Introduction: Although ANC services are increasingly available to women in low and middle-income countries, their inadequate use persists. This suggests a misalignment between aims of the services and maternal beliefs and circumstances. Owing to the dearth of studies examining the timing and adequacy of content of care, this current study aims to investigate the timing and frequency of ANC visits in Ethiopia. Methods: Data was obtained from the nationally representative 2011 Ethiopian Demographic and Health Survey (EDHS) which used a two-stage cluster sampling design to provide estimates for the health and demographic variables of interest for the country. Our study focused on a sample of 10,896 women with history of at least one childbirth event. Percentages of timing and adequacy of ANC visits were conducted across the levels of selected factors. Variables which were associated at 5% significance level were examined in the multivariable logistic regression model for association between timing and frequency of ANC visits and the explanatory variables while controlling for covariates. Furthermore, we presented the approach to estimate marginal effects involving covariate-adjusted logistic regression with corresponding 95%CI of delayed initiation of ANC visits and inadequate ANC attendance. The method used involved predicted probabilities added up to a weighted average showing the covariate distribution in the population. Results: Results indicate that 66.3% of women did not use ANC at first trimester and 22.3% had ANC less than 4 visits. The results of this study were unique in that the association between delayed ANC visits and adequacy of ANC visits were examined using multivariable logistic model and the marginal effects using predicted probabilities. Results revealed that older age interval has higher odds of inadequate ANC visits. More so, type of place of residence was associated with delayed initiation of ANC visits, with rural women having the higher odds of delayed initiation of ANC visits (OR = 1.65; 95%CI: 1.26–2.18). However, rural women had 44% reduction in the odds of having inadequate ANC visits. In addition, multi-parity showed higher odds of delayed initiation of ANC visit when compared to the primigravida (OR = 2.20; 95%CI: 1.07–2.69). On the contrary, there was 36% reduction in the odds of multigravida having inadequate ANC visits when compared to the women who were primigravida. There were higher odds of inadequacy in ANC visits among women who engaged in sales/business, agriculture, skilled manual and other jobs when compared to women who currently do not work, after adjusting for covariates. From the predictive margins, assuming the distribution of all covariates remained the same among respondents, but everyone was aged 15–19 years, we would expect 71.8% delayed initiation of ANC visit. If everyone was aged 20-24years, 73.4%; 25-29years, 66.5%; 30-34years, 64.8%; 35-39years, 65.6%; 40-44years, 59.6% and 45-49years, we would expect 70.1% delayed initiation of ANC visit. If instead the distribution of age was as observed and for other covariates remained the same among respondents, but no respondent lived in the rural, we would expect about 61.4% delayed initiation of ANC visit; if however, everyone lived in the rural, and we would expect 71.6% delayed initiation in ANC visit. Model III revealed the predictive margins of all factors examined for delayed initiation for ANC visits, while Model IV presented the predictive marginal effects of the determinants of adequacy of ANC visits. Conclusion: The precise mechanism by which these factors affect ANC visits remain blurred at best. There may be factors on the demand side like the women’s empowerment, financial support of the husband, knowledge of ANC visits in the context of timing, frequency and the expectations of ANC visits might be mediating the effects through the factors found associated in this study. Supply side factors like the quality of ANC services, skilled staff, and geographic location of the health centers also mediate their effects through the highlighted factors. Irrespective of the knowledge about the precise mechanism of action, policy makers could focus on improving women’s empowerment, improving women’s education, reducing wealth inequity and facilitating improved utilization of ANC through modifications on the supply side factors such as geographic location and focus on hard to reach women.
This study used secondary data from the 2011 Ethiopian Demographic and Health Survey (EDHS). We accessed the data from MEASURE DHS database at http://dhsprogram.com/data/available-datasets.cfm. The Ethiopia survey was conducted by the Ethiopian Central Statistical Agency as part of the International Demographic and Health Survey program known as MEASURE DHS, which is currently active in 90 countries and conducted under the auspices of the United States Agency for International Development (USAID) with the technical assistance of ICF International, based in the USA. The Demographic and Health Surveys (DHSs) are free, public datasets, though researchers must register with MEASURE DHS and submit a request before access to DHS data is granted. This data request system ensures that all users understand and agree to basic data usage ethics standards. Sampling procedures were published in the final report [26]. The 2011 EDHS samples were selected using a stratified, two-stage cluster sampling design to provide estimates for the health and demographic variables of interest for the country. The sampling frame consists of a total of 85,057 Enumeration Areas (EAs). A nationally representative sample of 17,817 households was included in data collection. The outcome variables of this study were- 1) Timing of first ANC attendance, and 2) Total number of ANC attendance. ANC visits are of critical important to avert pregnancy related complications, counselling for maternal and foetal health, preparedness for health-facility delivery [27]. WHO recommends the first ANC visit should take place within the first trimester of gestation, and at least four visits during the course of the pregnancy. According to these guidelines, the outcome variables are categorized as: 1) Timing of first ANC attendance (Within 3 months of gestation = early, and beyond 3 month = delayed), and 2) Total number of ANC attendance (<4 visits and 4 or more visits). Besides these, several individual and community level factors were considered as explanatory variables for their relevance in the uptake of ANC care. These were: Age (15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49), Type of place of residence (Urban, Rural), Highest educational level (Nil, Primary, Secondary, Higher), Wealth index (Poorest, Poorer, Middle, Richer, Richest), Occupation (Not working, Sales, Agricultural, Other Skilled manual), Frequency of reading newspaper or magazine (Not at all, Less than once a week, At least once a week), Frequency of listening to radio (Not at all, Less than once a week, At least once a week), Frequency of watching television (Not at all, Less than once a week, At least once a week), Sex of household head (Male, Female), Decision maker of respondent's health care (Respondent alone, Respondent and husband/partner together, Husband/partner alone). Data analyses were carried out using STATA 14. The dataset was checked for cases which fulfilled all the inclusion criteria: age being 15 years and above, having experienced at least one childbirth, availability of information on ANC visits. Basic characteristics of the participants were tabled using frequencies and percentages. Chi-square test was performed to examine the association between timing and frequency of ANC visits and the explanatory variables. Furthermore, multivariable logistic regression analysis was used to determine the odds ratios (with corresponding 95%CI) of delayed initiation of ANC visits and less than four ANC visits. Examining marginal effects, we explored the disparities in predicted probabilities across the factors, in which estimated effects were proportionately adjusted according to a weight for each level of the covariates. Based on the estimation of marginal effects, we predicted the probability of delayed initiation and inadequacy of ANC visits [28]. Thus; Where Set[E = e] reflects putting all observations to a single exposure level e, and Z = z refers to a given set of observed values for the covariate vector Z. Furthermore, p^ez is the predicted probabilities of delayed initiation of ANC visits and adequacy of ANC visits respectively for any E = e and Z = z. The marginal effects indicate a weighted average over the distribution of the covariates or confounders and are equal to estimates got by standardizing to the entire population. As a post logistic regression test, the exposure E is set to the level e for all respondents in the dataset, and the logistic regression coefficients are used to compute predicted probabilities for every respondent at their observed covariate pattern and newly exposure value. Since predicted probabilities are computed under the same distribution of Z, there is no covariate of the corresponding effect measure estimates. After obtaining results of the logistic regression model; the margins command was then used to compute the marginal effects of the factors in STATA [28]. Before each interview, all participants gave informed consent to take part in the survey. The DHS Program maintains strict standards for ensuring data anonymity and protecting the privacy of all participants. ICF International ensures that the survey complies with the U.S. Department of Health and Human Services regulations for the protection of human subjects, whilst the host country ensures that the survey complies with local laws and norms. Further approval for this study was not required since the data is secondary and is available in the public domain. More details regarding DHS data and ethical standards are available at: http://goo.gl/ny8T6X.
N/A