Modeling Determinants of Desired Number of Antenatal Care Visits among Pregnant Women in Ethiopia: Application of Multilevel Count Regression Models.

Abstract

Antenatal care visit is the service given to pregnant women in order to have a safe pregnancy and a healthy baby. Recently the technical working group of World Health Organization has recommended a minimum level of care to be eight visits throughout the pregnancy to reduce the maternal morbidity and mortality. The main objectives of this study was assessing the regional variation of number of ANC service visits per woman and identifying the factors influencing number of antenatal care visits based on 2016 EDHS dataset. The survey collected information from a total of 15,683 women aged 15-49 years out of which 7174 women were considered in this study. Multilevel count regression models were used to explore the major risk factors and regional differentials in number of antenatal care service visits per a child bearing woman in Ethiopia. Descriptive statistics results show that nationally about 2481(35%) of mothers did not take any antenatal care, which indicates excess zero and less percentage of non-zero counts and only 255(3.6%) attended ANC service follow-up eight times and above. From several multilevel count regression models (Poisson, NB, ZIP, ZINB, HP, HNB), multilevel hurdle Poisson model was selected using model comparison criteria like AIC, BIC and Deviance. Among multilevel hurdle regression models (null model, random intercept with fixed coefficient and random coefficient model), it was found that the random intercept with fixed coefficient model is the best model to describe the data set. At the stage of multilevel, HP model showed that predictor variables age, type of place of residence, wealth index, Mother educational level, husband educational level, frequency of watching television, distance from health facility, wantedness of pregnancy and pregnancy complication were found to be related with the antenatal care service follow-up. The multilevel analysis further showed that there are within and between regional variations per mother regarding to ANC service visits in Ethiopia. The findings of this study might help in planning and developing strategies for utilization of ANC visits among pregnant women in Ethiopia.