Pakistan, with a maternal mortality rate (MMR) of 297 deaths /100,000 live births, is one of six countries contributing to half of all maternal deaths worldwide.1 In an effort to decrease these unacceptably high rates the Government of Pakistan created a new cadre of skilled birth attendants, the community-based midwives (CMW).2 Critical to consider in maternal health service programming in Pakistan is the disparity in access to and uptake of crucial maternal health services based on socio-economic status. According to the national DHS (2007) survey, 70% of women in the wealthiest quintile reported skilled birth attendance compared with 16% in the poorest quintile. One goal of the CMW program is to reduce the inequities in poor and marginalized women's access to maternal health services. Recent research from Pakistan and elsewhere in South Asia shows that the large disparities in access to care between the rich and poor are not solely the result of economic poverty.3–6 Economic poverty is relational and embedded within power hierarchies influenced by both class and gender.5,6 Social exclusion is therefore a useful conceptual framework for understanding the multi-dimensional nature of the disparities in women’s ability or willingness to access maternal health services. Understood as both a process and an outcome, a social exclusion lens urges a critical analysis of the role of various actors and processes that directly or indirectly maintain exclusion.7,8 The present study aims to explore whether the CMWs are achieving the government objective of improving access to the full scope of skilled maternity care for poor, socially excluded women in Pakistan. Specifically, the study aimed to: (1) Assess the coverage of CMW maternity care. What proportion of antenatal and birth attendance is provided by CMWs? And are they providing services to the socially excluded? (2) Explore any challenges (programmatic, social, and financial) the CMWs are facing in establishing their practices and providing services to the socially excluded? What factors promote the establishment of their practices? (3) Map the social, financial and other barriers poor and socially excluded women face in accessing CMW services. Findings from this research will identify key challenges the nascent Pakistani CMW program may be facing in providing care to poor, socially excluded groups and will help to highlight potential leverage points at which solutions can be implemented to enable the CMWs to provide optimal and equitable maternity care. Methods A mix of quantitative and qualitative research methods were used. Data were collected in three overlapping modules over a 9-month period in 2011- 2012 in two districts, Jhelum and Layyah in Punjab. Module 1 in each district consisted of qualitative data collection through in-depth interviews with 38 community-midwives, 15 local dais (traditional birth attendant), 30 other health care providers in both the public and private sector, and a variety of program managers and policy makers (20) to understand the challenges individual CMWs are facing as well as the institutional challenges they may be facing in establishing their practices. Five CMW monthly reporting meetings were attended and over 20 hours of observation were carried out to document CMW training in labour room and obstetric wards in Layyah. Module 2 aimed to identify the social, financial, geographic and other barriers socially excluded women face in accessing CMW services. Interviews were conducted with 78 women of reproductive age (15-49) who had given birth in the last two years, 35 husbands of women who had given birth in the last two years, and 23 older women (aged 50 years plus) . Eleven interviews were conducted with women and their families who had experienced either childbirth complications or a maternal death. Finally 18 focus group discussions were held with men and women of varying socioeconomic statuses. Furthermore, informal interactions took place with a variety of community members (170). Module 3 aimed to quantitatively measure levels social exclusion through measures of material assets, poverty of opportunity, and caste and to investigate the associations between social exclusion status and uptake of CMW care. A cross-sectional, clustered and stratified survey was conducted in the two districts (n(total)=1457) with women who had given birth in the past two years. Results Overall, our findings suggest that poor, socially excluded women are not receiving any biomedical maternity care; they are 7 times more likely to report attendance a dai and 80% less likely by a physician compared to the richest, socially included women. They are 4.5 times more likely to deliver at home compared to the socially included. Our data also show that the CMWs have yet to emerge as a significant and relevant maternity care provider in rural Punjab. Only 3% and 11.7% of all births in Layaah and Jhelum, respectively, in the last two years were attended by a CMW. Amongst the small number of women who received CMW care, our data suggest that CMWs are providing services equally to socially included and excluded women. The qualitative data provided a nuanced understanding of reasons why the CMWs are performing sub optimally and why they are providing care equally to socially excluded and included women. It identified the barriers the CMWs face in establishing their practices and the barriers socially excluded women face in seeking their care. A key reason for CMWs not working was a lack of interest in practicing midwifery. Programmatic barriers such as poor quality training, uniformity of polices that result in neglect of the context-specific ground realities, and a failure to incorporate the gendered and social realities of CMWs’ lives into the design of the program further hindered the interested CMWs in establishing their practices. The few CMWs who are struggling to practice are largely providing care to their relatives and poor, socially excluded women. These two groups are not their stated target clients but are being served by default as the CMWs try to gain experience and exposure in order to reach their real target, the paying patients. Our research also provided examples of CMW characteristics that predict success in CMW functioning. These include a poverty-pushed desire to work, greater family support to overcome the gendered and social barriers, and individual CMWs’ professionalism and work ethic. The successful CMWs had developed linkages with other providers. Characteristics such as CMW age and marital status did not emerge as important predictors of success. Socially excluded women also face barriers in accessing CMW services. The data showed that they were unable to pay CMW fees and that, overall, a culture of disrespectful maternity care directed towards poor, socially excluded women deters them from seeking not only services from CMWs but from all biomedical services. Discussion The current study has identified several challenges the CMWs are facing in establishing their practices and, in turn, are limiting the efficacy of the broader Maternal, Neonatal, and child health (MNCH) program. However, a few CMWs are successful and their characteristics and initiatives provide potential leverage points the program can use to improve its functioning and achieve its stated goals. Based on these leverage points, we recommend the following: Challenges Policy recommendation 1. Large number of CMWs non-practicing; are not interested in practicing midwifery Improve CMW selection by (1) changing criteria such as increasing minimum age, focusing on personal characteristics of professionalism and work ethic, and giving preference to those with previous experience; and (2) changing recruitment process to include an interview and developing a system to verify CMW stated credentials 2. CMWs must adhere to the same gendered mobility restrictions that necessitated their appointment in the first place Consider establishing a system of support for escorting CMWs to deliveries and take into account physical and social geography by decreasing density of CMWs in more densely populated areas while increasing density of CMWs in less densely populated ones. 3. CMWs lack competency in providing domiciliary care (1) Incorporate a domiciliary care training component into the training via an internship program involving local domiciliary maternity service providers; and (2) develop portable protocols of care for CMWs to refer to for either routine or emergency domiciliary care Challenges Policy recommendation 4. CMWs are a part of a health system culture that provides disrespectful maternity care (1) Health care policy makers at the highest levels should consider highlighting respectful maternity care as an important dimension of quality; and (2) incorporate the notion of respectful maternity care into the CMW training program (3) Consider incorporating into the curriculum a module regarding the social determinants of health and encourage thinking around the targeting of poor and marginalised populations; 5. CMWs serving the poor means they are not able to financially sustain practice (1)Explore methods of subsidization to provide incentives to CMWs to target poor and vulnerable populations.