Menu
Menu
Menu
Menu
Introduction Antenatal care (ANC) is a vital mechanism for women to obtain close attention during pregnancy and prevent death-related issues. Moreover, it improves the involvement of women in the continuum of health care and to survive high-risk pregnancies. This study was conducted to determine the prevalence of and identify the associated factors of eight or more ANC contacts in Nigeria. Methods We used a nationally representative cross-sectional data from Nigeria Demographic and Health Survey—2018. A total sample of 7,936 women were included in this study. Prevalence was measured in percentages and the factors for eight or more ANC contacts were examined using multilevel multivariable binary logistic regression model. The level of significance was set at P<0.05. Results The prevalence of eight or more ANC contacts in Nigeria was approximately 17.4% (95% CI: 16.1%-18.7%). Women with at least secondary education were 2.46 times as likely to have eight or more ANC contacts, when compared with women with no formal education. Women who use media were 2.37 times as likely to have eight or more ANC contacts, when compared with women who do not use media. For every unit increase in the time (month) of ANC initiation, there was 53% reduction in the odds of eight or more ANC contacts. Rural women had 60% reduction in the odds of eight or more ANC contacts, when compared with their urban counterparts. Women from North East and North West had 74% and 79% reduction respectively in the odds of eight or more ANC contacts, whereas women from South East, South South and South West were 2.68, 5.00 and 14.22 times respectively as likely to have eight or more ANC contacts when compared with women from North Central. Conclusion The coverage of eight or more ANC contacts was low and can be influenced by individual-, household-, and community-level factors. There should be concerted efforts to improve maternal socioeconomic status, as well as create awareness among key population for optimal utilization of ANC.
We used a nationally representative cross-sectional data. The individual woman questionnaire in Nigeria Demographic and Health Survey (NDHS) was analyzed in this study. A total sample of 7,936 women of reproductive age who became pregnant and had given birth after the new guideline of eight ANC contacts was endorsed were included in this study. The 2018 NDHS is the sixth survey of its kind to be implemented by the National Population Commission (NPC). Data collection took place from 14 August to 29 December 2018. The sample was selected using a stratified, two-stage cluster design, with Enumeration Areas (EAs) as the sampling units for the first stage. The complete listing of households carried out in each of the 1,389 selected EAs, an approximate number of 30 households was selected in every cluster resulting to a total of 41,821 women were interviewed during the survey, yielding a response rate of 99%. A total sample of 7,936 women of reproductive age who became pregnant and had given birth after the new guideline of eight or more ANC contacts was endorsed by WHO [8], were included in this study. In particular, NDHS 2018 used a three-stage sampling stratification, in which respondents were first stratified by urban versus rural dwelling, and EAs were then selected randomly within each stratum. Finally, households within each EA were then selected for the survey using equal probability sampling. This three-stage sampling method was taken into account in the computation of survey weights, applied to ensure the representativeness of the sample with regard to the general population. The sampling frame used for the 2018 NDHS is the Population and Housing Census of the Federal Republic of Nigeria (NPHC), which was conducted in 2006 by the National Population Commission. The sample for the 2018 NDHS was a stratified sample selected in two stages. Stratification was achieved by separating each of the 36 states and the Federal Capital Territory into urban and rural areas. In total, 74 sampling strata were identified. Data for this study are derived from the individual female data for analysis. The DHS project, funded primarily by the United States Agency for International Development (USAID) with support from other donors and host countries, has conducted over 230 nationally representative and internationally comparable household surveys in more than 80 countries since its inception in 1984. The data is available in the public domain and accessed at; http://dhsprogram.com/data/available-datasets.cfm. Details of DHS sampling procedure has been reported previously [20]. The frequency of ANC contacts with doctors, nurses and midwives was measured dichotomously; less than eight ANC contacts vs. eight or more ANC contacts. The WHO ANC guideline recommendations mapped to the eight recommended contacts, presents a summary framework for the 2016 WHO ANC model in support of a positive pregnancy experience [8,21,22]. Family mobility: internal immigrant (if a respondent lived in the current location in less than 5 years) vs. native (if a respondent had lived in the current location at least 5 years). Religious background: Christianity, Islam and African Traditional Religion (ATR)/others. Literacy: cannot read at all, able to read only part of a sentence and able to read whole sentence. Total number of children ever born: 1–2, 3–4 and over 4 children. Women’s knowledge level was measured using; educational attainment, read newspaper/magazines, listen to radio, watch television and use internet [23]. Using Principal Component Analysis (PCA), the standardized z-score was used to disentangle the overall assigned scores to low, medium and high. Maternal educational attainment: no formal education, primary and secondary or higher education. Media use was measured dichotomously (yes vs. no) if a respondent used any or newspaper/magazine, radio, television or internet irrespective of the frequency levels, "almost every day", "at least once a week", and "less than once a week" as yes/use and the response level "not at all" as no/not use [24]. Maternal age: 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49. Wanted child when became pregnant: then, later and wanted no more. Health insurance: covered vs. not covered. Marital status: never in union, currently married/living with a partner and formerly in union. Employment status: working vs. not working. Family type: monogyny vs. polygyny. Intimate partner violence: yes (if a woman had physical, sexual or emotional violence) vs. no (otherwise). Women’s autonomy was measured using PCA for selected items: person who usually decides on respondent's health care, person who usually decides on large household purchases and person who usually decides on visits to family or relatives [25]. The standardized z-score was then used to disentangle the overall assigned scores to low, medium and high. Time to ANC initiation (in months). Sex of household headship was male vs. female. Household size was based on the total number of individuals who resided together and grouped as: 1–4, 5–8 and over 8 persons. Household wealth quintiles: PCA was used to assign the wealth indicator weights. This procedure assigned scores and standardized the wealth indicator variables such as; bicycle, motorcycle/scooter, car/truck, main floor material, main wall material, main roof material, sanitation facilities, water source, radio, television, electricity, refrigerator, cooking fuel, furniture, number of persons per room. The factor coefficient scores (factor loadings) and z-scores were calculated. For each household, the indicator values were multiplied by the loadings and summed to produce the household’s wealth index value. The standardized z-score was used to disentangle the overall assigned scores to; poorest/poorer/middle/richer/richest categories [26,27]. In creating household wealth index, rural-urban differences was adjusted for and used in the analysis. As a response to criticism that a single wealth index is too urban in its construction and not able to distinguish the poorest of the poor from other poor households, the new variable created to provide an urban- and rural-specific wealth index was utilized. We used EAs to represent communities prominently because the DHS did not collect aggregate-level data at the community level. Hence, community-level variables included in the analysis were based on women’s characteristics particularly those that have implications for accessing ANC. Cultural norms about wife-beating was created by aggregating responses from women in each community. Here, we used the items: “beating justified if wife goes out without telling husband”, “beating justified if wife neglects the children”, “beating justified if wife argues with husband”, “beating justified if wife refuses to have sex with husband” and “beating justified if wife burns the food”. Finally, a binary variable was created for acceptance of wife beating [28]. Maternal residential status was measured as: urban vs. rural. Geographical region was categorized thus: North Central, North East, North West, South East, South South and South West. Furthermore, aggregate community-level variables were constructed by aggregating individual level characteristics at the community (cluster) level and categorization of the aggregate variables was done as low or high based on the distribution of the proportion values calculated for each community. If the aggregate variable was normally distributed mean value and if not normally distributed median value was used as cut off point for the categorization. Community-level poverty was categorized as high if the proportion of women from the two lowest wealth quintiles in a given community was 43–100% and low if the proportion was 0–42%. Community-level media use was categorized as high if the proportion was 60–100% and as low if the proportion of women who use media in the community was 0–59%. Community-level illiteracy was categorized as high if proportion of women who cannot read at all was 67–100% and as low if the proportion of women who cannot read at all was 0–66%. Community-level urban residence was categorized as high if proportion of women who reside in urban area was greater than 1–100% and as low if the proportion of women who reside in urban area was 0%. Community-level women’s autonomy was categorised as high if the proportion of women who had at least moderate autonomy was 61–100% and categorized as low if the proportion was between 0–60%. This approach was used in a previous study [29,30]. In this study, we utilized population-based secondary datasets available in public domain/ online with all identifier information removed. The authors were granted access to use the data by MEASURE DHS/ICF International. DHS Program is consistent with the standards for ensuring the protection of respondents’ privacy. ICF International ensures that the survey complies with the U.S. Department of Health and Human Services regulations for the respect of human subjects. No further approval was required for this study. More details about data and ethical standards are available at http://goo.gl/ny8T6X. The survey (‘svy’) module was used to adjust for stratification, clustering and sampling weights to compute the estimates of eight or more ANC contacts. The prevalence of eight or more ANC contacts was explored using percentage. A cut-off of 0.7 was used to determine multicollinearity known to cause major concerns in the logit model [31]. Consequently, maternal literacy and knowledge were excluded from the model as they were found to have positive interdependence with educational attainment which was therefore retained in the model. Other significant variables from Chi-square test or student’s t-test at 25% level of significance were retained in the logit model in the absence of multicollinearity. A multivariable multilevel binary logistic regression model was used to estimate the fixed and random effects of the factors associated with eight or more ANC contacts. We specified a 3-level model for binary response reporting eight or more ANC contacts, for women (at level 1), in a household (at level 2) from an Enumeration Area (at level 3). We constructed five models. The first model, an empty or unconditional model without any explanatory variables, was specified to decompose the amount of variance that existed between community and household levels. The null or empty model is important for understanding the community and households’ variations, and we used it as the reference to estimate how much household and community factors were able to explain the observed variations. In addition, we used it to justify the use of multilevel statistical framework, because if the community variance was not significant in the empty model, it advised to use the single-level logistic regression. The second model contained only individual-level factors, the third model contained only household-level factors, and the fourth model contained only community-level factors. Finally, the fifth model simultaneously controlled for individual, household and community level factors (Full model). Statistical significance was determined at p< 0.05. The Bayesian and Akaike Information Criterions were used to select the best model out of the five models. A lower value on Akaike or Bayesian Information Criterion indicates a better fit of the model [32]. Data analysis was conducted using Stata Version 14 (StataCorp., College Station, TX, USA). The results of fixed effects (measures of association) were reported as adjusted odds ratios (AORs) with their 95% confidence interval (CI). The probable contextual effects were measured by the Intra-class Correlation (ICC) and Median Odds Ratio (MOR) [33]. We measured the similarity between respondents in the same household and within the same community using ICC. The ICC represents the percentage of the total variance in the probability of eight or more ANC contacts that is related to the household and community level, i.e. measure of clustering of odds of eight or more ANC contacts in the same household and community. The MOR measures the second or third level (household or community) variance as odds ratio and estimates the probability of eight or more ANC contacts that can be attributed to household and community context. MOR equal to one indicates no household or community variance. Conversely, the higher the MOR, the more important are the contextual effects for understanding the probability of eight or more ANC contacts. The ICC was calculated by the linear threshold according to the formula used by Snijders and Bosker [34], whereas the MOR is a measure of unexplained cluster heterogeneity.
N/A
