Addressing disparities in maternal health care in Pakistan: gender, class and exclusion

Disparities in maternal health care between countries are well documented. Recent evidence suggests there are equally large disparities within countries [9, 14]. For example, in Chad, rich women are 23 times more likely to have a skilled attendant present during childbirth compared to poor women; in Bangladesh the difference factor is 14 [9].

A growing body of research indicates that economic poverty alone does not explain the large disparities in access to maternal health care between the rich and poor and how such inequities might best be challenged and addressed [16]. Economic poverty is relational and embedded within power hierarchies [17, 18]. The concept of social exclusion may therefore be a more useful concept to draw upon to understand the disparities in access to maternal health services. Originally developed in French political discourse [19], social exclusion is a complex concept that is variably understood to range from individual ostracism or poverty to systematic societal structures that sharply define hierarchical group boundaries that serve the interests of a privileged group [20‐25]. In the latter concept of exclusion, termed the Monopoly paradigm, social institutions such as class, political power, economic structures, and other cultural distinctions perpetuate inequality and domination [24]. In South Asia, the concept of social exclusion has been expanded by Kabeer (2006) and others to include perceptions of social identity related to caste membership [26]. In these contexts, powerful groups construct and draw upon societal beliefs, norms and values to disparage, invisibilise and demean certain groups of people, thereby justifying the denial of full rights of participation in economic, social and political life [27]. Notwithstanding the important debates around the term and concept of social exclusion [27, 28], there is an emerging consensus that it is both a process and an outcome [24], and that the discrimination occurs in private social institutions such as households as well as public institutions such as the health, education and legal systems [29].

The social exclusion perspective is a useful guiding framework for the present research because it focuses attention on social relations and integrates various forms of disadvantage within a single analytic framework. The social exclusion discourse is particularly well developed in understanding the relational aspects of deprivation, how some groups are systematically denied the material and social resources that facilitate full participation in the society and the subsequent devaluation of classes thus created. It is also a useful framework to generate a greater understanding of how women in Pakistan, as a gendered group, may be socially excluded or, as some suggest, be adversely incorporated [29, 30] The framework will enable us to understand, in all its complexity, how the gendered social structures, processes and relations that lead to women being systematically denied resources and full participation in society become the basis for denial of maternal health services.

Historic experience from Europe and recent empirical evidence suggests health care systems are the key to reducing maternal mortality [8, 30, 31]. More recent, cutting-edge thinking suggests that health systems are not just mechanical structures that provide health care; they “are also purveyors of a wider set of societal norms and values” (p.143) [32]. According to Freedman, health systems should be considered “core social institutions, culturally embedded, politically contingent, and part of the very fabric of social and civic life” (pg. 22) [33]. More importantly, they are a common interface between individuals and the power structures that shape their broader society. The structures and operations of health systems may, therefore, signal and reinforce societal values and norms [33]. Neglect, abuse and exclusion within health care systems may thus essentially be a reflection of the experience of being poor and socially marginalized in that society [33, 34]. Feminist bioethics scholars have long examined the dynamics of gender and western health care [35‐37], often extending the analysis to matters of race [38], socioeconomic status, and related social categories. Similar inquiries regarding gender, power dynamics and health care in developing nations are still limited [39].

It is necessary, but not sufficient, for health systems to ensure that good quality maternity services are widely available; utilization of these services also involves the dynamics of users’ decision-making. A large body of literature addressing the three-delays-model [40] of maternal health care utilization in developing countries has tended to focus on the individual characteristics of women, their families and decision-making processes. This individualistic approach fails to consider how health system characteristics per se may shape the decision-making process, including the primary decision of whether or not to seek care at all. For example, health system characteristics such as abuse and under-the-table bribes may deter women and their families from seeking the care they need [41], yet avoidance is often regarded as a failure of the individual woman and her family to make appropriate decisions [8]. A better understanding of the ways in which health systems, as social institutions, may be systematically creating barriers to use thereby excluding particular sub-groups of the population, and how such exclusion can be challenged is urgently needed.

The structure of Pakistani society is acknowledged to be characterized by two separate institutionalized systems of inequality: class and gender [42, 43]. Class in Pakistan is closely related to zaat, a multi-dimensional identifier similar to the Hindu concept of caste [44, 45]. Severe inequities in the distribution of land, the primary resource, are interwoven with notions of zaat to create a hierarchical and feudal society. Ethnographic work from Northern Punjab shows that only members of the Kammi zaat work in low prestige occupations such as butcher, barber, tailor and carpenter. They are also considered “katia” (low class) and “kameenee” (low form of life) (pg.167) [44]. Individuals born in these zaats largely remain locked in such occupations, a division of labor that causes and reinforces differential access to and control over resources of all types, including social status [46].

Gender inequality is Pakistan is particularly acute. Gender roles are clearly demarcated; men are socially constructed as providers and women as dependents [47]. The social institution of purdah ² [48] further demarcates boundaries between men and women and sets standards of female morality. Consequently, large differentials exist between women and men in access to resources of all types [49]. This of course does not suggest that gender roles and relationships are non-negotiable or that there is no variation. Pakistan is a large and heterogeneous country, and women’s gendered experiences and social relations are shaped by socioeconomic, ethnic and regional variations [45, 50, 51]. Nonetheless, the coexistence of social stratification in relation to class and gender means a woman may be disadvantaged because she is a woman, but her disadvantage is compounded if she is a woman belonging to a lower status zaat [44, 52].

Ethnographic and survey data from Pakistan also suggest gender values and norms are tightly interwoven with maternal health-seeking behavior [44]. Moreover, the degree of adherence to these gendered norms varies by zaat and socio-economic class [44, 51‐53]. For example, gendered notions of purdah glorify women’s seclusion, but poor women often cannot afford to adhere to these norms when the practical needs of survival necessitate their mobility outside the home to collect water, work on the farm, or even relieve themselves in the fields. Studies show that poor women’s greater mobility outside the home renders them vulnerable to sexual exploitation by the higher status zaat or land-owning men [44, 53]. Clearly, these issues have major implications for poor women’s ability to access maternal health services, many of which are located at considerable distance. There is an urgent need to better understand how class and gender interact to create barriers to women’s access to maternal health care.

An evolving global maternal/reproductive health discourse has resulted in formulation of three maternal health/reproductive health policies in Pakistan between 1990 and 2001 [54]. The first two policies were silent on gender and class in reproductive health services [55]. The 2001 policy does identify gender equity as one objective without expanding what this will entail. The latest National Maternal and Neonatal Strategic Framework, 2005–2015 has, however, pledged to “ensure availability of high quality maternal and neonatal health services to all, especially the poor and the disadvantaged” [56].

Despite this apparent political commitment, available evidence suggests that delivery of maternal health services in Pakistan is patchy and fragmented, of poor quality, and inequitably distributed and utilized [55, 57, 58]. Notwithstanding a wide network of facilities, a survey of public sector health facilities in Punjab showed that only 13% of facilities designed to provide basic emergency obstetric services were actually doing so [59]. The corresponding rate in North West Frontier Province (NWFP) is 10% [60]. The paucity of services is reflected in the very low C-section rates: 0.4% in Punjab and 0.89% in NWFP, when rates of between 5 -15% are generally deemed essential for safe childbirth [60, 61].

In seeking to understand the persistently inadequate maternal health care services, research has largely focused on parameters of service delivery [58, 59]. A few studies focus on poor governance [60]. More critical analyses argue that although the most recent health policy in Pakistan focuses on women’s health, it fails to recognize and address the key gendered obstacles to uptake of maternal health care services [61]. These include a paucity of female health care providers, resulting in part from gender hostility in employment settings [62] as well as limited education opportunities for females. Clearly, a robust discourse around class, gender and exclusion in the use of maternal health services has yet to start in Pakistan.

A key challenge for describing the multidimensional class and gender systems is that deeply embedded social norms are rarely transparent to those who live within them. Therefore, we will conduct a village ethnography [63]. The ethnographic approach, with its emphasis on immersion in the research setting [64] is particularly suited to Module 1 research questions because it will allow us to develop a “thick description” [65] of the cultural understandings of zaat, class and gender and elicit subtleties of power and resource distribution. A critical ethnographic approach will be used, in that we will synthesize the traditional ethnographic focus on subjective meanings and beliefs of respondents with the insights gained from a broader historical and structural analysis [66]. Since respondents are unlikely to give concrete answers to direct questions about gender and caste, insights will be acquired by asking participants about maternal health and associated health care experiences as starting points. The data will be collected in four phases; insights from each phase will refine and inform the research questions and methods of the next, thus increasing focus and depth as the project progresses.

To reconnect with the villagers and gain acceptance into the community, the PI will conduct social mapping exercises, visit homes and meet with the village elders and power-brokers. Social mapping is a method most useful for producing a local perspective of the village geography and socio-economic and related power structures, depicting information that outsiders cannot see by direct observation [64]. As a participatory research method, the local informants are treated as knowledgeable experts. Four social mapping exercises will be conducted; two with 6–10 women and two with 6–10 men. Participants will be asked to draw maps of the village, the larger region, and the locations of available health services. The zaats biradaris and land-holding patterns in the village will be identified from these depictions.

Focusing on research questions 1 and 2 (Module 1), the two main methods used in this phase will be participant-observation and informal interviews [67]. Participant-observation, a key characteristic feature of the ethnographic approach, is a method of data collection in which the researcher observes the people as they go about their daily lives and participates in their activities [68]. Underlying the method is the belief that knowledge about a community is different from understanding it, and that the latter can only be determined and obtained by actually living in it [69]. This method is suited to our research question because not only are gender and class sensitive issues, they permeate all aspects of life [70]. The research team members will live in the study village and immerse themselves in the lives of the villagers by helping with daily chores and participating in community events. The team will observe and record the various activities that constitute the daily life of the villagers. Informal conversations will constitute “informal interviews”.

This phase will explore in greater depth how zaat, class, and gender hierarchies play out in social relationships, access to resources and social opportunities by using semi-structured in-depth interview and focus group discussion methods [69, 71]. Individual and group interviews are useful methods for not only exploring people’s understandings, interpretations and experiences of zaat, class and gender, but also of the language they use in constructing the discourse related to these concepts [72]. Sample size in qualitative research is difficult to ascertain and specify in advance. However, considering that 12 interviews are usually enough to reach data saturation [73], we estimate we will need to conduct about 20 in-depth interviews with women and 20 with men. The respondents will represent both the rich and poor groups, and all shall have had children in the past 10 years. In addition, ten older women will be interviewed since they are often the primary decision-makers around maternal health issues [54]. Eight focus group discussions, with 6–10 participants in each, will also be held, separately for women and men [74]. Representation of all zaats and socio-economic groups will be ensured. By this stage, the research team will be familiar with the villagers and will purposefully select the participants. Guidelines for the interviews and focus group discussion will be based on data collected during phases 1 and 2, with reference to the research objectives.

The final stage will be a tightly focused and in-depth exploration of the complex interactions among class, zaat, gender, and related distinctions and how they interact with maternal health-care-seeking behavior. Six longitudinal case-studies will be conducted with pregnant women belonging predominantly to disadvantaged groups [75]. Data will be collected via multiple open ended interviews, accompanying women to any antenatal care appointments or interventions, observing deliveries and subsequent post-natal care. Husbands, mothers-in-law and other family members will also be interviewed to enrich each case study. The specific focus within these longitudinal case studies will be the experiences, perceptions, rationales and decision processes that influence choices of birth attendants, the location of delivery (home or health centre), and whether to seek emergency care if a need was identified.

Each phase builds upon the next to achieve an increasingly detailed understanding of the subtle interactions among social concepts, individual experiences, and subsequent healthcare-seeking decisions. Because individual decisions are shaped by both social and personal factors, this in-depth and multi-dimensional exploration will provide a more realistic vision of the context in which maternal health policy and health service delivery must operate.

Data will be collected using audio recorders and note-taking depending on participant consent. Since the data will be collected in Punjabi, the local dialect, it will be translated and transcribed by native Punjabi speakers. The PI (who understand and speaks Punjabi) will double-check all interview transcripts based on audio-tapes to verify the translation and focus on conceptual equivalence. A database of the transcribed interviews, focus groups, and observation notes will be created in Atlas-ti, a qualitative data analysis software program [76]. Using a social constructivist, interpretative approach [77], data will be coded and themes identified. Initial coding will be guided by the stated research objectives and later by additional concepts as they emerge. Data analysis will be an on-going and iterative process through all phases of data collection, as early identification will allow a fuller probing of unanticipated concepts and variables in upcoming formal or informal interviews and focus group discussions. Interpretive accuracy will be assessed by triangulation of findings across the four phases, peer debriefing within the research team and other colleagues and respondent validation [78].

Module 2: Provision of Maternal Health Services (Empirical objective 2.2)

This module will address the following research questions, all of which are considered within the context of health care as an institution that mirrors and reinforces larger patterns of social exclusion:

Drawing upon the Hulton, Mathews & Stones (2000) framework for assessing quality of maternal health services in developing countries [79], both women’s experiences of care and key aspects of care provision will be examined. Women’s experiences of maternity care will include an assessment of their impression of the adequacy of human and physical resources, their understanding of the usefulness of the procedures they undergo, the respect they are accorded, their sense of dignity, equity and the emotional support they receive. Key aspects of the provision of care assessed will include the quality of human and physical resources; the quality of their referral links and information management systems; the use of appropriate technologies in caring for women and the facility/hospital’s adherence to internationally recognized good practice [77]. System-wide and policy-level contributors to women’s experiences of maternal health care will also be explored. To ensure maximal phenomenal variation [80], both the public and private health care sectors will be included. The public sector health care services are usually used by the poor, whilst the rich tend to use private sector care. In each sector, one primary health care centre/clinic and one hospital will be assessed as case studies, for a total of four health care settings. The following specific components will be undertaken using a range of methods:

Interview guidelines will be developed for each group of respondents to ensure a systematic approach, with flexibility to pursue new avenues of inquiry as insights accumulate. Because the PI is a physician herself, professional courtesies will facilitate appropriate patient contact and access to service providers and policymakers.

The data will be analyzed as follows:

Integration, synthesis and knowledge transfer (knowledge transfer objective 3)