria. There is a pro-rich inequality in the uptake of IPTp-SP in urban areas (Cn = 0.283; 95%CI 0.279 to 0.288). In contrast, a pro-poor inequality in the uptake of IPTp-SP was observed in rural areas (Cn = − 0.238; 95%CI − 0.242 to − 0.235). The result of the decomposition analysis indicated that geographic zone of residence and antenatal visits were the two main drivers for the concentration of the uptake of IPTp-SP among wealthier pregnant women in Nigeria. Conclusion: The pro-rich inequalities in the uptake of IPTp-SP among pregnant women in Nigeria, particularly in urban areas, warrant further attention. Strategies to improve the uptake of IPTp-SP among women residing in socioeconomically disadvantaged geographic zones (North-East and North-West) and improving antenatal visits among the poor women may reduce pro-rich inequality in the uptake of IPTp-SP among pregnant women in Nigeria. The study focuses on Nigeria. Nigeria, with an estimated population of 198 million in 2018, is the most populous country in Africa [17]. The population is predominantly young, with about 45% aged under 15 years and 20% under 5 years, while women of childbearing age (15–49 years) account for about 22% of the total population [18]. The country is divided into six geopolitical zones: North-Central, North-East, North-West, South-East, South-South, and South-West with each geopolitical zone comprising about six states [19]. Of the six zones, the northern geopolitical zones especially the North West and North East have the highest poverty rates in the country [20]. The health care system in Nigeria is largely public sector driven, with substantial private sector involvement in service provision. Most secondary- and tertiary-level health facilities are in urban areas, whereas rural areas are predominantly served by primary health care (PHC) facilities. There is a shortage of PHC facilities in some states [18] and less than 20% of health facilities in the country offer emergency obstetric care, despite that Nigeria accounts for one-quarter of all malaria cases in Africa [17]. Malaria accounts for 60% of outpatient visits and 30% of hospitalizations among children under five years of age in Nigeria. The malaria prevalence among children 6–59 months in the six geopolitical zones was as follows, South West (50.3%), North Central (49.4%), North West (48.2%), South-South (32.2%), North East (30.9%) and South East (27.6%) [21]. Data for the analyses were obtained from the latest Nigeria Demographic Health Survey (DHS) 2018, conducted between August 14, 2018 to December 29, 2018 [22]. The choice of the dataset is to ensure that the research finding represents current reality, which is nationally representative, and generalizable. The dataset contains survey information elicited from the women of reproductive age of 15–49 years in the six geopolitical zones in the country. The study used an Individual (women) Recode file that collected information on women’s and husband’s background characteristics, reproductive history, antenatal care, malaria prevention and treatment, household asset ownership, and type of toilet facilities among other information. The DHS survey uses a multistage sampling procedure, standardized tools and well-trained interviewers to collect comparable and reliable data on maternal and child health. The 2018 Nigeria DHS had a response rate of 99%. Further information about the survey has been provided elsewhere [17, 22]. The analysis of this study was restricted to all pregnant women age 15–49 in the women sample (n = 13,705). After dropping 227 and 1184 observations with missing information in the outcome and independent variables respectively, the final sample size for the study comprised 12,294 pregnant women. The dependent variable is the uptake of adequate (≥ 3) IPTp-SP, categorized following the WHO recommendation of IPTp-SP doses, women took during pregnancy in the year preceding the surveys. The variable was categorized as less than three doses of IPTp-SP, as inadequate uptake (i.e. < 3 doses = 0), and at least three doses or more ≥ 3 doses = 1, as adequate uptake [1, 6, 23]. The wealth index variable was used as a proxy for socioeconomic status. It was constructed using household ownership of selected assets (e.g. televisions and bicycles), materials used in housing construction, type of water access, and sanitation facilities data via a principal component analysis (PCA) [24]. Based on the current literature [6, 9, 23, 25, 26], the following variables were used as determinants of the uptake of IPTp-SP among pregnant women: age groups, the level of education, marital status, religion, occupation, place of residence (rural and urban), geopolitical zone, wealth index quintiles, husband/spouse level of education, distance to a health facility, and a requirement to obtain permission for self-medical help (defined as either a big problem or not a big problem) to access IPTp-SP [6, 23]. Additional file 1: Table S1 reports description of variables used in the analysis. This study used the concentration index (C) approach, an appropriate and most widely used measure of socioeconomic-related inequality [27–29], to quantify socioeconomic inequalities in IPTp-SP. The C is based on the concentration curve which graphs the cumulative share of the population on the x-axis and the cumulative share of the health outcome on the y-axis. The C index is defined as twice the area between the concentration curve and the line of perfect equality (45-degree diagonal) [29, 30]. The C can be computed using the convenient regression method as follows: where σr2 is the variance of the fractional rank, h is the healthcare variable of interest (i.e. IPTp-SP uptake) of i th woman, μ is the mean of the variable of interest, h, for the whole population, and ri=1N is the fractional rank of the i th woman in the distribution of socio-economic position, with i=1 for the poorest and i=N for the richest. The C is calculated as the ordinary least squares (OLS) estimate of β [29, 31]. The C ranges from − 1 to + 1, for continuous health outcomes. Since the outcome variable (IPTp-SP) of interest is binary, the minimum and maximum of the C are not between − 1 and + 1 but depend on μ [32]. Hence, the index can be normalized by multiplying the estimated C by 11-μ. The normalized concentration (Cn) index is used to quantify socioeconomic-related inequalities in uptake of adequate (≥ 3) IPTp-SP. If the value of the Cn is zero, it suggests that there is no socioeconomic-related inequality in health outcomes. A negative (positive) value of the Cn when the curve lies above (below) the line of equality indicates a disproportionate concentration of the health variable (i.e. IPTp-SP) among the poor (rich) [28, 29]. A higher value of the Cn corresponds to high socioeconomic inequalities [27]. The Cn is decomposed to quantify factors (demographic, geographic, and socioeconomic) that contribute to the observed socioeconomic inequalities in the uptake of adequate IPTp-SP following the Wagstaff, Van Doorslaer [33] approach. If there is a linear regression model to link the outcome variable (i.e. uptake of adequate IPTp-SP) h, to a set of k explanatory factors, xκ such as: where α and β are parameters that measure the relationship between each explanatory factor x and the uptake of adequate IPTp-SP and ε error term. Wagstaff, Van Doorslaer [33] showed that the C of h, can be decomposed into the contribution of determinants that explain the uptake of IPTp-SP during pregnancy as follows: where x- k is the mean of xk, and Ck denotes the concentration index for xk, a contributing factor. The GCε denotes the generalized concentration index of the error term, εi. Equation 3 shows that the overall inequality in the uptake of IPTp-SP has two components. The first term (βkxk-μ)CK denotes the contribution of factor k to socioeconomic inequality in the uptake of adequate IPTp-SP. It constitutes the deterministic or explained component of the IPTp-SP uptake of the concentration index. The second term GCεμ represents the unexplained component or the residual of the IPTp-Sp uptake [31, 34]. Based on Eq. 3, the product of the elasticity of each factor and its Ck gives the contribution of that factor to the inequality. The negative (positive) contribution of a predictor to the Cn suggests that the socioeconomic distribution of the predictor and the association between the predictor and the adequate uptake of IPTp-SP leads to an increase in the concentration of uptake of IPTp-SP among the poor (rich). A zero value of either elasticity or the Ck leads to the zero contribution of the factor to C [29, 35]. Applying the Wagstaff [32] normalization approach to the decomposition of the C can yield: The dataset was weighted using the primary sampling weight provided in the DHS to obtain estimates that are representative of all pregnant women in Nigeria. A survey logistic estimation on samples was conducted to check for collinearity before the decomposition analysis. Chi-square was used to test associations between socio-demographic characteristics and IPTp-SP uptake. The predictors of IPTp-SP uptake were considered statistically significant at p < 0.05. All data analyses were conducted using Stata/SE-13 software [36].