Poor coverage and low uptake of skilled maternity care are major contributors to maternal morbidity and mortality. India alone accounts for 17 % of the 289 000 annual global pregnancy-related deaths [1]. Safe motherhood requires adequate distribution of health services, access to emergency obstetric care, and skilled birth attendance [2].

Individual, household, community, and health system factors affect access to and utilisation of health care. At the individual and household levels, economic status is a key determinant. Analysis of Demographic and Health Survey (DHS) data from 45 developing countries has shown that wealthier women are much more likely than poorer women to have prenatal care and to deliver with a skilled attendant [3]. Other country-level studies support this trend. In Nigeria, women in the highest household wealth quintile are at least seven times more likely to deliver in a health facility than women in the lowest [4]. In Cambodia, the wealthiest women are almost 12 times more likely to do so [5].

Other determinants include maternal age, education, and parity. For example, younger, less educated women from lower socioeconomic groups in Brazil make inadequate use of prenatal care services [6]. In Bangladesh, skilled maternity care among married adolescents is associated with higher education and wealth index, urban residence, and lower birth order [7]. In India, women in northern states and rural areas use maternal health care less than others; barriers include low household economic status, caste, maternal and paternal education, higher birth order, Muslim faith, and less exposure to mass media [8, 9].

Some research suggests that women’s autonomy effects maternal care-seeking [10–12]. In Ethiopia, women who were ultimately responsible for decisions about birthplace were almost four times more likely to deliver at a health facility than those who were not [13]. In Tajikistan, women with financial decision-making power were more likely to attend at least one prenatal consultation (although less likely to attend four or more), deliver with a skilled provider, and seek institutional delivery care. However, associations are contextual; autonomy might be a weak predictor of care uptake in general [14], but more strongly associated with choosing private over public sector care [15].

Studies in diverse settings have shown associations between urban location and institutional delivery [5, 7, 16]. Urban residents benefit from a concentration of health infrastructure and proximity of services. However, population growth creates greater demand for health services. When these services are unevenly distributed, access becomes unequal. These inequalities adversely affect disadvantaged groups in underserved neighbourhoods [17]. Our previous research has shown a positive association between higher socioeconomic status and the use of private prenatal and delivery care by women from informal urban settlements (slums) in Mumbai [18].

India is the world’s fastest urbanising country; currently, 410 million Indians (one-third of the total population) live in urban areas. Mumbai, the country’s second largest city, has more than 16 million inhabitants [19], more than 40 % of whom live in slum areas [20]. The health care sector is characterised by a co-existence of medical systems and public and private providers. Public sector infrastructure includes teaching hospitals, specialist hospitals, general hospitals, maternity hospitals, and community-level health posts and dispensaries [21]. The private sector includes super-speciality hospitals, medium-sized facilities that provide both outpatient and inpatient care, and a substantial number of smaller practices that offer limited services. Most urban health care across socioeconomic groups, including the disadvantaged, is privately provided. The sector is virtually unregulated and many practitioners are underqualified or lack formal training [22, 23].

Because institutional prenatal and delivery services are often underutilised, discussions of maternity care in low- and middle-income countries have emphasised uptake of services, followed by a consideration of quality. While some research has documented the utilisation of public and private sector services [5, 24, 25], choice of specific types of facility within each sector has largely been ignored. Understanding how families in underserved urban communities choose among health providers is essential. Although the public sector is an important source of health care for the urban poor, private practitioners dominate in many low-income communities. Examining health care-seeking behaviours in these communities is key to developing effective strategies that address inequalities, improve access, and help protect the poor against unaffordable health costs [26].

Our objectives were to quantify the pattern, determinants, and choice of maternity care provider at the health facility level in the public and private sectors in Mumbai’s informal urban settlements, and to explore the reasons underlying these choices. We were interested in examining two aspects of choice that have appeared rarely in discussions: private sector maternity care for poor people whose substantial use of it has gone largely unnoticed, and the ways in which they decide which providers they will use. Our broad hypothesis was that the likelihood of institutional prenatal care, delivery, and private health care would all increase as maternal education, duration of residency, and economic status increased.

Our study shows that institutional delivery is the norm in Mumbai’s informal settlements. However, poorer and less educated women, and recent migrants were less likely to receive professional prenatal and delivery care. Tertiary public hospitals were a common source of maternity care across all socioeconomic groups. Private hospitals were popular with wealthier, more educated women.

We identified four conceptual processes central to choosing a health care provider: exploring the options, defining a sphere of access, negotiating autonomy, and protective reasoning. The overall aim was the selection of a suitable or best-option provider. Evidence of quality and positive outcomes encouraged women to seek care with certain providers while others were avoided or abandoned. Heath care decisions and provider choice were mediated by household socio-economic status, the cost of care, and the ability of women to negotiate their social and economic environment.

The dominance of tertiary public hospitals as a preferred site of maternity care across socioeconomic groups is a problem for the equitable delivery of health services to underserved areas: despite being located in proximity to poor neighbourhoods, poor perceptions of quality, limited services and understaffing in primary public health facilities often cause residents to bypass them in favour of tertiary hospitals. To most women, large hospitals symbolised comprehensive, integrated care, sophisticated equipment and technology, expertise and specialisation, where complications could be treated in one place. This made them attractive and convenient. At the same time, this preference exacerbates problems of overcrowding, longer waiting times, shorter consultations in tertiary facilities and loss of wages, dissuading some educated and wealthier people from utilising public sector health care [34].

Poor perceptions or experiences of care and fear of providers and practices were common reasons to avoid certain health facilities, especially in the public sector. Use of public sector service was often considered a consequence of “helplessness” or when “in trouble”. Several studies affirm the urban preference for the private sector [35–37]. Among the reasons for this are ease of accessibility, convenient timings, and a perception that the quality of care is better than in the public sector [38–40]. However, access to private health facilities is limited by the ability to pay; some women who had particularly poor perceptions or experiences of public sector care had either sought financial support from within the family or had taken a loan to avoid seeking care at a public hospital. Muslim women were more likely to seek prenatal and delivery care at private hospitals, reflecting a strong preference for female physicians [15, 41].

Uptake of institutional care was lower among recent migrants to Mumbai. Of women who had arrived within the last year, 24 % made fewer than 3 prenatal visits and 39 % delivered at home. A study by Stephenson and Matthews [42] found that rural–urban migrant women in Mumbai reported levels of prenatal care similar to urban non-migrants but substantially lower delivery care, suggesting that migrants assimilated the urban preference for institutional prenatal care while preserving the traditional practice of home birth. One explanation was that while social networks provided women with information to access prenatal care, they were also a resource for home-based delivery care [42]. In a study of two migrant groups in a Delhi slum, institutional maternity care became habitual when modern health services were available and considered effective. Lower exposure to health care in the place of origin and unfamiliarity with hospital care resulted in greater fear and distrust of institutional delivery. Conversely, greater autonomy and social interaction outside the home increased women’s knowledge of health services and confidence to use them [12]. In our study, recent migrants had limited knowledge of health facilities and quality of services, and weak social networks. This reduces access to information about available or appropriate care and made it difficult to mobilise support to choose from a wider pool of providers. Women often seek maternity care from specific, local private providers recommended by family and reported to offer good quality care.

Our study contributes to an understanding of disparities in the utilisation of institutional care in poor urban areas by considering the complexity of factors that influence uptake and choice of provider across public and private sectors. Its strengths were a relatively large sample and disaggregated data on utilisation patterns in both public and private sectors. Limitations included potential recall bias and ‘best behaviour’ bias regarding women’s use of prenatal and delivery care. We have no reason to suspect that women gave false information, and the reported proportions of institutional care were similar to those in Mumbai slums as a whole [43]. Since we excluded families that were absent after the third visit, we might have missed some women who gave birth in their natal homes. A qualitative limitation arose from the use of quantitative and qualitative methods in grounded theory: we found it difficult to reconcile analytical concepts derived from deductive (quantitative) and inductive (qualitative) methods. We are continuing to develop our analysis into a substantive theory of provider selection.

Socioeconomic differentials manifest as inequities in the availability, affordability, and utilisation of health services [17, 44, 45]. The poorest are less able to pay for care because of disproportional health care costs from greater spending proportional to income, most of which have to be covered by wage income rather than savings [46]. Poorer groups, for whom good health and wellbeing are crucial for economic and household stability, often turn to more accessible, lower quality providers for their health care needs. They tend to consult with less competent practitioners who make less effort [47] or who operate in the largely-unregulated private sector. This is of concern because of the potential iatrogenic effects of over-medication, inappropriate treatment, or ignoring minimum standards of care [38, 48].

In Mumbai’s informal settlements, institutional maternity care is the norm. Individuals and families, even in the most disadvantaged groups, choose among health providers in both private and public sectors. However, socio-economic inequalities limit people’s sphere of access and lead to differential utilisation across groups. Paradoxically, these inequalities make the selection of a suitable provider both more important and more difficult: more accessible practitioners are less likely to be fully qualified or trained, have lower competence and offer poorer quality care. Mitigating uncertainties about quality and safety compels many families to engage in a complex decision-making process, mediated by their ability to mobilise social and economic resources, in an attempt to ensure positive experiences and outcomes of care.

Addressing health care disparities in underserved communities requires a clear understanding of how families choose among health care options. In addition to questions of service uptake, research in pluralistic urban settings must disaggregate information by level of health facility and type of provider across sectors. Improving women’s choice and experiences of health care requires that health sector managers implement effective health system strategies, including high quality maternity services across sectors, a functioning regulatory mechanism, and monitoring of provider competences and behaviour.