While many nations have made significant progress in reducing maternal death, improvements have been uneven both among countries and populations. Worldwide, more than a thousand women still die each day from pregnancy-related causes [1]. More than 99% of these deaths occur in the developing world; 87% in sub-Saharan Africa and South Asia-most in just 11 countries. Not surprisingly, given its huge population, India leads the world in total deaths with 63,000 mothers dying each year [2].

While annual maternal mortality in India appears staggering at first glance, the world's second most populous country has made significant progress toward attainment of MDG Five [3]. With a 4.9% rate of decline in the maternal mortality ratio (MMR) between 1990 and 2008, it is one of the few countries achieving reductions close to those required for success in 2015 [4]. Regionally, India's MMR of 230 per 100,000 population ranks better than Nepal's, Bangladesh's, Pakistan's and Myanmar's with MMRs of 380, 340, 260 and 240 respectively [2]. As in other parts of the world however, overall gains obscure differences between populations [5]. Estimates from the Government of India National Family Health Survey (NFHS II, 1998-1999) suggest for instance, that maternal mortality could be as much as 132% higher outside of cities [6]. In spite of these disparities most areas of India have still seen steady if not spectacular progress in reducing maternal deaths [7].

Much of India's current success in rural areas may be attributable to the establishment of the National Rural Health Mission (NRHM). The goal of NRHM was to provide accessible and affordable primary health care to the non-urban poor of India [2]. This was largely accomplished by revitalizing and staffing more than 22,000 Community Health Centres (CHCs), 4,000 Primary Health Centres (PHCs), and some 150,000 Sub-Centres (SCs) to serve India's rural populations [3]. Additionally, the government has hired almost a half million Accredited Social Health Activists (ASHAs) to promote health programs, and extended hours of operation for CHCs and PHCs [4]. While many still complain of glaring deficiencies in NRHM efforts [6], the clear progress being made in reducing maternal mortality in rural areas speaks loudly for the current health strategy.

Attaining further declines in MMR in India however, is likely to become increasingly difficult as more traditional interventions are exhausted. Some progress appears possible through additional investment in healthcare infrastructure for the rural poor. A recent survey found that nearly 150,000 health centers still do not have a doctor according to India's health ministry [8]. While activists continue to complain about unevenness of healthcare resources, the NRHM has made significant progress in staffing CHCs and PHCs although there remains a severe shortage of physicians and nurses trained in emergency obstetric care [9,10]. As more rural health centers achieve adequate staffing levels however, further reductions in MMR can only come from reaching harder-to-access populations, many of them residing in more isolated rural areas.

Beginning in 1935, the Government of India created a schedule of disadvantaged castes and tribes living in remote areas of the country for additional protection and services [11]. Labeled by higher Hindu castes as "untouchables" or "Dalits", these groups were often severely impoverished and faced widespread stigma and discrimination [12]. Scheduled Castes and Scheduled Tribes (SC/ST) had the shortest life expectancy at birth, the lowest rates of female literacy, and the highest infant and maternal mortality rates [11,13]. Even today, these SC/ST are among the most economically deprived and marginalized groups in India [14]. Most are rural, scattered, difficult to reach [15,16]. While they make up only 24% of India's total population [8], a 2008 study by UNICEF concluded that they contribute more than half of the country's maternal mortality [17]. Karnataka state, where the study was carried out, is one of the top ten states in India for SC/ST [18].

The survey also found that a vast majority of women (93.4%) had received at least one antenatal checkup for their most recent pregnancy, a slightly higher percentage than the 88.6% reported for rural Karnataka in District Level Household Survey carried out in 2007/2008 (DLHS-3) [20]. In contrast, reported institutional deliveries were slightly lower (55.4%) compared to DLHS-3 (59.7%). These differences may reflect dissimilarities in demographic composition of the samples: DLHS-3 had only 18.1% SC/ST compared with 43.7% in our study.

Data from the survey also suggests that traditional birth attendants (TBA) still play an important role in deliveries in rural Karnataka. One out of every three births in sample villages occurred at home with the help of a traditional birth attendant. Not surprisingly, women who reported delivering at home were also significantly less likely to say they had received antenatal care. SC/ST reported both lower levels of antenatal care and poorer rates of institutional delivery.

Consistent with other studies, low income (as measured by possession of a BPL card) was associated with lower rates of institutional birth [21]. This finding suggests that simply making more maternity services available, particularly if they are for-profit in nature, is unlikely to solve the problem of low institutional delivery rates. Interestingly, our finding that SC/ST had lower odds of institutional delivery even after adjusting for possession of a BPL card, suggests that there might be even deeper social and cultural reasons for low uptake of services among these groups. Saroha et al., in their study of the relationship of caste and use of maternal health services, suggested that because maternal care involves physical contact with health service providers who often belong to general or other castes, lower caste women often elect home deliveries with TBA from their community out of fear of being stigmatized and discriminated against [16]. Such a finding may also suggest the need for targeted health promotion to increase the uptake of maternal health services. There is some evidence to suggest that SC/ST will uptake health interventions if they are culturally appropriate and involve marginalized communities in program planning and implementation [22].

There is little doubt that the Government of India will have to make substantial investments in promoting antenatal care and institutional delivery among marginalized communities such as SC/ST. Social and cultural norms are deep-rooted in rural Indian society. It will take much more than improved healthcare access and affirmative action legislation to change hundreds of years of ongoing discrimination and social isolation based on caste [23]. Despite that, this study underscored the need to address what are clearly inequities in the provision maternal health services among these marginalized segments of Indian society.