(In)visibilising pregnancy loss in Southern Malawi

Abstract

Miscarriages, stillbirths and neonatal deaths have received limited attention in global health programmes and research, even though pregnancy loss is common, traumatic and stigmatised. This paper seeks to illuminate lived experiences of pregnancy loss in southern Malawi, drawing on findings from semi-structured interviews and focus groups with women who have experienced loss, health professionals and community members, and observations of maternity care. Combining thematic and discourse analysis, we show how societal and medical discourses frame women as responsible for (failed) reproduction, and restrict possibilities to speak about, and respond to, loss. Some accounts and (care) practices invisibilise loss and associated suffering. However, invisibilisation may also be intended as support, and underscores rather than denies the social significance of parenthood. Other accounts (e.g. women emphasising faith and acceptance) constitute moral survival strategies to avoid the acquisition of a ‘spoiled identity’. We conclude that societal and medical discourses of loss enact stigmatised, subaltern subject positions for women experiencing pregnancy loss, create social suffering, and amount to a form of structural violence. Programmes and interventions should change these discourses.

Keywords:

• Pregnancy loss
• discourse
• stigma
• structural violence
• Malawi

Introduction

Maternal mortality has been the focus of global health programmes and studies since the launch of the Safe Motherhood Initiative in 1987 (Maclean 2010). The Millennium Development Goals (MDG 5) and the Sustainable Development Goals (SDG 3) increased the focus of attention given to maternal deaths. Safe Motherhood, however, is about more than mothers’ survival. Obstetric morbidities, often paired with stillbirth (after 28 weeks gestation) and early neonatal death (ENND, within 1 week after birth), can affect women’s long-term health and well-being (Mooij et al. 2021; Storeng, Akoum, and Murray 2013). Approximately 15% of pregnancies end in miscarriage (loss before 28 weeks’ gestation), two million babies are stillborn every year, and ENNDs constitute 73% of all neonatal deaths (World Health Organization n.d.). Most of these losses occur in the Global South (UNIGME 2020). Yet, miscarriage and stillbirths are not routinely recorded and excluded from the MDGs and SDGs (Lawn et al. 2011). Midwifery training rarely addresses care for mothers of stillborn babies (Homer, Malata, and Ten Hoope-Bender 2016). Hence, as Frøen et al. (2011, 1353) state, ‘in times of global focus on motherhood, the mother’s own aspiration of a live-born baby is not recognised on the world’s health agenda’.

In a corrective move, we seek to illuminate experiences of, and responses to, miscarriage, stillbirth and early neonatal death in one Global Southern setting, southern Malawi, drawing on ethnographic data, literature, and our personal lived experiences in Malawi. We explore pregnant women’s biosocial subjectivities (Berry 2010), that is, how women’s experiences and understandings of the self and one’s place in the world, are influenced by biological processes (such as birth), social relationships, and societal discourses (Berry 2010; Foucault 1978). Hence, we explore how the discursive, social worlds in which loss takes place shape lived experiences of loss and create certain subjectivities for women who encounter loss.

Existing studies emphasise that pregnancy loss in the Global South is traumatic and stigmatising (Baffour-Awua and Richter 2020; Haws et al. 2010; Pollock et al. 2020). Like infertility (Bornstein et al. 2020; Dierickx 2022), pregnancy loss may lead to a ‘tainted self’ (Goffman 1963), since it violates strong parenthood mandates and easily invokes suspicions of intentionality, since boundaries between intended and unintended loss are blurred (Van der Sijpt 2014). Recurrent pregnancy loss may lead to divorce, extramarital affairs or polygamy (de Kok, Hussein, and Jeffery 2010), women may be reluctant to share their experiences (Haws et al. 2010), and many will ‘weep in silence’ (Kiguli et al. 2015).

Aside pronatalist ideologies, socio-cultural interpretations of personhood shape meanings of, and responses to, pregnancy loss. Van der Sijpt (2010) shows how in Cameroon, pregnancy loss can be classified as an event or ‘non-event’, depending on the ascribed ontological status of what was lost: blood, water, spiritual being or an actual human. As anthropologists have shown, personhood is socially constructed (Morgan 1997; Layne 2022), depending on the socio-cultural context and the pregnancy’s desirability (Van der Sijpt 2010; Andipatin, Naidoo, and Roomaney 2019).

Furthermore, critical anthropologists have noted how the political economy and structural violence shape interpretations and responses to loss. In a classic ethnography of child death in Brazilian favelas, Scheper-Hughes (1993) argues that structural, everyday violence like dire poverty and the church’s institutional neglect of early child death limits the space for grief. The institution of biomedicine may also silence pregnancy loss. Analysing South African women’s narratives of miscarriage, Andipatin, Naidoo, and Roomaney (2019, 553) argued that biomedical discourse and practice frames loss as a medical rather than emotional affair, involving ‘reproductive or pathological waste’. This normalises loss, overlooks the potential (constructed) personhood of the early foetus, and neglects women’s support needs.

Loss stigma, resulting from pronatalist norms, is aggravated by societal explanations of loss such as attributions of loss to a spiritual curse or witchcraft (Bakari et al. 2021; Kiguli et al. 2015; Van der Sijpt 2014; Sturrock and Louw 2013); to the use of contraception (Kopp et al. 2018); or to extramarital affairs (De Kok 2019). Conversely, women are deemed able to prevent reproductive mishaps by behaving ‘well’ (Bakari et al. 2021). These attributions frame women as responsible for successful reproduction and to blame for loss (Baffour-Awua and Richter 2020; De Kok 2019; Pollock et al. 2020).

Societal discourses strongly govern and control women’s conduct and reproduction. These forms of ‘bottom up’ reproductive governance (Morgan and Roberts 2012) push women to procreate, avoid reproductive risks, cry in silence when reproduction fails, and place them ‘on the defensive’, thus shaping discursive dynamics and accounts of loss. Van der Sijpt (2014) for instance, describes how women in Cameroon may emphasise their suffering to portray themselves as innocent and worthy Christians, thus maintaining their social status.

There have been calls for an exploration of the social processes generating stigma (Link and Phelan 2001), including stillbirth associated stigma (Pollock et al. 2020). In response, we explicate how societal and medical discourses enact stigmatised, disempowered subject positions for women experiencing miscarriage, stillbirth or early neo-natal death, whilst women also mobilise counter-narratives that uphold their moral standing. To explore these discursive processes, we paired thematic analysis with a hybrid form of discourse analysis, described below.

Methodology

During 13 weeks of fieldwork in two districts in southern Malawi, we conducted 37 semi-structured interviews with women (23) who had experienced a miscarriage, stillbirth or neonatal death; nurse-midwives (11); clinical officers (2); and a medical assistant. We also conducted 15 focus group discussions (FGDs) with health care practitioners (4 FGDs, 23 participants) and community members (11 FGDs, 87 participants).¹ The FGDs were mostly naturalistic: participants knew each other as colleagues (teachers), neighbours, or friends. We also observed antenatal and intra-partum care in two health centres, a district and a referral hospital. Data were collected in 2013, and complemented by recent publications, current social media discussions of pregnancy loss on Malawian media, and the Malawian authors’ on-going lived experiences and observations.

Participants were recruited through Under-Five and Antenatal Clinics, health surveillance assistants (community health workers) and personal contacts. Purposive and snowball sampling yielded a range of participants in terms of socio-economic status and residence (rural/urban). In town, four women with a loss experience had received tertiary education, four had obtained secondary education, and three primary education. Three were housewives; the others had small businesses, or worked as a hospital guard, teacher or researcher. In the rural area, two women had not received any education, the others had partly completed primary school. Most were subsistence farmers and housewives. Women’s ages ranged from 16 to 50, most were in their 30s. Complicated reproductive histories were common: eight women had experienced multiple losses (up to six), mostly between 6 weeks to 3 years prior. Three women were divorced, one had re-married.

In town, most FGD participants had received tertiary education. In the rural area, half had received secondary education, half primary school or no education. Most interviews and FGDs lasted over an hour, but some interviews were only 15 min in duration.

Ethical approval for the study was obtained from Malawi’s National Health Sciences Research Committee and Queen Margaret University, Edinburgh. We obtained oral and, in some cases (practitioners), written informed consent. We explained the study aims to participants, that they would not receive individual benefits and that they could withdraw at any time. Scarcity of mental health professionals prevented us from offering professional psychological support. We gave participants our phone numbers and identified health surveillance assistants who could talk to participants if desired. We ended the interviews on a more neutral topic and did not leave participants in (visible) distress.

Data were transcribed verbatim, pseudonymised, and translated from Chichewa to English by a professional translator, with extracts included in the paper double checked by the authors.

We began with a thematic analysis, inductively and deductively identifying and describing content-based recurring themes (Braun and Clarke 2006). Thematic coding was carried out with the aid of Nvivo. We coded the individual interviews and FGDs as one data set, but developed a separate coding tree for the observations made (although the codes overlapped).

FGDs illuminated shared understandings and societal discourses concerning reproduction, loss, marriage, and gender relations. Interviews generated individual narratives of lived experiences and enabled analysis of subject positions made available by societal discourses, and discursive strategies used by participants to build identities.

Our thematic analysis identified conflicting patterns that marked reproductive loss as painful and as a ‘non-event’. A hybrid form of discourse analysis illuminated this paradox. We used Foucault’s (1978) notion of discourse as a historically specific way of speaking, thinking and knowing about phenomena such as sexuality, or pregnancy loss. For Foucault, discourse includes speech, but also silence, images, architecture and practices (Rose 2016). Foucault’s work helped us unpack how societal discourses, infused with power, create particular forms of knowledge about pregnancy loss (e.g. as reflecting moral transgressions) and shape subjectivities. Discourse does not just impose itself on powerless victims, however, and Gee’s (2014) situated discourse analysis helped us explore how speakers use language to build actions (e.g. justifying behaviour) and identities (e.g. being a respectable wife), in a context where loss is socially costly. Finally, Gumperz (1982) interactional socio-linguistics enabled us to focus attention on how details of utterances (‘contextualisation cues’) signal contextual presuppositions and nudge the interpretation of statements.

Findings

Some accounts of loss seem to minimise its importance, yet others describe the social suffering it generates. Invisibilisation of loss reproduces stigma and underscores its significance. Furthermore, since loss risks spoiling a person’s identity and creating relational difficulties, subsequent accounts of that ‘loss’ can be used to repair a ‘tainted self’ (Goffman 1963).

Invisibilising loss

De-personalising the lost baby as a thing

Whilst the literature emphasises that loss is traumatic, narratives, especially those produced by poorer women, were often surprisingly brief and neutral, with common reference to the miscarriage or stillborn as ‘a thing’ rather than a person. Tiferanji for instance, 38 years and from a poor rural community, had a stillbirth. She said: ‘So when that happened, that thing was not properly born, it was born how, already dead’. However, nobody referred to a deceased newborn as ‘thing’, all referred to the neonate as a ‘child’.

How can we understand these neutral accounts and impersonal conceptualisations? Women in an urban FGD told us that they described a miscarriage or stillborn as a ‘thing’, ‘because it has not been seen in the outside world’, and ‘you are not acquainted with what is coming’. ‘Thing’ was deemed an appropriate term for something unseen and unknown, as other studies in Africa have found (Ayebare et al. 2021; Sisay et al. 2014).

A foetus may also be considered thing-like because ultrasound visualisations, that imbue the foetus with personhood (Layne 2022), are uncommon in rural Malawi. During the fieldwork, ultrasound was used only in the largest, tertiary referral facility, and women could barely see the ultrasound screen. Travelling technologies are socially appropriated by the contexts in which they are used (Muller-Rockstroh 2012); unlike in Euro-American settings, ultrasound scans were used in ways that did not foster the emergence of the foetus-subject or, in parallel, the mother-subject.

Neutrality in loss narratives may also be linked to poverty and structural violence. As Scheper-Hughes (1993) has suggested, if families struggle to meet immediate basic needs, this may reduce the space for grief. Scheper-Hughes (1993) warned against treating neutral responses to death as a cultural veneer, masking ‘true’ feelings of grief; this would be a problematic imposition of ‘Western’ psychological models of ‘normal’ mourning responses to child death. Equally, women’s brief, neutral accounts do not necessarily mean they are not suffering after pregnancy loss. There is much evidence to the contrary, as we detail later. For now, let us note how female focus group participants attach emotions to ‘the thing’.

R2: Waiting for 9 or 10 months, you get rid of the thing, and then the thing falls through

R3: And God takes it away from you.

Several: Aah it pains a lot.

R1: Will you see/have another one?

R2: There isn’t, for you to see another person like that one, it might be after two years, mmh, it becomes a problem.

Here, R2 describes a stillbirth as a thing (chinthu) and as very painful. Later, she refers to it as a ‘person’ (munthu). Hence the thing is imbibed with emotions and a sense of personhood. Other words also constituted the newborn or stillborn baby as human and suggested a feeling of loss, even in extremely brief accounts. For instance, women used the euphemism kumwalira, to ‘pass away’ rather than kufa, to die, normally used for non-human deaths.

There are additional reasons for the neutrality of loss accounts. Talking about pregnancy loss is complex, certainly in interviews. As Cecil ([1996] 2020, 2) notes, ‘feelings concerning simultaneous birth and death, the death of one who never was, may be virtually impossible to convey’. Moreover, discussing pregnancy loss is taboo in many African settings due to stigma, but also for practical reasons: talking or even crying about early loss is sometimes seen as endangering future reproduction (e.g. ‘closing the womb’ through witchcraft, Haws et al. 2010). More generally, in her study of maternal deaths in Malawi, Putter (2003) found that forgetting was seen as helpful for the grieving process. However, remembering, and thus narrating, was harmful for the individual and the community.

Displaying neutrality thus appears a contextually appropriate response to loss. Public silence, however, is not the same as absence of suffering.

Minimalist burials and professionals’ neglect

Funerals can facilitate grief and coping (Putter 2003; Sturrock and Louw 2013) by validating the social value and personhood of the deceased (Putter 2003). However, interlocutors in this study told us that death rituals and burials for late miscarriages, stillbirths and neonatal deaths are minimal, and parents may not take part in them, as other studies in Africa have shown (Ayebare et al. 2021; Sisay et al. 2014). This happened with Talimba’s second stillbirth:

• I: Okay, what happened here at the hospital? Did the doctor explain to you what happened?

• R: They just mentioned that ‘aah here it seems like what? There are problems. You have delivered a dead thing’. Me, ‘aah okay’. And they will also bury it here (at the hospital).

• I: Hmm.

• R: That’s all, just coming out of hospital and just going home.

Talimba relates matter-of-factly that health staff informed her that the ‘dead thing’ she delivered would be buried at the hospital. Similarly, Zione told us how her stillborn baby was buried by ‘strangers’.

Minimalist burials still enact a social meaning for the deceased. Based on fieldwork in Ecuador, Morgan (1997) argues that burials outside or on the edge of the graveyard enact foetuses and newborns as liminal persons, on the brink of personhood. In a sense, this minimises the importance of the loss. However, at the same time, minimalist burials confirm the personal and social significance of the loss, marking it as a bad and ominous death, that may be tsempho (a contagious illness). Other studies note how excluding the mother from the burial and minimalising public grieving about loss is considered a way to protect the mother from psychological and spiritual harm, and preserve her fertility (Ayebare et al. 2021; Sisay et al. 2014).

Like burials, health care practices invisibilise loss. Like Talimba, many interlocutors were not informed about what possibly caused the loss and generally received little attention. In crowded antenatal clinics (ANC), interactions were brief, impersonal and standardised. Previous loss experience was literally a tick box in the maternal health passport, only discussed when it had occurred more than twice. In the rural referral hospital, a student nurse told us how a woman whose baby had died was awaiting labour to commence after induction, surrounded by women with new-born babies and ignored by staff. This, he emphasised, was not a workload issue, but negligence. Another client whose baby had died intrapartum was wheeled into the labour ward. The nurse nor the accompanying guardian expressed ant sympathy; neither touched the woman or helped her when she slowly began to undress herself, whilst the nurse focused on technical procedures only (De Kok 2015).

Signifying loss

Loss pains the heart

Pregnancy loss was painful for women and men, of low and high socio-economic status, in rural and urban areas. Grace, 35 years old, and Agness, 40 years old, were both educated to tertiary level and living in town. Grace described her two miscarriages and stillbirth as ‘traumatic’. Agness told us how she felt when she miscarried at five months:

• Agness: ↑Ahm it was ↑² bad. Because you know, when you are pregnant it means you are expecting that you will receive a child in your hands (…) You feel sad. It was not easy. You feel sad because there is that kind of attachment between you and the child though the baby is still unborn, but the attachment is there. I felt that I have lost a child. Yah but then I had to cope with the situation.

• I: Yah and how was it like to go through that experience of having the baby…?

• Agness: It’s not a good experience, honestly speaking, miscarriage is not a good experience, better to have a child because at 5 months, that’s a ↑child … You know at 5 months that’s a child, I had to go through labour(.) Yah at the hospital, I laboured and I knew that this labour is in vain, it’s just because we want to, the baby to go out.

Agness highlights the impact of her loss by saying ‘honestly speaking’, a ‘contextualising clue’ that frames her account as the truth, that may contravene expectations. She underscores her sadness through repetition (‘You feel sad’, ‘It’s not a good experience’), reference to her attachment to the unborn, and the construction of her five-month pregnancy as a child (with high pitch for emphasis). Similarly, Khumbizeni, living in a village, highlighted the impact of losing her seven-month stillborn twins by emphasising they ‘were people’: ‘I felt pain in my heart because what was lost were people, it was my blood from my body, I just laboured, going aside [miscarrying], I didn’t gain anything’. Others emphasised the pain by noting its enduring nature. Chikondi, from a rural community, told us how three years after her loss: ‘Even now I feel pain. If that boy was still with me, he would have grown up as I am speaking now’.

Adding to kupweteka mtima or ‘paining of the heart’, is the fact that couples invest in a pregnancy, emotionally, physically, practically, and financially (Simwaka, de Kok, and Chilemba 2014). In an urban FGD, a female participant told us how, ‘you spend money to buy things waiting for the thing that is to come and use. Now you find out that the thing is lost and you just look at the things that were bought, that brings a lot of sadness’. Items bought (a basin, clothes), do not just reflect ‘wasted’ money. As Layne (2022) explains, purchasing items in preparation for the baby are ‘constitutive acts’ that ‘call forth’ or enact the baby and ‘the child-to-be’s gift-giving network’, including fathers and mothers (Han 2013, 168). When these roles do not materialise, items acquired become painful reminders of expectations shattered.

Loss thwarts personal and social expectations. In African settings, procreation is a core purpose of marriage, a marker of adulthood, and a source of respect (Bornstein et al. 2020). Tellingly, the Chichewa word banja denotes both marriage and family; societal discourse equates marriage with children. Failure to reproduce (regularly) can then become a source of embarrassment and mockery, and a woman may be deemed an unworthy, ‘useless’ wife by the spouse or in laws, as a male FGD participant told us.

Discourses of reproduction and marriage thus produce a spoiled identity (Goffman 1963) for women who lose a pregnancy. Women would like to hide their ‘tainted self’. Grace told us: ‘I didn’t want people even to know that I have miscarried again (…) Especially with the last one I felt bad about it (.) because the pregnancy had already (.) you know shown. And I felt bad.’

Causal attributions often blamed women for loss, adding to moral judgement and stigma (De Kok 2019). For instance, community members and health professionals attributed loss to women’s extra-marital affairs, or women not pushing correctly, or being ‘lazy’ during labour (De Kok 2019). In a rural FGD, arguments were identified as possible cause of miscarriage, hence ‘what is expected of a pregnant person is that she should be quiet. She should not be involved in quarrels’. Expectations that women remain quiet and avoid arguments indicate that they are framed as responsible for minimising reproductive risks to a rather extreme degree (Cecil [1996] 2020).

Loss: catalysing relationship trouble

Discourses of reproduction and marriage that circulate in Malawi society create the subject position of tainted, devalued, ‘useless’ woman and wife for those experiencing pregnancy loss, through speech but also actions. For instance, men having affairs ‘to prove that they are fertile’ (male FGD participant), arguably devalue the wife. In polygamous households, women may acquire a marginalised position following loss. As one male respondent in a rural FGD explained, if a second wife ‘is giving birth properly, (.) this one def- definitely this one is ignored like that. The other one (second wife) will be more loved (…) and probably well cared for’. Women in a rural FGD vividly described the relationship trouble a pregnancy loss can trigger:

• R1: Staying in the family is about cooperation with your husband. So when you bear a child, meaning, saying that that family is what, it is fixed, yes, you feel open two people, saying that we have found our ‘counsellor’ [the child]. So when you marry you marry and bear a dead child then you know that your house/marriage will not go well no.

• I: Oooh, she has just explained that you bear a child, when you marry and then bear a child and that child dies then things won’t go well in that house [marriage], why?

• R2: Because if the child doesn’t live you start to quarrel, there is little love, the husband always shouts [at you].

• R3: Peace, you lack it.

• (…)

• R4: The husband and wife start despising each other. They start pointing fingers at each other. The husband says this is happening because of the wife’s home or the wife again says ahh not from our home but from whom, from yours, husband.

Here, I women construct the child as the cement which fixes ‘the house’ or marriage. Loss however leads to spouses ‘despising each other’ and ‘pointing fingers’. Whilst both partners may start by ‘pointing fingers’, usually men blame their wives. Tamandani told us how after her stillbirth her husband ‘was just furious in the heart’ and said ‘that’s it, let’s go, to yours, something like this can’t just happen all over a sudden.’ Chikondi’s husband blamed her for her miscarriage, telling her ‘you have taken medicine [to cause this]’. When their newborn also died he responded as follows:

He came here with some things that he had bought. … he had heard that “your wife is at the hospital, she has borne a baby”. When he came here with the things … he had bought, he entered … and he said “where is the baby?” When he heard that the child had died, that’s it …, he got up and left without even talking to each other.

By aborting his visit and taking away the valuable items (soap and a cloth), the husband enacts Chikondi as unworthy of his support. Just like material items ‘call forth’ and enact the baby and parents to be (Layne 2022), withdrawing them un-makes these enactments and breaks the bond that was being forged through the pregnancy.

Women of higher socio-economic status also risked blame and divorce. Grace, a university graduate, married a lawyer. He was supportive at first, but got angry after the third miscarriage, and told her, ‘it seems you are the one who has a problem, but me I don’t have a problem you know’. He then decided to leave her.

Loss may also disturb wider family relationships or aggravate existing tensions. The earlier reference to ‘finger pointing’ implied witchcraft accusations. Witchcraft is omnipresent in Malawi (Adolffson and Moss 2023), and several of our interlocutors thought that their in-laws had bewitched them, or their husband. Chisomo attributed her miscarriage to her mother-in-law’s ‘magic’, because she refused to pray in her church. Chisomo’s mother-in-law told her ‘you will bear us a child that doesn’t pray (…) in our church so you should go away’. Chisomo stayed, since ‘the husband still wanted me’, but they moved to another village. Ultimately the marriage dissolved due to the mother-in-law’s interference:

She came to take her son but he did not want to go. However, my mother in law insisted, ‘let’s go home, I will kill you, let’s go home, this is not a place (to stay)’. Later she took her son away, the family/marriage ended there, he never came back again’

Chisomo’s narrative underscored how disharmonious extended family relationships can be; relatives are deemed capable of killing a baby or spouse. Geschiere (2003) calls this ‘the dark side of kinship’. Since witchcraft suspicions reflect and aggravate social tensions, they will affect women’s biosocial subjectivities, their sense of place in the world, adding to the ‘cross’ they carry. One participant in a rural FGD explained this well:

There is sudden death (…) A child may have a fever and by evening we hear that the child has died. And that illness is not an illness of peace. That kind of illness breeds finger pointing. Who did these things? “Ahm, perhaps it is thatIe… Or it is the grandparent. It is so and so”… And so the consequence for the marriage is what, it is disturbed, not living well anymore. And (…) for the woman, for her to have safe motherhood with all the freedom, it is not possible. She becomes very troubled. The heart is tormented because of t’oughts’

Witchcraft suspicions and blame lead to a troubled heart and mind, and hinder uchembere wabwino, or safe motherhood. This underscores the additional hardship caused by the blame for pregnancy loss, and how safe motherhood is about a lot more than survival; positive social relationships matter too.

Surviving loss: repairing a spoiled identity

Discourses of reproduction, marriage and kinship make available tainted subject positions, but women are not passive sufferers and engaged in various moral survival strategies (Erviti, Castro, and Collado 2004; Van der Sijpt 2014). When Chisomo, 32 years, told us ‘the husband still wanted me’, ‘he did not want to go’, she construed herself as a desirable wife. Others engaged in similar identity projects by countering their spouses’ indictments. Grace recounted how she told her husband ‘‘look, it just happens, I didn’t, I don’t even want to lose these pregnancies. It is just happening. There is nothing that I can do about it’. Here, Grace framed herself as an innocent victim of ‘happenings’. Similarly, Chikondi (32 years) countered her husband’s accusation that she had ‘put [taken] medicine’:

I said no, I was well well really. Things started when I went to mow grass. So I came home and then I rushed to the hospital, they examined me and said the child was okay but had moved a little from the side (…). Later I came back, I didn’t understand, so I went back. That’s when I miscarried at the hospital.

Chikondi rejects blame by declaring, with repetition and thus emphasis, that she was ‘well well’ until engaging in dutiful behaviour of cutting grass. Moreover, she rejected her husband when he changed his mind and returned to her:

I told him that, ‘aah you didn’t take care of me during the time I was in trouble. I did not want those problems. They just happened to me… So you want me again now? It’s better for me to stay alone with my poverty. I will stay alone and if I get a real man who is serious, I will get married to him’.

Here, Chikondi shifts the blame to her husband (he did not take care of her when she needed it), frames him as unworthy of her, and says she is better off alone, even in poverty. Clearly, women who are poor and had limited education can also talk back and speak out.

Narratives were replete with references to faith and trust in God’s will, usually paired with acceptance. For example, when we asked Tadala whether her loss had worried her, she responded: ‘Not at all, since it’s God, what could we do?’ Such references to God construe the loss as ‘meant to be’, and this may provide solace. Yet, invocations of religion do not merely reflect participants’ faith. Women could be upset and tearful during interviews, despite expressed belief in God. Religious accounts are also moral survival strategies that enabled our interlocutors to construct a positive identity (Erviti, Castro, and Collado 2004; Van der Sijpt 2014). As Layne (1997, 306) remarks, women referring to faith in God utilise

an authoritative discourse to transform themselves from objects of shame, pity, and scorn into women deserving respect and praise. In other words, they are dealing with the problem of a spoiled identity.

Framing oneself as accepting God’s plans attests to one’s religiosity, a valued subject position in Malawi, especially for women whose failure to adhere to societal norms of reproduction endangers their moral position.

Midwives mobilised a similar religious discourse of God’s will and acceptance. Nurse Christine, for instance, described how she tries to comfort women:

You encourage her, don’t worry, it happens. But don’t think God doesn’t love you. He still loves you … Anyway you have come for admission, what they are going to tell you in the ward just accept it… maybe they will do the induction for that foetus… But they have told you that it’s gone, just accept it.

For health practitioners, religious discourse can be a tool to console and counter women’s self-blame (Andipatin, Naidoo, and Roomaney 2019). It can also protect their moral status by diverting attention away from their own acts and omissions as potential causes of death, similar to health providers’ and women’s common statements that the baby had died kale, or ‘earlier’, in the womb (De Kok 2019).

Access to and control over discourse is a specific source of power (Van Dijk 2015). Health professionals have a privileged position: medical knowledge is authoritative knowledge (Jordan 1997). Thus, health professionals’ attributions of loss, including those that frame women as responsible (e.g. attributing loss to her incorrect pushing or disobedience (De Kok 2019)), are more likely to be accepted. ‘’

Discussion and conclusion

As in other settings in Africa, pregnancy loss in Southern Malawi is associated with stigmatisation and blame, especially of women, disturbed relationships, and profound social suffering. Reproductive loss frequently involves multiple losses: of an aspired identity (mother, worthy spouse), social status, conjugal and family relationships and economic security (de Kok, Hussein, and Jeffery 2010; Sturrock and Louw 2013). Thus, pregnancy loss can fundamentally change women’s biosocial subjectivities, that is, their understanding of themselves and their place in the world (Berry 2010).

In this study, discourse theory allowed us to explore how social, discursive processes generate loss stigma. Attributions of loss to the woman’s (sexual) behaviours or laziness, designations of the bereaved woman as a ‘useless’ wife, unworthy of gifts and support, are elements of gendered discourses that reproduce stigmatised subject positions for women that disempower them in the marriage, family, health facility and society at large.

Silence, practices and use of space are elements of discourse (Rose 2016). Not speaking about loss and grief, and minimal burials outside the graveyard co-constitute societal discourses that frame pregnancy loss as a stigmatised death and create a stigmatised, spoiled identity (Goffman 1963) for women experiencing loss. Others have critiqued limited public grief and minimalist burials for devaluing loss (Ayebare et al. 2021; Froen et al. 2011). Yet, when funerary practices minimalise a loss to that of a liminal person or even a ‘thing’, this could be considered intended support for ‘moving on’ to attain the valued ideal of motherhood. However, support was absent in hasty ANC consultations that treated loss as a tick box, and when health professionals offered no explanation or consolation following loss. Such silences in medical discourse invisibilise and neglect loss and women’s support needs, and constitute disrespectful maternity care (Bohren et al. 2019).

Societal and medical discourses lead to stigmatisation and neglect of loss and suffering, but also to the control of women, their reproduction, and response to reproductive failure. Hence, these discourses surmount to structural violence (Castro and Farmer 2005; Galtung 1975): indirect forms of violence embedded in societal structures that prevent individuals in less powerful positions from realising their potential. Structural violence includes symbolic violence, or misrecognition, in this case of loss and women’s value as human beings-more-than-mothers (to be). Spouses and in-laws appear key actors in the reproduction of structural violence; narratives regularly described their harsh condemnations and abandonment. Clearly, we must not romanticise marriage and kinship relations (Geschiere 2003).

Our analysis highlights that loss-associated suffering is social suffering. Kleinman (1997) details how suffering is social since it often involves ‘transpersonal engagement’ with misery in social relationships, and societal constructions offer cultural ‘guides’ for experiences of suffering. Furthermore, professional discourse can transform suffering into a bureaucratic category and a legitimate object for intervention or treats it with bureaucratic indifference, as is often the case with pregnancy loss.

However, women are not passive sufferers but employ moral survival strategies. Through language they repaired a ‘spoiled identity’, casting themselves as worthy and desirable wives, or their husbands as ‘not real men’. References to religious faith enabled women to displace blame and re-establish moral status (Van der Sijpt 2014; Layne 1997), rather than reflecting fatalism (Frøen et al. 2011).

Our analysis has implications for policy and practice. Stigma, blame and marginalisation will compound already complex maternal grief (Flenady et al. 2014). Since societal and medical discourse produces social suffering, programmes and interventions should be ‘multi-levelled’ (Gerrits et al. 2023; Link and Phelan 2001) and go beyond offering psycho-social support to individual women and their spouses. Acknowledging the role of discourse in structural violence has a liberatory and emancipatory potential (Logie and Gibson 2013), since discourses are produced by people and can change. When we change how we talk in communities and health facilities about loss and its causes, its impact can shift. Observed counter-narratives that position women as not to blame need amplification, for instance through radio shows, community theatre and vlogs (popular Malawian vlogger Alice Andrews (2023) already critiques the invisibilisation of loss). Peer support (e.g. through social media) might beneficial, as suggested by infertility support groups (Gerrits et al. 2023).

Interventions need contextualisation and alignment with prevailing discourses. The literature suggests that burials and ‘memory making activities’ (e.g. creating footprints, photographs) can help the grieving process (Baffour-Awua and Richter 2020; Frøen et al. 2011). However, grief theory is based on Global North research, and we should explore what kind of bereavement support works in settings such as Malawi. In a Nigerian study, Yoruba women did not wish to be involved in the burial (Popoola, Skinner, and Woods 2022). Our interlocutors did not express regret about the minimalist burials, whilst more extensive funerals may increase stigma by visibilising these kinds of ‘bad’ deaths.

Furthermore, health professionals need bereavement support training (Homer, Malata, and Ten Hoope-Bender 2016; Quenby et al. 2021; Simwaka, de Kok, and Chilemba 2014) that provides health professionals with a repertoire to support couples experiencing loss (Andipatin, Naidoo, and Roomaney 2019). As part of ante-natal care, previous loss experiences and anxieties should be discussed. Framing loss as something that ‘just naturally’ happens limits women’s space to express their grief (Andipatin, Naidoo, and Roomaney 2019), as does health professionals’ quick invocation of God’s will (Popoola, Skinner, and Woods 2022). Asking women how they are feeling and could be supported might be more helpful (Popoola, Skinner, and Woods 2022). Explaining what might, and might not, have caused loss to the woman and her spouse and relatives may protect women from blame. Yet, mobilising religious frames may be health professionals’ unconscious strategy to reduce their own accountability (De Kok 2019). Interventions must also address professionals own care needs. For them too, loss can be upsetting (Lappeman and Swartz 2019).

Miscarriages, stillbirths and neonatal deaths are not ‘non-events’ and require more attention. Our analysis enables development of theory based interventions grounded in the acknowledgement that societal and medical discourses create stigmatised subaltern subjectivities for women who encounter loss. These discourses must change. This paper is a modest contribution to that.

Acknowledgements

We are deeply grateful to the research participants. We also thank Address Malata who has provided invaluable advice from the start of the project, as well as Patricia Jeffery, Claire Wendland, Alfred Maluwa and Olive Goba. We thank Steven Russell for his editorial assistance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Bregje de Kok received an Early Career Fellowship from the Independent Social Research Foundation for this research.

Notes

1 In this paper, we present an analysis of a sub-set of our data: as part of a larger study, we also interviewed spouses, relatives and traditional birth attendants.

2 ↑ signifies heightened pitch. (…) indicates that some text has been omitted. [text] is additional text, added when needed for clarity.