Utilization of optimal antenatal care, institutional delivery, and associated factors in Northwest Ethiopia

Optimal antenatal care visits (ANC4+) and institutional delivery are essentials to save lives of the women and the baby during pregnancy and delivery. Though focused antenatal care visits and institutional delivery is recommended by World Health Organization, Ethiopia has sub-optimal antenatal care and lagged facility delivery. A community-based cross-sectional study was conducted among 811 lactating women in Northwest Ethiopia. Multivariable logistic regression analysis was performed using 95% confidence level and p < 0.05. The prevalence of optimal antenatal care visits and institutional delivery were 39.6% (95%CI: 36.2–43) and 62.6% (95%CI: 59.2–66), respectively. Maternal education (AOR = 2.05; 95%CI: 1.14, 3.69), home visiting by health extension workers (AOR = 1.57; 95%CI: 1.01, 2.29), and early antenatal care booking (AOR = 11.92; 95%CI: 8.22, 17.31) were significant predictors of optimal antenatal care. Exposure to mass media (AOR = 1.65; 95% CI: 1.02, 2.65); intended pregnancy(AOR = 1.68; 95%CI:1.12, 3.63); parity of one (AO = 3.46; 95% CI: 1.73, 6.89); 1–3 antenatal care visits (AOR = 2.17; 95% CI: 1.29, 3.63); and ANC4 + (AOR = 3.57; 95% CI: 2.07, 6.14); history of pregnancy-related complications(AOR = 1.63; 95%CI: 1.04, 2.57), and access to transportation to reach a health facility(AOR = 1.58; 95%CI: 1.00, 2.45) were significant predictors of institutional delivery. Addressing the modifiable factors identified in this study could improve optimal antenatal care visit and institutional delivery. This study was conducted in two districts of the Central Gondar Zone, Northwest Ethiopia. The districts are located 658 km from Addis Ababa, the capital city of Ethiopia. There were 16 health centers and 88 health posts in the districts. The total population was 524, 907 (female = 260,879 and male = 264,028) at the time of the survey, of which 122,303 and 16,745 were women in the reproductive age group and surviving infants, respectively. The study period of this study was from October to November, 2020. A community-based cross-sectional study was conducted. All lactating women who gave birth within the past 12-months and living in the study districts were the target population. A sample of lactating women who gave birth within the past 12-months who were permanent residents of the selected kebeles were the study population. Both the single population proportion and the double population proportion formulas were considered in estimating different sample sizes. For the single population proportion formula, the assumptions used were: prevalence of ANC attendance7, prevalence of completion of four or more antenatal care visits7, the proportion of mothers who received delivery service from health care providers after completing antenatal care four and above and the proportion of women retained in the continuum of maternal care7, considering 95% CI, the margin of error = 5%, design effect of 2 and 10% non-response rate. For double population proportion formulas, the sample size was computed using the STAT-CALC program of Epi-info version 7.0. software using the following assumptions: 5% level of significance (two-sided), 80% power, and a 1:1 ratio of exposed to non-exposed and odds ratio. Factors such as being a model household in the community14, distance to reach a health facility15, and antenatal care visits16 were considered predictor variables. The design effect of 2 and 10% non-response rate were used during different sample size estimations. Of the different sample sizes estimated, the maximum sample size, 811, was obtained from a single population formula and considered in the study. A multi-stage cluster sampling technique was employed to reach study participants. Wogera and Gondar zuriya districts were selected randomly among the six transformation woredas in the central Gondar zone. Specifically, the primary sampling units were kebeles with the respective health centers while the secondary sampling units were women who gave birth during the past 12 months. The health centers (or the respective kebeles) were selected with a simple random sampling technique. To select sample participants, a sampling frame of mothers was developed using a list of eligible women from the community health information system family folder(pouch). Then the sampling interval was obtained by dividing the source population (8393) by the estimated total sample size (811). The first study participant (the 2nd mother) was selected by a simple random sampling technique from the first sampling interval, and all other mothers were selected systematically by taking every 10th mother in the frame. When two or more women were found at a single household level, one of them was selected randomly and included in the study. Seriously ill or women who were unable to speak were excluded from the study. For those women who were selected for an interview and were not available during data collection, we waited for them for three to five days to get them. If they were not still accessible, they would be considered non-respondent. The outcome variables of this study were optimal antenatal care (ANC4 +), and institutional delivery. The optimal antenatal care was classified as “yes” if the women had optimal antenatal care. i.e., attending four or more antenatal care visits during pregnancy, or “no” if they had less than four antenatal care visits. Similarly, women who gave birth at health facilities (health centers or hospitals) were categorized as “yes” or “no” if otherwise. Explanatory variables were the age of the respondents, education status of respondents, education status of partner, intendedness of the current pregnancy, being a member of community-based health insurance, mass-media exposure, health extension workers' home visits during pregnancy, being a member of Women’s Development Army, early ANC booking, pregnancy-related complications, planned pregnancy, frequency of ANC visits, parity, access to transportation, and household wealth index. For wealth index assessment, principal component analysis was employed to reduce data that measure the same construct together, and the data were recoded into binary variables. The wealth quantile was categorized into five categories: poorest, poor, middle, richer, and richest. Interviewer-administered questionnaire was used to collect the household data. The tool was prepared in English after a thorough literature review and translated into the local language (Amharic) and then finally returned to English. Since experts’ views were sought for the psychometric properties of the face and content validity, experts were invited to review the relevance of each question in the tool. The tool was refined according to the experts view and piloted out of the study setting and validated before data collection. The proportions of the rated item content validity index and scale content validity index were 0.98 and 0.81, respectively. Six data collectors and two supervisors were recruited for data collection and supervision. The data were collected using interviewer administered questionnaire via face-to-face interview. A two-day training was given to data collectors and supervisors regarding the objective of the study, procedures of the data collection, data collection tools and handling of the data. During the data collection period, daily supervision was conducted by the supervisors. Data were entered into EpiData 4.02 and then exported to STATA 14 software for statistical analysis. Principal Component Analysis (PCA) was constructed to assess household wealth status, and the items were recoded into two categories (0 = no, 1 = yes) to reduce the amount of data that measure the same construct. Twenty-one items composed of household productive and non-productive assets and utilities were entered for analysis. The frequencies of the assets greater than 95% and less than 5% were excluded from the analysis as it would not help to identify the richer or poorer category. The scale reliability coefficient and Kaiser–Meyer–Olkin measure of sampling adequacy were employed to assess the satisfaction of the assumptions for PCA. The wealth quantile was categorized into five categories: poorest, poor, middle, richer, and richest. To identify factors associated with response variables, bivariable logistic regression analysis was fitted. Likewise, the model fitness was checked using the Hosmer and Lemeshow goodness of fitness test. The correlation of independent variables was checked by using the Variance Inflation Factor (VIF), in which VIF > 10 indicates substantial multicollinearity, and there was no multicollinearity effect among predictor variables. Descriptive statistics such as frequency and cross-tabulation were computed. In the bivariable logistic regression model, a p < 0.25 was used to recruit candidate variables for multivariable model adjustment, and the selected variables were entered sequentially by using backward stepwise regression. An association between outcome variables and explanatory variables was presented using an adjusted OR with a 95% confidence interval. A significance level of p < 0.05 was considered to declare the level of association with women’s optimal antenatal care visits and institutional delivery. Study approval and ethical clearance were obtained from the University of Gondar ethical review board (R.NO. V/P/RCS/05/2020). A formal letter of approval was taken from Amhara national regional health state bureau and central Gondar zonal health department. Written informed consent was taken from participants and their parents and assent obtained from the minor/participant. They were informed about the objective, importance of the study, procedure and duration, risk and discomfort, benefits of participating in the study, confidentiality, and the right to refuse or withdraw during data collection. After obtaining the relevant information, participants were counselled on the benefits of attending maternal healthcare services and the consequences of missing maternal health care services. All methods were carried out in accordance with relevant guidelines and regulations.