Statistical analysis of number of antenatal care visits among pregnant women in rural Ethiopia

Abstract

The ANC service is used to ensure a normal pregnancy with delivery of a healthy baby from a healthy mother. Even if WHO recommends a minimum of four ANC visits, existing evidence from developing countries including Ethiopia indicates that few women utilize it due to different determinants such as lack of education, awareness, nearby health post, residence, etc. Objective: The main objective of the study was statistically to analyze the determinants of the barriers in number of antenatal care service visits among pregnant women in rural Ethiopia. Methods: A cross sectional data from EDHS-2011 was used and 1127 pregnant women who had 9 months of pregnancy was included to the study. Several count models were fitted to select the model which best fits the data, these are: Poisson, NB, ZIP, ZINB, HP, and HNB regression models. Each of these models was compared by likelihood ratio test (LR), Voung test and the information criteria’s. The data were Analyzed using SAS version 9.2. Results: In this study there were excess zeros, 51.5%; the variance of the data, 7.196, was much higher than its mean, 1.85. Women from better progressed regions were about one (OR=0.8048) times more likely to have positive ANC visits than women from other regions, educated women were one times (OR=1.44) more likely to have positive ANC visits than non educated ones, women who had seen danger signs of pregnancy was one times (OR=1.42) the rate of positive ANC visits than women who hadn’t seen it, women with heavy workload had less likely to visit positive ANC attendance (OR =0.36) than women who had no workload problems, there is a greater likelihood of a positive number of ANC visits for rich women than poor women (OR=1.15), and women who had nearby health post have a greater likelihood of a positive ANC visits than women with lack of nearby health post (OR=0.48) holding all other predictors constant. Conclusion: Though the government effort is to improve access to modern ANC visits during pregnancy, it was low in rural Ethiopia than the national value. Lack of awareness, absence of education, heavy workload, poverty, and shortage of health post were significantly associated with not attending ANC visits. Hence, institutions that act on maternal and children’s health care should do well to apply the minimum of four ANC visits scheduled by WHO mainly on rural areas so that all perpetrators of maternal care shall be brought to book to deter others from repeating such absences and thus move the country closer to MDG targets for maternal health by 2015. Hurdle Poisson regression model was found to be better fitted with data which is characterized by excess zeros and high variability in the non-zero outcomes. Key words: ANC; Poisson regression model; Negative binomial (NB) regression model; Zero-inflated Poisson (ZIP) regression model; Zero-inflated negative binomial (ZINB)regression model; Hurdle Poisson regression model; Hurdle NB regression model.