Maternal healthcare utilization and full immunization coverage among 12–23 months children in Benin: a cross sectional study using population-based data

Background: Maternal and child health are important issues for global health policy, and the past three decades have seen a significant progress in maternal and child healthcare worldwide. Immunization is a critical, efficient, and cost-effective public health intervention for newborns. However, studies on these health-promoting indicators in low-income and middle-income countries, especially in sub-Sahara Africa are sparse. We investigated the association between maternal healthcare utilization and complete vaccination in the Republic of Benin. Methods: We analysed data from the 2018 Benin Demographic and Health Survey (BDHS). Specifically, the children’s recode file was used for the study. The outcome variable used was complete vaccination. Number of antenatal care visits, assistance during delivery, and postnatal check-up visits were the key explanatory variables. Bivariate and multilevel logistic regression analyses were carried out. The results were presented as unadjusted odds ratios (uOR) and adjusted odds ratios (aOR), with their corresponding 95% confidence intervals (CIs) signifying their level of precision. Statistical significance was declared at p < 0.05. Results: The prevalence of full immunization coverage in Benin was 85.4%. The likelihood of full immunization was lower among children whose mothers had no antenatal care visits, compared to those whose mothers had 1–3 visits [aOR = 0.11, 95% CI: 0.08–0.15], those who got assistance from Traditional Birth Attendants/other during delivery, compared to those who had assistance from Skilled Birth Attendants/health professionals [aOR = 0.55, 95% CI: 0.40–0.77], and mothers who had no postnatal care check-up visit, compared to those who had postnatal care check-up < 24 h after delivery [aOR = 0.49, 95% CI: 0.36–0.67]. With the covariates, religion, partner’s level of education, parity, wealth quintile, and place of residence also showed significant associations with full immunization. Conclusion: The study has demonstrated strong association between full immunization and antenatal care, skilled attendance at birth, and postnatal care check-up visit. We found that full immunization decreases among women with no antenatal care visits, those who receive assistance from Traditional Birth Attendants during delivery, and those who do not go for postnatal care visits. To help achieve full immunization, it is prudent that the government of Benin collaborates with international organisations such as WHO and UNICEF to provide education to pregnant women on the importance of immunization after delivery. Such education can be embedded in the antenatal care, delivery and postnatal care services offered to pregnant women during pregnancy, delivery, and after delivery. The study used data from the 2018 Benin Demographic and Health Survey (BDHS). Specifically, the birth recode file was used. The survey focuses on essential maternal and child health markers, including immunization [32]. To ensure accuracy of data collection, data collection was done by survey staff who were trainees and were given instructions in standard DHS procedures. These procedures included general interviewing techniques, conducting interviews at the household level, and review of each question and mock interviews between participants. To ensure participants understood the questions being asked, the definitive questionnaires were first prepared in English and subsequently translated by experts into the major local languages at the various data collection points. Interviews were also conducted in local languages. As part of quality assurance, pretest training and field practice of the DHS survey protocol and instruments were done. Field staff were further given training before the actual data collection to ensure that they were able to gain accurate understanding of the data collection instruments. The DHS questions were also standardized, making it possible to do cross-country studies. The surveys employed a two-stage stratified sampling technique, which makes the survey data nationally representative [33]. The first-stage involved a listing of primary sampling units (PSUs) or enumeration areas (EAs) that covered the entire country and usually were obtained from the latest national census. These EAs are also known as the clusters. Each EAs was further subdivided into standard size segments and a sample of predetermined segments were selected randomly with probability proportional to the number of households in each EA. In the second stage, households were systematically selected by surveying the personnel from a list of previously enumerated households in each selected EA segment, and in-person interviews were conducted in selected households to target populations: women aged 15–49, men aged 15–64, and children under 5. In this study, a total of 4156 married and cohabiting women who had complete information on all the variables of interest were considered as the sample size. The dataset can be accessed at https://dhsprogram.com/methodology/survey/survey-display-491.cfm. We relied on the Strengthening the Reporting of Observational Studies in Epidemiology’ (STROBE) statement in conducting this study and writing the manuscript. The study used complete vaccination as the outcome variable. In this study, complete and full immunisation coverage are used interchangeably. The information on vaccination coverage was collected from either immunization cards or from mothers’ verbal responses to these questions “Did (NAME) ever receive vaccination against Measles?”, “Did (NAME) ever receive vaccination against Polio?”, “Did (NAME) ever receive vaccination against BCG?”, and “Did (NAME) ever receive vaccination against DPT?”. Responses were “Yes”, “No” and “Don’t Know.” These responses were coded as “No” = 0, “Yes = 1″ and “Don’t Know = 8″. For the purpose of the analysis, only women who provided definite responses (either “Yes” or “No”) were included in the study. According to the WHO guideline [34], “complete or full immunization” coverage is defined as a child that has received one dose of BCG, three doses of pentavalent, pneumococcal conjugate (PCV), oral polio vaccines (OPV); two doses of Rota virus, and one dose of measles vaccine. We recoded each variable (vaccinations) as “0″ and “1″ for children who didn’t take the recommended doses and those who took respectively. The complete vaccination was obtained by creating a composite variable which comprised all the vaccines administered. To provide a binary outcome, the two responses were coded as follows: “Incomplete” = 0, “Complete = 1″. The study used three indicators as key independent variables. These indicators are number ANC visits, assistance during delivery and the period when postnatal checks were done. The choice of these variables was based on their consideration as the key components of maternal healthcare in the DHS as well as their significant associations with full vaccination in previous studies [35–41]. These variables were generated from the questions, “how many times did you attend ANC during pregnancy?”, “who assisted in the delivery of your baby?” and “when after birth did PNC checks occur?”. Number of ANC visits was coded as ‘0, 1-3 and 4+’. Delivery assistance was coded as ‘Traditional Birth Attendants (TBA)/others and Skilled Birth Attendants (SBA)/health professional. The PNC check-up visits time was coded into ‘No, <24 hours, and >=1 day’. Eighteen control variables were considered. These variables were categorised into two broad factors. These factors were individual level (i.e., sex of the child, size of the child at birth, type of delivery, twin status, maternal age, marital status, occupation, mother’s education, partner’s education, religion, ethnicity, parity, frequency of reading newspaper or magazine, frequency of listening to radio, and frequency of watching television), and contextual level factors (i.e., wealth index, place of residence, and region) These variables were considered because of their statistically significant relationships with the full vaccination in previous studies [42–45]. Data were processed and analyzed using Stata version 14.0 with the use of both inferential and descriptive statistics. Prior to the data analysis, data cleaning was done and missing data in the form of blanks. The screening processes explained in the DHS as representing not applicable for the respondent either because the question was not asked in a particular country or because the question was not asked of this respondent due to the flow or skip pattern of the questionnaire were observed for ANC (167 cases), postnatal care (244), partner’s educational level (213), and birth size (43). These missing data were deleted through listwise deletion in order to get respondents with complete cases for the analyses. The analyses were done in three steps. In the first step, descriptive statistics (frequency and percentages) were used to describe the characteristics of the respondents (see Table 1). Next, a bivariate logistic regression analysis was conducted to examine the unadjusted relationship among maternal healthcare utilization, individual, contextual level factors, and full immunization. The results were presented as unadjusted odds ratios (uOR). All the variables that showed statistical significance (p < 0.05) were moved to the third step of the analysis. The third step was the use of multilevel models to examine the association between maternal healthcare utilization and full vaccination, while controlling for the individual and contextual level factors. The unit of analysis was married and cohabiting women who had at least one birth at the time of the survey. Five multilevel models (Model 0, I, II, III and IV) were built using only variables which were significant from the bivariate analysis in the second step of the analysis. Model 0 was regarded as the empty model that showed the variations in full immunization without any of the explanatory variables. Statistical significance at this level provides the basis for the use of multilevel models. In Model I, the key independent variables (ANC, assistance during delivery and PNC check-up visits) were included. Model II contained only the individual level variables, Model III had only the contextual level variables, and Model IV had the key independent, individual, and the contextual level variables. To ensure the accuracy of the results, a sensitivity analysis was performed using single women (never married, widowed, separated, and divorced). The results were presented in supplementary Table S1, with adjusted odds ratios (aOR) and their corresponding 95% confidence intervals signifying the level of precision. Statistical significance was declared at p < 0.05. For comparing models, the Akaike’s Information Criterion (AIC) and the log likelihood tests were used. The lowest AIC and the highest log likelihood ration were used to determine the best fit model. Sample weight was applied and the survey command (svy) was used to account for the complex sampling design of the survey. Full immunization in 12–23 months children in Benin and unadjusted Odds Ratio by explanatory variables (Weighted) uOR unadjusted Odds Ratio, CI Confidence Interval 1 = reference category * p < 0.05, ** p < 0.01, *** p < 0.001 Source: 2018 Benin Demographic and Health Survey The 2018 BDHS report indicated that ethical approval was granted by the ICF Institutional Review Board. Both written and oral Informed consent was sought from all the participants during the data collection exercise including the emancipated adults (i.e those below 16 years). We requested for the dataset on 10th March, 2020 and was granted access. It was downloaded and kept safe from third parties using ‘my lock box’ after permission was granted.