Maternal deaths in Pakistan: intersection of gender, caste, and social exclusion

Complications of pregnancy and childbirth remain the leading cause of death and disability for childbearing women in many low and middle income countries. Of the estimated 529,000 maternal deaths globally each year, 99% take place in the developing world [1]. over half in just six countries [2]. The obvious remedy is to provide competent maternal health care during pregnancy and childbirth [3–5]. Contemporary policy espoused by the World Health Organization and the United Nations Population Fund (UNFPA) promotes this medicalized approach, which subsequently has grounded the national maternal health policies of many developing countries, including Pakistan [6, 7].

However, despite significant international investment under the Safe Motherhood Initiative and strategies to achieve Millennium Development Goal No. 5, countries with the highest burden of maternal mortality are failing to reduce their mortality rates [8–10]. This suggests the medicalized approach may be necessary but is insufficient to achieve the desired reductions in maternal mortality. A large, but parallel body of public health and development literature demonstrates that women’s use of reproductive health services is determined not only by the physical availability of such services, but also by gendered inequities that shape women’s access to education, ability to travel, financial and social resources, and decision-making authority in key aspects of their lives [11–13]. These gendered inequities are further intersected by class and caste-based hierarchies, resulting in large differentials in access to services between rich and poor women within countries. Consequently, maternal health care is one of the most inequitably distributed health resource even within poor countries. A recent analysis of Demographic Health Survey data from 33 low and middle income countries shows that the odds of having a skilled birth attendant at delivery for women in the poorest quintile are 94% lower than for women in the highest wealth quintile [14]. Women with primary education are five times more likely to have skilled birth attendance compared to women with no formal education [15–17].

Generally lacking from this body of literature, however, is insight into the underlying political and social dynamics that create and reinforce these inequities in access to care. An understanding of the root causes, and the political and social dynamics that reinforce these inequities remains a critical gap in our knowledge. In particular, we know very little about the ways in which economic class and gender influence healthcare policy formation, resource allocation, the design and delivery of maternal health services, and disadvantaged women’s access to health services. Material poverty alone does not explain the large disparities in access to maternal health care between the rich and poor [18]. Poverty is relational and embedded within power hierarchies [19, 20]. We therefore drew upon the concepts of social exclusion and adverse incorporation [21–25] to understand how power hierarchies operating at family, community and state levels conspire to prevent women from accessing timely maternity care despite the physical accessibility of maternity services [19, 20]. Three frameworks grounded our inquiry: the Monopoly paradigm describes cultural distinctions such as class, political power, and economic priorities as perpetuating inequality and domination [26]. Kabeer expands this approach to include perceptions of social identity related to caste membership [27], and the “social relations” analytical framework addresses gender [28].

The current paper presents early findings from an innovative research study in rural Punjab, Pakistan that aimed to address these issues. Pakistan’s maternal mortality rate stands at 297/100,000 live births, among the six worst rates in the world. Pakistan is also characterized by deeply entrenched gender inequities [29] and vast inequities in access to maternal health care services, illustrated by the fact that while 77% of women in the highest wealth quintile report receiving skilled care at delivery, this figures is just 16% in the lowest wealth quintile [30]. Two broad questions were addressed by the research 1) how are the axes of social exclusion constructed, mobilized and experienced in Pakistan, particularly in relation to class and gender divisions? 2) how do these forces affect access to maternal health services?

Through the case histories of two women who died shortly after childbirth that we present below, we can begin to see how gender and socio-economic hierarchies, underpinned by societal discourses that sytematically devalue particular sub-groups and legitimize discriminatory behaviors, and shaped by wider ideologies and economic structures, contribute to tragic outcomes in a location where maternal health services are physically available.

Our study adopted a critical ethnographic approach, which allowed us to synthesize the traditional ethnographic focus on subjective meanings and beliefs of the respondents with the insights gained from a broader historical and structural analysis [31]. The ethnographic method, with its emphasis on immersion in the research setting enabled us to develop a “thick description” of the cultural understandings of caste, class and gender, and to elicit subtleties of power and resource distribution [32].

The findings presented in this report are based on data generated in the first six months of field work in the village of Gaind-Pind, Northern Punjab, Pakistan (note that the name of the village has been changed). Data were collected in four overlapping phases by a team of four researchers (three women, one man). The primary investigator (PI) (Zubia Mumtaz) lived in the village for three months and maintained regular contact with the team in the field. In Phase One, four social mapping exercises and informal home visits enabled us to introduce ourselves to the villagers and familiarize ourselves with village life. Phase Two consisted primarily of informal interviews and observation/participation in the daily life of the village. Investigators digitally recorded detailed, twice-daily summaries of their activities, observations and conversations with villagers, supplemented by their personal impressions of the settings and behaviors in order to provide commentary on the context and meaning of narratives and not merely their face-value content. Phase Three focused attention on pregnant women and members of the lowest castes in the village, continuing to use mixed methods. Phase Four focused on constructing detailed case studies of the five maternal deaths that had occurred in the village within the past five years (one during the field work period). Interviews were conducted in Potohari and Urdu, and digitally recorded if permission was granted. All recordings were translated and transcribed by native Potohari and Urdu speakers under close supervision of the PI (ZM). The transcripts were open coded manually and recategorized. The two case studies we present below were pieced together from information gathered from different sources and the data was searched for patterns and insights regarding the events surrounding the two maternal deaths and their inter-relationships with gender and socio-economic class.

The objective of this paper was to understand the underlying contextual factors that explain why some women experience adverse maternal outcomes, while others do not. We argue poor maternal health and lack of access to maternal health care services is a social phenomena and a form of deprivation. Despite the different details of caste, family dynamics and specific circumstances, the stories of Shida and Zainab illustrate how rigidly structured economic hierarchy, the gendered devaluing of females, and the reinforced lack of control that many impoverished pregnant women experience, conspire to keep them from lifesaving health services that are physically close by and should be at their disposal.

Both women belonged to the lowest castes and were impoverished. Poverty underlay the decision to let Shida die and poverty was the key reason no health care was sought for Zainab despite physical availability of services. Our data show that poverty is not merely inherited or accidental, but is closely interwoven with the caste system. Indeed, of the five maternal deaths that occurred in the five years prior to our fieldwork in the village, four were Kammis. The data also shows that the Kammis are systematically denied access to social and economic resources (land ownership, cash paying jobs, and education opportunities) and the possibility of upward social mobility. However, they are not completely excluded from the structures of society, rather by performing essential work that others refuse to do, the lowest castes are adversely incorporated into the local economy that would collapse without their efforts. The indentured servitude of the bhattas is a clear example of how rigidly adverse incorporation can be enforced. Gender inequities further interacted with caste and poverty to kill the two women. Domestic violence is a gendered phenomenon and violence particularly during pregnancy is well documented [33, 34]. It serves to perpetuate patriarchal power and is sustained by a culture of silence and denial of the seriousness of the health consequences of abuse [35]. None of the respondents, the marital family, the neighbors, not even the health care providers who conducted the Zainab’s maternal death inquiry made the possible link between the violence and subsequent death.

The sharply demarcated gendered roles in which men are socially constructed as providers of economic resources means that women may become particularly vulnerable if their husbands are unable to provide. Poverty and gender values that perceive women as not “worth” spending money on for medical care [36], in this case maternal health care, means that women whose husbands are unable to provide are doubly vulnerable.

This research began as a collaborative project involving the University of Alberta, Canada; the Health Services Academy, Pakistan, and Sheffield Hallam University, UK. Researchers at University of Alberta and Sheffield Hallam University conceptualised the research with support from the local partners in Pakistan. The PI had a joint appointment at the University of Alberta and the Health Services Academy. However, by the time research funding was secured, a change in management at the Health Services Academy and the loss of one local co-investigator who left to pursue her PhD studies meant that the originally planned institutional collaboration could not be actualized to the degree planned. Module 1 of the research project was therefore conducted independent of the local partner. The PI – a person of Pakistani origin - has extensive knowledge and connections on the ground; it is possible that researchers without such extensive local knowledge and relationships may well have had their project collapse under such circumstances. A key lesson learned from this was the importance of individual researchers in maintaining institutional partnerships.

Knowledge translation was a key objective of this research. In the absence of local partner involvement, the PI (ZM) has developed direct and ongoing engagement with policymakers in the Pakistan government, specifically the Maternal and Neonatal Child Health program and bilateral aid agencies (DFID). Acknowledging that there is an inadequate understanding of the deeply rooted multi-dimensional inequities embedded within the societal structures and the role they may play in the high maternal mortality rates, the policymakers appreciate the research as addressing an important gap in knowledge. Nonetheless, personal one-to-one engagement with key stakeholders suggests that the sensitive and contentious nature of the topic mean that discussions involving gender, caste and related social distinctions will remain uncomfortable. Despite the increasing awareness, health policymakers feel that it is difficult for them to address the deep-seated social and economic inequalities that are reinforced by multiple hierarchies.

We conclude that until the complex, underlying dynamics that determine healthcare access are better understood, programs aimed primarily at expanding health services are likely to continue to fail to address inequities and to reduce the level of maternal mortality. The findings of this research have relevance to global health leaders and to maternal health advocates in particular, for they demonstrate that a purely health systems strengthening approach to reducing maternal mortality is inadequate. It is equally important to understand and address gender, caste-and class-based inequities.