Résumé
For assuring safe maternal and newborn health, institutional delivery was given paramount importance. In India, in spite of several efforts, lesser than 40% deliveries are conducted at health facilities, mostly at private sector. The present cross-sectional study aimed to fi nd out the determinants of preference for delivery at government hospitals in rural areas of Lucknow, a district in Uttar Pradesh. Multistage random sampling was used for selecting villages. From them, 352 recently delivered women were selected, following systematic random sampling. Overall, 84.9% of deliveries were conducted at health institutions. Out of them, 79.3% were at government hospitals. Applying multivariate logistic regression, Hindu women (odd’s ratio [OR] = 3.205), women belonging to lower socio-economic class (OR = 4.630) and late registered women (OR = 2.320) were found to be more likely to deliver at government hospitals. Attention should be given to religion, social status and timing of registration for ensuring higher fraction of deliveries at government set-up.