Background: High quality evidence on the cost-effectiveness of planned birth in alternative settings (at home, in a midwifery unit or an obstetric unit) has been lacking, and is a priority area for maternity policy. Aim: To provide evidence about the efficiency of the configuration of maternity care in England and to estimate the cost-effectiveness of alternative settings for intrapartum care for ‘low risk’ women, thereby providing guidance for commissioners, clinicians and for pregnant women and their families. Methods: A literature review of existing evidence was followed by four stand-alone empirical studies using different methods to determine the efficiency and cost-effectiveness of alternative settings for intrapartum care. Data from the Birthplace in England Programme of Research were analysed to explore whether there are differences in the efficiency of maternity units when they are stratified according to the type and scale of unit. Incremental cost-effectiveness ratios were used to estimate the short-term cost-effectiveness of different planned settings for birth for ‘low risk’ women and to develop a template for the design of decision-analytic models to estimate life-long cost-effectiveness for the mother and baby dyad. Findings: The larger obstetric units (OUs) tended to be more efficient than the smaller OUs. Less than half of free-standing midwifery units (FMUs) were operating at full efficiency. The cost of intrapartum and after birth care, and associated related complications, was less for births planned at home, in a free standing midwifery unit (FMU), or in an alongside midwifery unit (AMU) compared with planned births in an obstetric unit (OU). Planned birth in a FMU or in an AMU compared with an OU will generate incremental cost savings but with uncertainty surrounding the outcomes for the baby. Planned birth in all non-OU settings generated incremental cost savings and improved outcomes for mothers. For ‘low risk’ women having a second or subsequent birth, planned birth at home was found to be the most cost-effective option.

Background: There were approximately 63,097 known asylum seekers in England in 2002 (Heath et aL 2003). Women asylum seekers may be more seriously affected by displacement than men, leading to increased isolation, poverty, hostility and racism (Burnett and Peel, 2001a). In England, Black African including asylum seekers and newly arrived refugees had a seven times greater chance of maternal modality than White women (CEMACH, 2004 p244). Furthermore, women from ethnic groups other than white are twice as likely to die as women in the white group. In the CEMACH report a large number of women who died spoke little English. Access to local information about asylum seekers and refugees was difficult to obtain and suggested a disorganised service provision for this group of women. Aim: It was the intention of this study to explore and synthesise the experience of midwifery care by women asylum seekers and refugees in one large maternity unit in England. Design: Longitudinal exploratory case study research utilising a series of interviews. Sample: Four women from: Afghanistan, Rwanda, Somalia and Zaire. Three women were asylum seekers and one was a refugee. One woman spoke fluent English. Setting: Liverpool Women's Hospital and the women's homes. Years: The study took place from December 2002 - July 2003. Methodology: The research was developed from a constructionist paradigm which identifies that multiple realities can exist for individuals who experience 10 a similar phenomenon. Truth within this perspective is constructed by the individual. The researcher (LB) was the main instrument of data collection. Interpretation was generated via the researcher and was verified by the women at the final interview. The underpinning foundation for this study emerged as symbolic interaction theory (Mead, cited in Morris, 1967 p43; Blumer, 1969). Methods: Following ethical approval consent was obtained with the help of professional interpreters. Exploration was facilitated by in-depth interviews at five time points throughout the antenatal and postnatal period. Photographs taken by the women themselves were used as a prompt for conversation. Analysis: The researcher's interpretation of the data identified emerging themes and categories. The process of analysis involved decontextualisation, display, data complication and re-conceptualisation (Miles and Huberman, 1994 p10). Three key themes were generated: the influence of social policy, understanding in practice and the perception of 'self. Results: Synthesis of the results suggested that social policy directly affected the lives of the women. At times, "taken for granted" communication created a barrier to understanding for the women. Stereotype was socially constructed and pervaded the care environment. The women perceived 'self as a response to social interaction. The midwife-woman relationship relied heavily on gestures and symbols and the women's descriptions are related to symbolic interaction theory. Midwives capable of understanding the subtle cues in communication may be able to negotiate negative stereotypical images generated by society. Women had little or no information around childbirth. Recommendations: Midwifery care would benefit from a deeper understanding of how the women in this study perceived 'self. An advisory post may provide the link between maternity care and broader public health issues. Midwives should engage with asylum seekers and refugees to develop partnerships in care. A collaborative partnership may assist in creating relevant information around childbirth for other asylum seekers and refugees. Innovative methods of dissemination of information related to childbirth should be facilitated by a collaborative approach with non-governmental and community organisations.

The purpose of this embodied phenomenological inquiry was to explore the impact of the so-called Capacitar Body-Mind-Spirit Practices Training for the transformation of individual and community trauma. A general philosophical framework of transpersonal psychology, particularly drawing on the concept of interrelatedness, has been implemented. This research took place in the U.S.-Mexico borderlands with 14 female co-researchers (Mexican, Mexican-American, Mayan, White American) who live in the twin cities of El Paso (Texas, U.S.) and Ciudad Juárez (Chihuahua, Mexico), who were thus exposed to the constant threat of lethal violence and so-called femicide, to discrimination and the socioeconomic problems that are peculiar to this region. The women’s embodied experiences of change as a result of the Capacitar Training were investigated with semi-structured multiple interviewing multilingually (English / Spanish / Mayan). Data analysis procedures combined imaginative variation and embodied interpretation that resulted in a general meaning structure with its variations. The contribution to knowledge made by the phenomenological results consists of the confirmation of past research on the impact of the Capacitar Training that showed the potential of the body-mind-spirit practices for transforming trauma with culturally and spiritually diverse individuals. The most significant research findings of the present study suggest that: (1) the majority of the co-researchers’ experiences of bodily change through body-mind-spirit practices initiated further integration of past negative (traumatic) and / or positive experiences in an embodied way, including interrelatedness to spirituality, culture and nature; (2) the initial bodily felt shift led to the co-researchers’ desire for more change; (3) the experiences of change were independent of the cultural or spiritual background of the co-researchers; (4) a desire to support others’ change emerged for the co-researchers based on their own experiences of improvement; and (5) ambiguity arose for a minority of co-researchers in the beginning of the training related to cultural and religious barriers, and self esteem issues; and at the end of the Capacitar Training linked with the question of commitment to time and to the engagement with the practices. To enhance an embodied understanding of the phenomenological results, biographical information from the co-researchers’ life experiences related to the borderlands had been compiled during the interviews, which mirrored the body of knowledge on issues in the U.S.-Mexico borderlands and has been presented as a cultural-spiritual narrative composite. Furthermore, evaluative information about the co-researchers’ experiences of the conveyance of the Capacitar Training has been synthesized, adding their critical reflections about the conduction and the further development of the Capacitar approach.

This thesis is composed of four essays that make empirical contributions to impact evaluations of health sector interventions in low- and middle-income countries, in light of the interventions introduced under Egypt's Health Sector Reform Program (HSRP) between 2000 and 2014. We are mainly interested in the effects on family planning and maternal and child health. Different methods are used in this context: difference-in-differences (DD), DD propensity score matching (PSM), fixed effects (FE), random effects (RE) and pooled ordinary least-squares (POLS). In chapter 3, we estimate the effect of improving the quality of health care through facility accreditation on the family planning, maternal health and child health outcomes that we expect to reflect the effect of compliance with quality standards, policies and procedures. We found that accreditation had multiple positive effects, especially on delivery care and child morbidity prevalence. No significant effects were observed, however, with respect to most antenatal care (ANC) outcomes. In chapter 4, we estimate the medium-term effect of introducing user fees on the utilization of family planning, ANC and delivery care services, women's access to health care, and child health status. With respect to ANC, we found that the positive effect of increased willingness to pay for an improved quality of service outweighed the negative effect of the price elasticity of demand. Introducing user fees was associated with a higher likelihood of receiving ANC by skilled health personnel, a higher likelihood of receiving at least four ANC visits and a higher likelihood of receiving iron supplements during pregnancy. However, the two effects offset each other with respect to the outcomes that reflect the utilization of family planning and delivery care services, women's access to health care, and child health status. No net effect at all was observed on these outcomes. Chapter 5 complements the analysis of chapter 4 by allowing us to estimate the net effect of combining user fees and two quality improvement interventions: facility accreditation and performance-based financing (PBF). Again, we observe positive effects on both the utilization and the quality of ANC services. More notably, a positive effect on access to care was observed during our first study period that is more likely to reflect the effect of quality improvements. These effects, however, were reversed during the second study period that is more likely to reflect the effect of user fee introduction. The positive effects reported in chapters 4 and 5 were mainly with respect to ANC. No effects were reported on the outcomes that reflect the utilization of family planning and delivery care services, and child health status. In chapter 6, we estimate the effect of discontinuing provider incentives on health outcomes that reflect the health services targeted by the PBF scheme as well as the quality of these services. We found that discontinuing the incentives had a negative effect on four out of seven health outcomes: knowledge of contraceptive methods, receiving ANC by skilled health personnel, receiving iron supplements during pregnancy, and more importantly, under-five child mortality. Our findings, first, suggest that improving the quality of care through facility accreditation could be particularly effective in improving delivery care and child health. However, a high level of commitment from the central government is indispensable to sustain the positive effects of quality improvement interventions. Second, introducing user fees will not necessarily have negative effects on access and utilization of family planning, maternal health and child health services. However, user fees are ineffective, in general, as a stand-alone policy. Third, negative effects of introducing user fees in low- and middle-income settings on the utilization of healthcare services can be mitigated by officially exempting the poor from any fees at the point of service. More importantly, this exemption should be known to the population. Fourth, combining quality improvement interventions with user fees will not necessarily add to the few positive effects obtained when user fees are introduced as a stand-alone policy. Finally, provider incentives should be introduced carefully in low- and middle-income countries as negative effects are observed when these incentives are discontinued.

The UK has one of the highest young pregnancy rates in the developed world (Unicef, 2001). An association between socio-economic environment and young pregnancy is evident, with deprived areas having higher under-18 conception rates and lower proportions of abortions than affluent areas (e.g. Uren, Sheers & Dattani; 2007). Presently, the processes underlying this association are unclear (e.g. Lee, Clements, Ingham, & Stone, 2004). The forthcoming studies therefore have set out to explore the mechanisms and processes through which the socio-economic environment influences young people's sexual and reproductive decisions and behaviours. A theoretical framework drawn from critical psychology, social psychology and epidemiology was used to provide a detailed exploration of this association. The role of cultural and social factors as possible processes by which socio-economic environment influences young people's sexual, reproductive and pregnancy decisions were explored using social representations theory framework (e.g. Moscovici, 2000; 1976) and cultural and behavioural explanations of social inequalities (such as suggested by Townsend & Davidson, 1983). Young mothers (N=16) and fathers (N=5) from a mix of socioeconomic environments were interviewed about their experiences. These data were used to design future studies. The first of these, addressed social acceptance of young pregnancy in areas characterised by different levels of deprivation (N=570), while the second used questionnaires (N=49) and focus groups (N= 10) to explore the antenatal ,. and postnatal needs of young parents as well as the reasons for their low attendance. These studies suggest that the image of pregnancy and parenthood offered to young people through the process of intra-cultural communication differ from their actual experiences - this discrepancy between the imaged and real picture for young parents' impacts upon their emotional well-being and future plans. Social inequalities influence young pregnancy outcomes through social representations. Within each economic subgroup (more deprived and more affluent areas and families) different social representations produce varying values and beliefs concerning sexual and reproductive behaviour, gender roles, role models and social acceptance of young pregnancy and abortion. Several issues warrant further exploration - these include socio-economic variations in contraceptive use and reasons for non-use, the influence of gender roles, the need for ward-level comparisons and the influence of culture and subcultures on sexual, reproductive and pregnancy decisions.

Background: Maternal mortality and morbidity still pose a significant challenge in Sudan, where no significant improvements in maternal health have been achieved despite the focus on the Millennium Development Goals. Under-utilisation is a major public health concern even though Sudan is among the African countries that have registered poor maternal and child health. Health services in Sudan are generally limited and with poor quality and disparate access. Therefore, there is a need for better understanding of the barriers to the provision and utilisation of maternal health services in order to improve the health and survival of Sudanese mothers. Objectives: This study sought to assess the maternal health system functions and influences on utilisation as well as the social, cultural, and women’s characteristics that may constitute barriers to utilisation. Methodology: The study used an explanatory sequential mixed-methods design. A comprehensive analysis was conducted using several quantitative and qualitative data sets, guided by a new framework, the Maternal Health System Performance framework (MHSP) developed as part of this work in order to assess both the three objectives and four functions of the health system on both macro and micro levels. Findings: The study findings provide clear evidence that the Sudan health system is not currently capable of achieving an adequate level of attainment of the health goals or equitable distribution, due to dysfunction of the four health system functions. In addition, the findings draw attention to the important role of the stewardship function in health system performance. This function can play a key role in health system reform, as it influences management of the health system and should work across all elements of the system to ensure a well-functioning health system and efficient use of resources. The findings also underline the important role of health system related factors rather than simply population factors (such as individual, household, and community factors) in the low service utilisation among women in poor settings. While it shows that certain population characteristics such as household income and education do have a significant impact on the utilisation, the health system functions, and in particular the stewardship function, are also demonstrated to be of considerable importance. Implication: These findings have implications for policy and practice, indicating that simply blaming women for not using maternal health services is unhelpful and inappropriate and indicate that decision makers should focus more fully on improving the performance of the health system. According to the comprehensive assessment of the health system performance, the study proposes several recommendations for each health system function to enhance the performance in the context of limited resources, ultimately to improve women’s and community health in Sudan.

Optimizing maternal health and improving reproductive outcomes are widely acknowledged as major challenges in the health care system. Care during the antenatal period has been the focus of improving maternal health and reproductive outcomes. Yet, evidences have shown that antenatal care alone is not enough. Initiating care before conception or preconception care could be potentially effective to further improve maternal health and reproductive outcomes. Preconception care encompasses a range of health promotion, risk assessment, preventative and curative interventions for women of reproductive age to reduce risks that potentially affect reproductive outcomes. It aims to provide prospective parents information and support with regards to preconception interventions that are beneficial for the parents and future children. Primary care providers are often being urged to provide preconception care as part of primary care services. In support of preconception interventions, there has been increasing evidences for such interventions. However, existing reviews or studies of preconception interventions have been limited by being risk specific, for example; focussing on folate supplementation or women with diabetes. Adding to this, interventions were reported mainly carried out in the secondary care settings. There is still paucity of evidence that comprehensively evaluate the impact of providing preconception care as a systematic approach involving multifactorial risk factors and, in particular, in primary care. Preconception care involved a range of risk assessment; assessment of genetic risk is no exception. The aim of preconception care for genetic risks is to allow women or prospective parents the opportunity to have informed reproductive decisions of future pregnancies. However, experience of offering preconception care in addressing genetic risks is yet less explored. This thesis specifically sought to evaluate the potential impact of preconception care involving assessment of reproductive genetic risk. Further, this thesis also aims to provide evidence for effectiveness of preconception interventions on multifactorial risk factors in the primary care settings. As primary care providers especially GPs are increasingly being recognised to provide such care, it was thus important to explore their views. For this, this thesis aimed to explore the opinions and attitudes of GPs in the United Kingdom towards providing preconception care that involved assessment of reproductive genetic risk in current general practice. This study took place within the Primary Care Trusts of Nottinghamshire and Derbyshire. The findings from this thesis are expected to help inform a strategy for the implementation of preconception assessment of reproductive genetic risk in the general practice in the United Kingdom. The aim of this thesis was achieved by carrying out three components of work. These components of work involve three domains that could assist in the implementation; the interventions; the settings; and exploring attitudes and opinions. 1. The first component involved carrying out a systematic review of literatures on the effectiveness of preconception care interventions in the primary care settings. 2. The second component involved carrying out a systematic review of literatures on the effectiveness of preconception assessment of reproductive genetic risk. 3. The third component involved a postal questionnaire survey of GPs practicing in the Nottinghamshire, Nottingham City, Derbyshire and Derby City Primary Care Trusts, exploring their attitudes and opinions. A new questionnaire was developed as the study instrument for this study. The first component of work has synthesized the evidence of the effectiveness of preconception interventions in the areas of maternal knowledge of pregnancy-related risks; self-efficacy and health locus of control; risk behaviour modification (for example, folate and alcohol consumption); adverse pregnancy outcomes (for example, congenital anomalies and preterm birth); and psychological consequences. The review has identified that both risk specific interventions or interventions involved multifactorial risks, both demonstrated significant improvement in maternal knowledge, self-efficacy and health locus of control. There was positive evidence for risk specific interventions in the areas involving risk behaviour modification. However, the effects for adverse pregnancy outcomes and psychological outcomes remained unclear. The second component of work sought to find evidence the effects of preconception assessment of reproductive genetic risk. The scope of literature search included family history and ancestry assessment, pre-carrier test education or consultation and carrier testing or screening. It was not possible to draw clear conclusion regarding its effectiveness as only two studies involving assessment of cystic fibrosis and haemoglobinopathies were identified. Nevertheless, the studies have provided information on potential benefits of preconception assessment of reproductive genetic risks on reproductive decisions, knowledge and understanding of carrier risk as well as psychological benefits. The third component of work involved self-administered postal questionnaire survey. The impact of this survey is restricted due to low response rates. Nevertheless, the results of this survey indicated that a substantial proportion of GPs were already offering or providing preconception assessment on reproductive genetic risk opportunistically, in particular, with women planning pregnancy and women with known family history of genetic conditions. Even if they are not offering of providing preconception assessment on reproductive genetic risk at present, majority of them indicated that they are prepared to offer and provide the service, especially when consulting women planning a pregnancy or women at-risk. Their primary concern was how to reach these women as not many would come to consult GPs for preconception advice. This study has demonstrated that family planning clinic was the most preferred primary care setting to offer preconception assessment on reproductive genetic risk. In the United Kingdom, family planning clinics serve a large proportion of women of reproductive age group, thus, this setting may provide opportunities to introduce preconception care and reproductive risk assessment including genetics. While there is paucity of evidence from the systematic reviews in my thesis that could impact on the direction or implementation of offering preconception care addressing genetic risks, many factors other than scientific evidence can influence the implementation process. Observational studies have demonstrated potential benefits of preconception care specifically preconception assessment of genetic risk interventions such as early antenatal diagnosis to informed reproductive decisions. Broad interests from the international organization such as in the United States and Netherlands have a role in the implementation. Similarly, interest from the stakeholders in particular individuals of reproductive age groups and the primary care providers also may influence the development of the interventions. In this context, the GPs that participated in the survey have provided important information on opportunities and barriers, and potential ways to facilitate its development. Nevertheless, analysis of the data has identified some areas that were not fully addressed in this thesis and this is discussed in the final chapter.

Indonesia has one of the highest rates of maternal mortality in Southeast Asia. Community participation has been known to reduce maternal mortality in some areas in Indonesia. However, in Aceh Province, the prevalence remains higher than the general Indonesian maternal mortality rate. The aim of this research was to gain an understanding of pregnancy and childbirth experiences from multiple perspectives, in relation to the use of maternal health services in Aceh, Indonesia. The conceptual framework was based on the importance of community engagement in improving maternal health. A qualitative study design with a grounded theory approach was utilised. This approach was chosen in order to gain an understanding of the social processes and ways in which experiences of pregnancy and childbirth are related to maternal death incidents. The process inherent in the method enabled the emergence of important theoretical concepts. A theoretical sampling strategy was employed. The data collection used multiple methods that involved a series of in-depth interviews, observations and focus group discussions with women, family members, a village leader and health professionals. The sample size was determined by data saturation (19 women, 15 family members, 7 health professionals, 3 kaders, 4 student midwives and 1 village leader participated). Ethical approval was gained and the research setting was in the two villages of Aceh Besar District, Aceh Province, Indonesia. Data were coded and analysed by following a constant comparison process. The emergent core category, entitled “maternal death: the elephant in the room” explains the views of the community about maternal death incidents in the research setting. The research findings highlighted that despite the maternal mortality rate still being high in the region, maternal death was not focused upon, as a problem within the community. The research findings revealed that maternal mortality was a hidden problem within the community, and was related to inadequate maternity practices in the village. The core category “maternal death: the elephant in the room” was found to consist of four interrelated categories. The categories of the value of midwifery in the community, desicion-making of maternity care, social control of the childbearing and distancing of maternal deaths; explain maternity practices in the community. Understanding of social processes related to maternal health can assist in informing strategies to improve the quality of maternal healthcare in Aceh Indonesia.

United Kingdom (UK) health policy over the past thirty years has been predicated on a partnership model focusing on empowering service users to be fully involved in their care. Within maternity care partnership relationships have been conceptualised as empowering women to have continuity, choice and control (Department of Health (DoH), 1993), within a relationship of personal autonomy between the woman and her carers. In this study I sought to identify the extent to which the Government agenda for partnership working and choice is realised or desired by women during pregnancy and childbirth. In addition, I wanted to examine the level of alignment between the views of midwives with that of women accessing the maternity services. This study took a qualitative approach, drawing on the principles of grounded theory. In the first phase of the study a purposive sample of sixteen pregnant women were recruited and invited to complete a diary and to take part in two interviews. Women maintained diary entries following appointments with the midwife during pregnancy and childbirth. Semi–structured interviews were undertaken at 36 weeks of pregnancy and four weeks after the birth, based on the diary entries. In the second phase, four focus groups were undertaken with two groups of community midwives and birth centre midwives from two National Health Service (NHS) Trusts. Quotes from the diary-interviews from phase one were utilised to develop three vignettes which acted as a prompt during the focus group interviews. Following a thematic analysis of the data, I analysed women’s views on partnership working and choice. Most women in this study did not feel that they developed a partnership relationship with the midwife. This was associated with a lack of continuity of care and insufficient time to engage in meaningful discussion in an environment which was not conducive to shared decision making. Women described wide variations on the midwives role in supporting decision making. This ranged from decisions being dictated to midwives guiding choices and for some women, being facilitated to make informed choices. Many women described input of family and friends and widespread use of the internet as an information source. Women depicted their antenatal midwifery care as medicalised and felt that whilst their bio-medical needs were met their psycho-social and emotional needs were not. Women described the visits frequently as ‘in and out’ or ‘ticking the boxes’ to describe this approach to care. A small number of women (n=5) did experience a partnership relationship. Three of these women knew the midwife from a previous pregnancy; the remaining two women attended a midwifery led unit for all of their care. In relation to the choice agenda, most of the women who participated in this study were not aware that they had a choice about who provided their care or where they would have their care. The midwife focus groups concurred with the women’s findings and suggested that a lack of time was a significant factor hindering the formation of a partnership relationship. Midwives felt that this was exacerbated by the paperwork they were required to complete in order to audit care and meet the ‘payment by results’ agenda (DoH, 2003b). During the focus groups midwives identified strategies which could be implemented to enhance midwifery led care, including offering antenatal care to small groups of women and undertaking an antenatal home visit towards the end of pregnancy, to provide women with the time to discuss any issues that they wanted to explore in more depth. The findings from this study contribute to the current body of knowledge on midwifery led care particularly in providing the women’s perspective on partnership working. Women want to experience midwifery care that meets their psycho-social needs as well as bio-medical needs through a model of care that provides continuity. In contrast to previous research findings, the women in this study described community based care as mechanistic, clinically focused and time bound, more in line with an obstetric model of care than a midwifery model. However, midwifery led care offered within a birth centre was perceived by women as providing a more holistic, social model of care. Whilst continuity of care is not a new concept, what this study contributes is that despite successive administrations supporting partnership working and informed choice over the past twenty years, most of the women in this study did not experience this level of care. The findings from this study resulted in the development of a midwifery partnership model as a theoretical framework that could be utilised in future research studies to evaluate the extent to which a partnership relationship exists within a range of midwifery care settings.

The thesis examines care-seeking for first births in low-income settlements of urban India. Care-seeking is framed as a dynamic, social process. The thesis shifts the research focus from non-use of maternity services to a more holistic notion of care-seeking strategies, and examines how they are shaped by patterned social relationships and their content. The study combines a prospective, qualitative design with multiple household perspectives. Seventy-seven in-depth interviews were conducted in 16 households. Matched data were collected for primiparous women and other household members, and interviews were conducted prospectively during pregnancy with a follow-up after birth. The study was conducted in Indore, a large city in the central Indian state of Madhya Pradesh, where a range of maternity care providers operate in a complex urban health system. This population could be characterised as strategic care-seekers, aware and discriminating across the range of care options available. Managing perceived risks was central to strategies, but solutions differed due to variation in perceptions of risks and their management. The notion that childbirth requires medical management was dominant. Yet, health facilities were also regarded as a potential source of risk. Strategies were plural and contingent, combining different providers across and within sectors, giving households control and flexibility in dealing with unfolding circumstances. Local narratives apportion responsibility for care-seeking to the household in which the woman is staying for the birth. The value placed locally on household-level ‘responsibility’ contrasts with the focus on women’s autonomy in the literature on maternal health. A corollary of responsibility is blame in the event of an adverse outcome, which impels households to seek care that meets expectations among their social ties. The thesis generates new insight on an issue that has previously been examined largely with static approaches, underpinned by individual rational actor assumptions. Findings reveal care-seeking strategies that go beyond a decision on whether or not to use a health facility. This partly derives from a complex urban health system providing choice, but it is also a response to the challenges households face in negotiating the health system to receive care they perceive to be ‘safe.’ The findings have implications for the policy goals of increasing births with a skilled attendant and improving quality of care.