Poor, rural women in India contribute disproportionately to the nation's high maternal mortality ratio. In response to this problem, the Indian government launched a conditional cash transfer scheme, "Janani Suraksha Yojana (JSY)," in 2005 to increase poor women's access to maternal health care. The state of Tamil Nadu reorganized public health system resources and the state of Gujarat contracted with private providers to implement the scheme in rural areas. This study investigated the role of JSY/government assistance, and other health care sector and household factors in predicting poor, rural women's utilization of maternal health services in the two states. Health care sector factors included receipt of JSY payment, availability of a primary health center with round-the-clock services, and connection to a health facility by an all-weather road. Household factors included maternal education, paternal education, age at first birth and parity. Use of four maternal health care services was examined: adequate antenatal care, institutional delivery, private facility delivery and Cesarean section. State findings were compared and contextualized by examining health polices/practices and health infrastructure within each state.

          The study employed secondary data analyses using District Level Household Survey (DLHS)-3 data, with a sample of 2,267 rural women from the lowest two wealth quintiles.  Multivariate logistic regression analyses examined associations between identified factors and maternal health care utilization in the two target states.  Overall, Tamil women had better access to maternal health care services than Gujarati women.  JSY payment predicted use of private facility deliveries in Gujarat, which incentivized use of private providers, but not institutional deliveries in Tamil Nadu where women also received cash regardless of the place of delivery.  JSY payment did not predict adequate antenatal care, which was not incentivized.  Access to an all-weather road predicted institutional deliveries in both states and adequate antenatal care by Tamil women.  Maternal education was a significant predictor of maternal health services utilization in Tamil Nadu, while paternal education predicted such usage in Gujarat.  Parity significantly predicted poor, rural women's use of all services.  Implications of the findings for strengthening conditional cash transfer schemes and improving maternal health care services are discussed.