Introduction

Common perinatal mental disorders (CPMDs; depression, anxiety and other disabling distress states occurring in pregnancy and the postnatal period) are a significant health concern globally. A meta-analysis of quantitative studies found that the prevalence of CPMDs is higher in low- and low-middle-income countries than in high-income countries (Fisher et al. 2012). In part, this may be explained by greater exposure of women in low-income settings to health and psychosocial risk factors including obstetric complications, HIV and other infectious diseases, malnutrition, intimate partner violence, poverty and gender inequality (Fisher et al. 2012; Sawyer et al. 2010). CPMDs in low- and middle-income countries (LMIC) have been found to be associated with poor birth outcomes and impaired infant health, growth and development (Surkan et al. 2011), although evidence for this from studies conducted in sub-Saharan Africa is mixed (Adewuya et al. 2008; Bindt et al. 2013; Guo et al. 2013; Hanlon et al. 2009a; Harpham 2005; Medhin et al. 2010; Ndokera and MacArthur 2011; Ross et al. 2011; Stewart et al. 2010; Tomlinson et al. 2006).

Only a limited number of studies have used qualitative methods to explore perinatal mental health in LMICs and in sub-Saharan Africa specifically. Qualitative methodologies are appropriate for exploring local understanding of perinatal mental health problems, the lived experience of women in the perinatal period and the social context of pregnancy and childbirth. In studies from sub-Saharan Africa, mental health problems in both pregnancy and the postnatal period were recognized and tended to be viewed as understandable reactions to stress rather than biomedical illness (Bass et al. 2008; Hanlon et al. 2009b; Rochat et al. 2011). Hanlon et al.’s (2010) study in rural Ethiopia found that pregnancy was seen as a time of vulnerability to natural and supernatural threat and that it could be a burden in conditions of poverty and low support. Furthermore, pregnancy was seen as a source of shame if occurring out of wedlock or soon after a previous delivery. Postnatally, distress states were understood as a reaction to interrelated stressors including poverty, lack of support and unmet expectations regarding traditional perinatal practices (Hanlon et al. 2009b). In Zambia, Mwape et al. (2012) identified fear and worry as key themes especially in relation to the risk of death and complications in childbirth, and HIV. Lack of social support or the strain of having an unsupportive, unfaithful or abusive partner also had important implications for maternal mental wellbeing. Similar stressors were identified amongst women in The Gambia (Sawyer et al. 2011) where the risk of being left isolated and unsupported was seen as being heightened by societal expectations that childbearing and childcare are the exclusive domain of women. In Tanzania, Kaaya et al. (2010) identified antenatal distress states occurring in the context of the “problematic pregnancy” marked by medical, supernatural, economic and interpersonal stressors.

No qualitative studies exploring perinatal mental health and associated stressors have been conducted in Malawi. In a quantitative study, we found that antenatal depression was associated with experience of intimate partner violence and lack of close confiding relationship (Stewart et al. 2013). In a study of women attending a child health clinic with an infant due for measles vaccination, factors associated with increased CPMD symptomatology were lower socioeconomic status, lack of a confiding relationship with partner or relative, recent infant illness and HIV infection (Stewart et al. 2010).

The purpose of this study was to use qualitative methods to expand and enrich the evidence from these quantitative studies and to inform the development of a psychosocial intervention for CPMDs in Malawi. Specifically, we aimed to explore local perceptions of the stressors experienced by women in the perinatal period and to investigate how perinatal mental health problems are recognized, described and understood.

Method

Ethical approval for the study was given by the College of Medicine Research Ethics Committee, Malawi (P.06/11/1089). Data collection took place from October 2011 to May 2012.

Setting

The study was conducted in Mangochi district which lies at the southern end of Lake Malawi. The district population is approximately 800,000, 90 % of whom live in rural areas (National Statistical Office, ICF Macro 2011). The largest cultural group in the district is the Yao. The Yao are predominantly Muslim, speak Chiyao as their first language and practise polygamy. Sixteen percent of men in the district have more than one wife (National Statistical Office, ICF Macro 2011). The remaining population is mainly Chichewa speaking, monogamous and Christian. The main occupations in the district are subsistence farming, fishing, tourism and small enterprise. The literacy rate for men is 74 % and for women is 52 % (National Statistical Office, ICF Macro 2011). Antenatal and child health services are provided free at government and NGO-run clinics and a district hospital in Mangochi town. In 2011, Malawi was estimated to have an adjusted maternal mortality ratio of 460/100,000 live births, infant mortality rate of 53/1,000 live births and adult HIV infection prevalence of 10.8 % (UNICEF 2013).

Sample

The focus group discussion (FGD) participants were parous women living in villages in Mangochi District (Namwera, Lungwena, Kafucheche, Nsanyira, Namakango). Purposive sampling was used. Participants were identified by the interviewing team or by field workers working on a study of a mother-infant nutritional intervention in the study area. For each FGD, field workers were asked to recruit mothers from different villages under a group village headship. A 50 % over recruitment of participants was recommended. Before the interviews, the interviewers screened the participants to make sure that no closely related women were in the same group. The FGDs conducted with women were a subgroup of wider data collection that included FGDs with men and key informant interviews conducted with traditional birth attendants and nurse-midwives. To allow detailed presentation of the women’s perspective, in this paper we present data from the women’s FGDs only. FGDs were conducted in a variety of settings chosen to reduce the risk of interruption or the conversation being overheard, e.g. a classroom, within a fenced compound, in a health centre office, at a designated place for community meetings (under a big tree).

Protocol

At the outset of each FGD, the purpose of the research was explained. Each participant was invited to read the information sheet and consent form. These were read out to illiterate participants. Written consent was obtained where possible. If a participant was consenting but unable to write, verbal consent was documented by the interviewer. The interviews were recorded on digital audio recorders. The interviewers were a male research psychologist (EU) with significant experience in conducting qualitative research (including in Mangochi District) and two female psychology graduates trained by EU. All interviewers were fluent in Chiyao, Chichewa and English. Interviews were conducted in the interviewees’ language of choice, with some FGDs conducted in a mixture of Chichewa and Chiyao. The topic guide was adapted in response to data collected following discussion amongst the research team. It consisted of questions concerning the following areas: the thoughts and feelings experienced by women during pregnancy and the postnatal period, the role of traditional and Western health beliefs and practices, how a woman might expect to be supported, causes of worry and distress, how exposure to these stressors might affect women’s health and wellbeing, and whether mental disorder in the perinatal period is recognized and how it is understood. The final topic guide is included as electronic supplementary material.

Data analysis

Interviews were simultaneously transcribed and translated into English, by the interviewer or a colleague from the research team. EU verified the translations by comparing the vernacular and English transcripts. Special attention was given to terms or concepts that did not have an immediate English equivalent. Any unusual or problematic expressions used were reviewed and discussed. Preliminary data analysis was conducted as data collection progressed and discussed by the field study team. In subsequent FGDs, the same broad topic themes were kept but additional probes were used to clarify certain emerging information. After completion of data collection, thematic content analysis was conducted aided by NVivo computer software (QSR International Pty Ltd 2012). One author (RS) coded the data and identified themes and subthemes. Other authors (KB and EU) reviewed the coding, and all authors discussed and agreed the final themes.

Results

We conducted 11 FGDs with a total of 98 parous women. Characteristics of participants are shown in Table 1.

Table 1 Characteristics of participants in focus group discussions (FGDs)
Full size table

Three major themes were identified: pregnancy as a time of uncertainty, the husband (and others) as support and stressor, and the impact of stressors on mental health.

Pregnancy as a time of uncertainty

For the participants, falling pregnant was generally seen as a positive event brought about by God’s will, and the expected child was referred to as a “gift”. However, pregnancy was also seen as a time of uncertainty both about the outcome of the pregnancy and whether the husband would be faithful and supportive. Uncertainty about the pregnancy focused on the risk of pregnancy-related illness, of miscarriage and of complications and death in childbirth. These concerns were seen to be heightened for women who had previous complications and for first-time mothers.

“You have worries, since they say you are “between” (pakati) that means you are between life and death. When it’s the day of delivery you think of those who died while giving birth, like: that woman died while giving birth, so will I make it through, will I not die?” (participant, Nsanyira FGD 1)

The uncertainty was exacerbated by a sense that women only had limited control over the outcome of the pregnancy. Firstly, delivery of a healthy baby and survival in childbirth were seen as being in God’s hands.

“The day of delivery you ask God for a safe delivery because He is the one who put it inside you. So, if God is with you, the baby is born normally without any problem but if God wishes to trouble you, He does so and that’s when you go to Mangochi hospital even for operation.” (participant, Namwera FGD 1)

Secondly, although there was a recognition that there are things a woman could do to increase the likelihood of a positive pregnancy outcome—to eat well, not to overwork, to be in a relationship that is “at peace”, to be prepared for the delivery and to attend antenatal clinic—she was highly reliant on her husband for these things. Traditional beliefs regarding the avoidance of certain foods or behaviours that could lead to the baby being harmed were acknowledged, but there was an awareness of a cultural shift away from traditional practices and sources of support to a greater reliance on the hospital, although there were concerns about the costs of transport and some women described having been treated rudely at the hospital.

“In our village we don’t allow people to go to the traditional doctor but they should go to the hospital, we should leave the old beliefs” (participant, Namwera FGD 2)

HIV testing at antenatal clinic was seen as a cause of anxiety. A woman who found she was HIV-infected would fear for her own life and for that of her child. She may also face rejection and violence if she disclosed her status to her husband.

“Others when they go to the hospital they are told that they are positive…when they go home to tell the husband his response is not good, he even tells you that he doesn’t want to be with you anymore” (participant, Namwera FGD 3)

HIV-infected women continued to be stigmatized, although some participants thought that attitudes toward HIV were changing as antiretroviral (ARV) medication became more readily available.

“For those who run away from testing for HIV, they do so because friends around you scream at you saying, “She has HIV”. But nowadays things have changed and being HIV positive is not something new and you should not be running away” (participant, Namwera FGD 1)

Witchcraft was seen as a very real danger to a mother and her baby. Witchcraft could be used to cause delay and complications in labour, make a foetus disappear or cause a woman to give birth to a cloth or stone instead of a baby. Witchcraft was usually used by somebody with whom one had argued and so it was important to avoid conflict when pregnant. Postnatally, there were concerns that the child might either be harmed by, or tutored in, witchcraft.

“When you are pregnant you shouldn’t argue with other people because during that time you carry two souls and you don’t know what the future might bring. So if you have quarreled with someone, the person can curse you” (participant, Namwera FGD 1)

“They are teaching people’s children…taking them at night to school, a night nursery school. The children are taught witchcraft and they are told to do something to their mother” (participant, Nsanyira FGD 1)

Whereas falling pregnant was regarded as bringing with it inherent worry, giving birth to a healthy child was seen as a time of great relief and usually of celebration for the mother, her partner and family. Being a mother is a source of pride, and the relationship between a mother and her baby is a warm and positive experience, even in some difficult circumstances.

“I become happy when the baby is born even if the husband should have an affair somewhere, that I have a baby and I did not die. I become happy, he can continue what he is doing, at least I have a baby” (participant, Namakango FGD)

Husband (and others) as support and stressor

Whether or not a woman is able to experience the positive conditions that increase her likelihood of having a successful delivery and raise a healthy child was understood as being highly dependent on the support she receives from her husband, and to a lesser extent from family and friends.

Participants said that it was hoped that a husband would show increased love and affection during pregnancy as this would show commitment to the relationship and the child. The husband is also expected to provide financial and practical support to his wife during the perinatal period. He is expected to provide her with food to meet her increased nutritional needs and to provide money for a basin, razors and plastic sheets for delivery and “chitenjes” (traditional cloths), blankets and soap for the new baby. He is also expected to arrange transport to the hospital for delivery.

“Sometimes after the husband gets you pregnant you want many things that the husband should help you, you should dress nicely, eat nice food, and you should get some love from your family. Both you and your husband should love one another” (participant, Namwera FGD 4)

It was recognized that in conditions of poverty, a husband might be unable to provide despite wishing to.

“…possibly there are others who are poor, the husband does not even cheat but he just does not have a means to help the family. Up to the point you know that the husband is trying hard.” (participant, Lungwena FGD)

However, it was felt that men often willfully failed to live up to their responsibilities. The perinatal period was seen as a time when a husband was more likely to be unfaithful, take a second wife or abandon his wife. If unfaithful, the husband might infect his wife with sexually transmitted diseases (STDs) including HIV.

“Sometimes the husband goes outside and sleeps with a woman who has STDs then he comes home and gives it to the wife so when it is time for delivery it is hard since he gave the wife some disease” (participant, Nsanyira FGD 1)

Because of the financial dependence of women on men, a husband’s failure to support his wife may be a serious threat to her health and that of their child. In extreme cases, men might actively withhold food and money, be physically and emotionally abusive or force the woman to have sex against her will.

“The abuse which you face when you are pregnant is that the husband sometimes beats you and the end result you deliver a dead baby” (participant, Namwera FGD 2)

Infidelity and abandonment in the perinatal period were seen to be, in part, a consequence of the taboo against a couple having sex in late pregnancy and for up to 6 months postpartum. Women felt that they were in a bind in that, if they had sex during the prohibited time, they might harm the baby or become pregnant again, but, if they did not, this would increase the chances that the husband would be unfaithful.

“The worries that happen for women like us, when the baby is born, we are supposed to stay six months without having sex with our husbands, then the husbands goes to other women without leaving food in the house” (participant, Nsanyira FGD 2)

Falling pregnant again too soon after having a child or when a mother felt unable to adequately care for her existing children was regarded as a significant worry.

“When the baby is still young to have a sibling and you a mother becomes pregnant, and that you did not follow any family planning method, the baby suffers, she doesn’t grow healthy…she has worries that my mother has stopped me from breastfeeding before the right time.” (participant, Namwera FGD 2)

Women recognized the problems associated with their financial dependence on men and saw the benefit of having small business to give some financial independence.

“In today’s marriages, if you depend a lot on the husband you do have worries, so you should do small businesses” (participant, Namwera FGD 2)

The practical support of family was seen as important, particularly if the relationship with the husband is poor. Family members were expected to bring gifts and help with chores if possible. A woman who was abandoned or abused might also return to her parental home. Being able to confide in a close friend was also seen as helpful.

“When you have a good friend you sometimes call her and sit with her “my friend, I have this story that is disturbing me, what should I do so that me and my husband should agree on things?”” (participant, Namwera FGD 2)

Relations with wider friends and acquaintances, however, could be problematic. A mother who was unable to prepare materially for the delivery and new baby, who was diagnosed HIV-positive or who became pregnant again soon after delivery might be gossiped about or ridiculed by others in the village.

“Also when you are pregnant, and you get pregnant again before the other baby walks, not even crawling, so wherever you pass women, there is a lot of gossip there about you” (participant, Nsanyira FGD 2)

For a woman who is struggling, the stress of poverty or a poor relationship might be exacerbated by envy of those who were seen to be better off. Comparing oneself to others was seen as a cause of despondency.

“Another problem is if the husband is jobless you do not have food, even cloths for you to wear when going to hospital. You worry because you envy your friends because they are dressing nicely” (participant, Namwera 2 FGD)

Impact of stressors on mental health

Participants described how exposure to stressors in the perinatal period could result in a condition of “nganisyo” (thinking too much). Somebody who was thinking too much would have characteristic thoughts and feelings and be recognizable because of their altered behaviour (Table 2). It could lead to severe outcomes including “refusing the baby”, more severe mental disturbance and suicide.

Table 2 Symptoms of “nganisyo” (thinking too much) described by participants
Full size table

“Sometimes you go crazy because of thinking” (participant, Kafucheche FGD 1)

“Sometimes you commit suicide because of thoughts” (participant, Kafucheche FGD 1)

Severe mental disturbance associated with stress in the perinatal period was termed “manunu” in Chiyao. There was some variability in how this term was used; some differentiated between manunu and masoka/misala (madness) but not all. Key features of manunu were talking to oneself, collapsing or fainting, poor concentration and giving answers contrary to the question. Manunu was almost always related to worries about husband’s behaviour, although learning that one is HIV infected was also described as a trigger.

“When walking the person speaks alone, and clapping her hands like the manunu has begun….your head becomes crazy but not madness that you were born with but the one that has come because of a man, home, friends but mainly it is because of the husband, so when walking and the car is coming you do not hear anything until the car hits you.” (participant, Namakango FGD)

Participants of one FGD also described another condition resulting from a husband’s behaviour, namely being affected by “majini” (Jinns, supernatural spirits). In this case, the condition was understood as having a function of communicating distress about the husband’s behaviour.

Really, the disease has got a name in Chiyao and it is called majini. If women are worried, they deliberately create them in such a way that they scream so that people should hear and gather in order to hear your murmurings and when the husband comes, you add some words but he denies.” (participant, Namwera FGD 1)

Discussion

This was the first qualitative study to investigate women’s experience of perinatal mental health problems, stressors and protective factors in Malawi. We found that there is a recognition that women in the perinatal period may experience depressive and anxious states, some with severe and disabling symptoms, in response to a range of psychosocial and health stressors. The results of this study support the validity of the methodology and the findings of quantitative studies in Malawi (Stewart et al. 2013, 2010) and are consistent with similar studies from other countries in sub-Saharan Africa (Hanlon et al. 2010; Kaaya et al. 2010; Mwape et al. 2012; Sawyer et al. 2011).

A strength of the study was the range of participants included in the FGDs; there was a mixture of Chiyao and Chichewa speakers, Muslim and Christian believers, and rural and semi-urban residents. Interviewers were fluent in both Chichewa and Chiyao, allowing respondents to express themselves in whichever language they felt most comfortable. The women’s FGDs were a subset of the data collected that also included FGDs with men and key informant interviews with nurse-midwives and traditional birth attendants. This allowed us to explore the experience of the perinatal period from different viewpoints and to triangulate the data. Although not presented here, these other sources largely confirmed the findings in this paper.

There were a number of limitations. Transcription and the translation into English of the interviews were done simultaneously rather than in a two-stage process; this increases the risk on inaccuracies or mistranslation. However, a sample of the transcripts was checked by one of the study team (EU), and any unusual or problematic expressions used were reviewed and discussed. A proportion of the FGDs were conducted by a male interviewer. This may have lead to a reluctance of participants to discuss gender-sensitive issues, although there was no evidence that the responses differed between interviews conducted by male or female interviewer. Care should be taken in generalizing the results to other areas of Malawi that differ in language, religion and rural/urban distribution.

Pregnancy as a time of uncertainty

Amongst the women in this study, pregnancy was seen to place a woman into a state of uncertainty about her own survival and that of her unborn child; this finding is similar to qualitative studies from other sub-Saharan Africa countries, all of which have high maternal mortality rates (Hanlon et al. 2010; Mwape et al. 2012; Sawyer et al. 2011). In Malawi, the word used for pregnancy is “pakati” meaning “between”, and several participants referred to pregnancy as a time when a woman is “between life and death”, a phrase also used by women in rural Ethiopia (Hanlon et al. 2010). Malawi has seen a significant decline in maternal mortality over recent years, although it remains very high compared with high-income countries; lifetime risk of mortality in childbirth in Malawi was estimated at 1:36 in 2010 (UNICEF 2013). In part, this reduction in mortality has been attributed to a strong government drive to promote facility-based delivery in place of home birth (Bowie and Geubbels 2013). In this study, participants reported that hospital care and advice have begun to take over from traditional practices and beliefs as the preferred basis for ensuring a successful pregnancy and birth. However, as was found in a qualitative study of cultural practices in Malawi (Malawi Human Rights Commission 2005), this is a fluid and ongoing transition in health beliefs and practice, which itself may add to the uncertainties of the perinatal period.

As was found in Zambia (Mwape et al. 2012), HIV infection remains stigmatized and a major source of worry to women in the perinatal period, although the increasing availability of ARV medication was noted by some participants as reducing the fear of infection. A reduction in the salience of HIV as a stressor in Malawi was also suggested by our earlier findings that HIV was associated with CPMD symptoms in women with young children in a study conducted in 2006 (Stewart et al. 2010) but was not amongst antenatal women in 2011/12, 98 % of whom reported being on ARV medication (Stewart et al. 2013). In the current study, it is possible that there was some “social desirability bias”, with participants endorsing government health initiatives and emphasizing their rejection of traditional practices, knowing that the study team was from the University of Malawi.

In addition to health risks, participants described being vulnerable to harm through witchcraft. Witchcraft beliefs are strongly held in Malawi (Chilimampunga and Thindwa 2011), and similar supernatural concerns are shared across traditional cultures in sub-Saharan Africa. Vulnerability to witchcraft in pregnancy was described in studies from Mozambique (Chapman 2004), Kenya (Harkness 1987), Ethiopia (Hanlon et al. 2010) and Tanzania (Kaaya et al. 2010).

Whereas pregnancy carries inherent uncertainties, delivery of a healthy child was seen as a great relief and joy to a mother (although worries related to child health and husband behaviour may subsequently accumulate). This finding is consistent with prospective cohort studies from Ethiopia (Medhin et al. 2010) and Cote d’Ivoire (Guo et al. 2013) that found that the prevalence of CPMDs dropped between pregnancy and the postnatal period. At the time of writing, no longitudinal studies with repeat CPMD symptom measurements over the perinatal period in Malawi have been published.

Husband (and others) as support and stressor

The relationship of a woman with her husband was identified as the critical factor affecting her mental wellbeing in the perinatal period, as a good relationship would bring both emotional and economic security. In rural Malawi, there is marked gender inequality; women carry the burden of domestic work and lack financial and decision-making independence (Government of Malawi 2012). Women in the study highly valued emotional support from the husband as it demonstrates commitment to the relationship and the pregnancy. However, in the context of gender disadvantage, poverty and high maternal mortality, breakdown of a relationship not only removes the coping benefits of intimacy and shared endeavour but also increases the magnitude of the objective threat to a woman’s health and life, e.g. having insufficient food, not being able to access transport to hospital for delivery and being infected with HIV.

Interpersonal problems outside of the relationship with the husband were also identified as being associated with distress in the perinatal period, namely envy and the experience of being gossiped about. These experiences of negative social comparison have not been measured in quantitative studies of PCMDs in LMIC and warrant further investigation in this context.

Impact of stressors on mental health

The perinatal distress states recognized by participants in this study are characterized by affective, cognitive, biological and behavioural symptoms and are consistent with the concept of common perinatal mental disorder. The general term used was “nganisyo” (thinking too much). This is a term used to identify a disabling distress state at any time and not specific to the perinatal period; it is similar to the term “kufungisa” used in Zimbabwe (Patel et al. 1995) and “kusononeka” used in Tanzania (Kaaya et al. 2010). Using a mixed method approach, Bass et al. (2008) in the Democratic Republic of Congo found that a depression-like syndrome “Maladi ya Souci” was recognized in the postpartum period. In rural South Africa, women identified as having antenatal depressive disorder used psychological language to describe their own symptoms; the authors argued that this indicated that the diagnostic concept was culturally sensitive (Rochat et al. 2011).

Although there was some variation in its use, the term “manunu” is a chiyao word that appears to refer to a disordered mental state specifically associated with severe stressors in the perinatal period. The marked functional disability might be consistent with a severe depressive episode, although this study should be regarded as exploratory and further research is needed. Behavioural disturbance attributed to possession by “majini” (Jinns) was recognized by some participants as a culturally sanctioned way for a woman to seek help from the community in dealing with an abusive husband. That some participants interpreted spirit possession in this way indicates a society in transition. This transition was described by Peltzer in a study of spirit possession in Northern Malawi as a move from “the heart to the head” (Peltzer 1989). That behavioural disturbance can have a communicative function in a context of marked gender power imbalance demonstrates the importance of careful assessment of phenomenology and context when assessing a woman presenting with symptoms of CPMD in Malawi.

Conclusion

In rural Malawi, mental health problems in the perinatal period are recognized and seen as disabling states that warrant intervention. In a context of poverty and marked gender inequality, physical threats such as the failure of the husband to provide materially, HIV and unsafe childbirth are potent stressors. However, interpersonal and psychological factors including poor communication with husband, stigma, envy and witchcraft beliefs are also important. In a meta-analysis of the effectiveness of interventions for PCMDs delivered by non-specialist personnel, Clarke et al. (2013) found that preventive interventions that focused on reducing health and psychosocial risk factors were effective, but interventions that included specific psychological components, such as the Thinking Healthy programme in Pakistan (Rahman et al. 2008), had a greater effect. Given the prevalence of PCMDs in Malawi (Stewart et al. 2013, 2010), we suggest that the development and evaluation of a psychosocial intervention, informed by the results of this study, is warranted.