Utilization of Maternal Health Services in Uttar Pradesh: A Comparison between Logistic and Multilevel Regression Analysis
Divya Tiwari, Jawaharlal Nehru University
India accounts for about one-quarter of maternal deaths worldwide with the highest number of deaths occurring in the state of Uttar Pradesh. Utilization of antenatal, natal and postnatal care is important for reducing maternal morbidity and mortality. Uttar Pradesh, however, is characterized by low uptakes of maternal health services. Using RCH-DLHS-2 (2002-2004) dataset, present study attempts to investigate the importance of community and individual level variables in explaining factors affecting the utilization of antenatal, delivery and postnatal services in Uttar Pradesh by using techniques of simple logistic regression analysis and multilevel regression analysis. Multilevel analysis has number of advantages over logistic regression analysis. From the analysis it is evident that the standard error of the coefficients under both the techniques present different picture. When data has hierarchical structure more and more community level variables emerge as significant associates of antenatal, natal and postnatal services.