These probabilities were estimated from survey-weighted multivariable logistic regression whose primary exposure was number of ANC visits which was treated as a count and adjusted for children ever born, maternal employment status, sex of a child, type of toilet facility, stunting, place of residence and province.