Determinants of timing, adequacy and quality of antenatal care in Rwanda: a cross-sectional study using demographic and health surveys data

This study is a cross-sectional study using secondary data from three waves of Rwanda demographic and health survey (RDHS). The three waves include RDHS 2010, RDHS 2015 and RDHS 2020. The RDHS is a cross-sectional study using a stratified two-stage sampling design in which rural and urban place of residence are regarded as strata [15, 28, 29]. The census enumeration areas are considered as clusters and a full list of all households was later used as a sampling frame to choose which households should be interviewed. A nationally representative household sample is finally collected. The response rate has been high, above 99% for women across the three waves. The RDHS collects data on maternal and child health services covering a period within the preceding 5 years of the survey. Details on sampling design, sample size, study instruments, data collection, informed consent, and other associated procedures can be found elsewhere [15]. The RDHS data are accessible from the Measure DHS website at http://​dhsprogram.​com/​data/​available-datasets.​cfm.

For the purpose of this study, the 2010, 2015, and 2020 RDHS individual recode (IR) datasets were merged based on established guidelines for managing DHS data. Women aged 15 to 49 years’ old who had a live birth in the five years prior to each survey and answered questions about ANC were included in this sample. Women with missing values or invalid responses to the key exposure, outcome, and possible confounders, such as “don’t know”, were removed. 18,034 of the 41,802 women who took part in the survey met the requirements for inclusion. More information is available in Fig. 1.

All the statistical analyses were conducted using Stata v14.0 [39]. Descriptive statistics for the sociodemographic characteristics of the study participants were generated by means of frequency and percentage as shown in Table 1.We used chi-square tests to identify demographic and socio-economic factors associated with each outcome variable. Crude odds ratios were generated by means of bivariate analyses to determine the odds of each outcome variable with explanatory variables. Potential factors with p < 0.20 [40, 41] in the bivariate analysis were retained for multivariate logistic regression in the final model. When covariates found to be collinear (r >  = 0.8, using Pearson correlation test), the variable that was most correlated with the outcome variable of interest was retained. To account for clustering, stratification, and sample weight, we weighted all analyses using the survey module “svyset” stata commands.

Table 1 summarizes a sample of 18,034 women aged 15–49. About 9159(51%) of the respondents were aged between 25 and 34 years, 14,486(80.39%) of the women were employed. A larger number of respondents 12,472(69.38%) had a primary level education and 11,194(62.04%) had access to media at least once a week. Rural habitants were 14,553(84.22%) and 7833(43.58%) were poor while the rich were 6756(36.77%). The percentage for married or partnered women is 14686(81.62%) and 11,659(79.29%) of respondents were involved in making decisions about healthcare. Mistimed and unwanted pregnancies was among 7425(41.15%) of the mothers.

Timely first ANC visit of the women was 9364(51.73%) whereas only 7751(42.85%) had adequate ANC visits. Women who received services components of ANC were 3621(23.75%) and only 1494(8.88%) achieved the high-quality ANC. A greater number of women 16,526(92.26%) had their ANC at a health center.

Figure 2 shows that mothers who have secondary and higher education level had 50% more odds of timely first ANC visit (aOR:1.50;95% CI:1.29,1.74) compared to the uneducated mothers. Mothers who attended their ANC visit at a clinic had above three times the odds of timely first ANC visit (aOR:3.32;95%CI:2.32,4.75) compared to those who went to the health center to seek ANC services and those who had their ANC visit at a referral hospital had nearly 50% more odds of timely first ANC visit(aOR:1.52;95%CI:1.09,2.12) than those who had ANC visit at a health center. Rich mothers and those who perceived the distance to the health facility as not being a problem had above 10% more odds of timely first ANC visit (aOR: 1.20;95%CI:1.09,1.32& aOR: 1.16;95%CI:1.07,1.25) than poor mothers and those who perceived the distance to the health facility as a big problem respectively. Mothers whose husband’s had a secondary or higher education level had 37% more odds of timely first ANC visit (aOR: 1.37;95%CI:1.19,1.57) than the mothers whose husbands were uneducated.

Mothers who had mistimed and unwanted pregnancies had less odds of timing their first ANC visit (aOR:0.67;95%CI:0.62,0.73 and aOR:0.76;95%CI:0.68,0.85 respectively) compared to the mothers who wanted their pregnancy. Higher birth order of the child was negatively associated with the timing of the first ANC visit, and the higher the birth order, the less was the timing of the first ANC visit (aOR:0.63;95%CI:0.55,0.72 and aOR:0.50;95%CI:0.43,0.57 for 3^rd and 4^th or more respectively).

Figure 2 shows that early middle-age was positively associated with the adequacy of ANC when referred to teen mothers 15–19 years; aOR:1.67;95%CI:1.19,2.34 and aOR:1.57;95%CI:1.11,2.22 for women ages of 35–39 and 40 + years older. Mothers whose education level were secondary or higher had 19% more odds of having adequate ANC visits (aOR:1.19;95%CI:1.03,1.38) than uneducated mothers. Mothers who perceived the distance to health facility as not a problem had 13% more odds of going for 4 or more ANC visits (aOR:1.13;95%CI:1.05,1.23) compared to those who perceived the distance to health facility as a big problem. Rich women had 20% more odds of having adequate ANC visits (aOR: 1.20;95%CI:1.09,1.31) compared to the poor women. Women who went for ANC at a referral hospital had twice the odds of completing the recommended ANC visits (aOR:2.07;95%CI:1.51,2.83) compared to those who went at a health center, similarly the mothers who took their ANC visit at a clinic had 3 times the odds of completing the recommended ANC visits (aOR:3.38;95%CI:2.48,4.59) than those who went at the health center.

Mothers who mistimed and those who had unwanted pregnancies had less odds of completing 4 or more ANC visits (aOR:0.73;95%CI:0.67,0.79 and aOR:0.78;95%CI:0.70,0.87 respectively).

Birth order of the child was negatively associated with completing the recommended ANC visits (aOR:0.68;95%CI:0.60,0.77 and aOR:0.54;95%CI:0.48,0.62 for third and fourth order respectively), the higher the birth order of the child the less the odds of completing recommended ANC visits.

Figure 3 shows that only 24.36% of the women who participated in the study got all services components of ANC. Maternal education increased the odds of receiving services component of ANC (aOR:1.44;95%CI:1.18,1.76). The increase in the husband’s education increased the odds of receiving services component of ANC for the woman. The odds of receiving services components of ANC by women whose husbands had secondary education level were 30% higher than that of women whose husbands did not have formal education (aOR:1.30;95%CI:1.07,1.57). Rich women had 18% more odds of receiving the services components of ANC (aOR:1.18;95%CI:1.04,1.34) than the poor women. Child unwantedness was found to be negatively associated with receiving services components of ANC services (aOR:0.67;95%CI:0.57,0.79). Older age was negatively associated with receiving services components of ANC (aOR:0.73;95%CI:0.42,1.27& aOR:0.44;95%CI:0.25,0.77) for 35 to 39 years old and 40 years and above respectively.

Maternal education was found to be positively associated with high-quality ANC. Mothers with secondary or higher education level had 50% more odds of receiving high-quality ANC (aOR:1.50;95%CI:1.15,1.96) than uneducated women. The husband’s education was found to be positively associated with receiving high-quality ANC (aOR:1.38;95%CI:1.07,1.77). Women in the rich tercile were positively associated with achieving high-quality ANC (aOR:1.26;95%CI:1.06,1.49). Women in southern (aOR:1.70;95%CI:1.36,2.14) and northern (aOR:1.60;95%CI:1.26,2.04) provinces had more odds of receiving high-quality ANC than women in other provinces in Rwanda. Mothers who mistimed (aOR:0.69;95%CI:0.59,0.79) and those with unwanted pregnancies(aOR:0.65;95%CI:0.51,0.82) had less odds of receiving high quality ANC.

Figure 4 shows that all outcome indicator variables have been increasing over the survey period from 2010 to 2020. We found that only the timing of ANC was above 50% in 2015 and 2020. The high-quality ANC is very low.

Figure 5 shows that except for tetanus injection which did not steadily increase, there was an increase in ANC services provided over the survey years.

This is the first study to make a comprehensive assessment of the quality of ANC in Rwanda that accounts for not just the timing of the first visit and adequacy of ANC visits but also the ANC services delivered to the women and the qualification of the ANC providers. A major strength of the study is using a nationally representative population-based dataset with socio-economic variables on healthcare-seeking behavior. We excluded women who missed data across confounders, this might have produced a missing data bias. However, the bias is negligible since missingness is below 5%. It is possible that women may not have recognized if they received any of the services component of ANC but may have disclosed them because of social acceptability bias. The variable tracking the place of ANC also could not adequately account for potential variations in the health facility that women might have visited. To make the data easier to analyze, all the women who might have had ANC visits at different types of health facilities were grouped into one category. Similar to this, the ANC provider utilized in the analysis might not have been the one the women visited exclusively during the course of the pregnancy. The last stages of data collection took place during COVID-19 and the process had to stop for more than two months [15]. With the country in a total lockdown to contain the pandemic, women missed ANC visits and this might have slightly affected the findings. Findings related to inadequate, delays and lower quality ANC highlight the areas requiring health policy prioritization. This study did not include variables such as facility readiness, interpersonal relationships between providers and women, transportation and other cultural norms and beliefs that could have been determinants of ANC utilization. This study being cross-sectional could not determine cause and effect and were limited to the assessment of associations. However, the pooled analysis of three surveys may provide a robust estimate of those influencing factors. Further research would conduct a longitudinal study design to explore the risk factors of the timing, the adequacy and the quality of antenatal care in Rwanda.