An escape from agony: A qualitative psychological autopsy study of women's suicide in a post-conflict Northern Uganda

Abstract

We set out to investigate suicide among women in a post-conflict context in Northern Uganda using qualitative psychological autopsy interviews. Three to five relatives and friends for each of the three suicides recruited were interviewed (N=11). Through interpretative phenomenological analysis (IPA) we found that the women all had been through traumatic experiences attributable to the protracted war/conflict between the rebel groups and Ugandan Government armed forces. Nevertheless, the decision of self-inflicted death seemed to have been due to a combination of unpleasant experiences/events that prevailed within the last 3 months prior to the suicide. These experiences are summarized in two broad themes: No control in life and No care. Changes in the traditional gender roles, men's quest for their lost masculinity, and women's attempt to fight for their rights that was perceived as a cultural transgression contributed to the women's suicides.

In Africa, like the rest of the world, there are few analytical studies on suicide which focus on men and/or women separately (e.g. Adinkrah, 2012; Kizza, Knizek, Kinyanda, & Hjelmeland, in press; Ndosi, 2009). Furthermore, female suicidality compared to male is under-researched (Beautris, 2006), particularly in sub-Saharan Africa (Lester, 2008). Yet there are known gender differences in suicidal behavior (Edwards & Holden, 2001), not only in terms of epidemiology but also in terms of the reasons behind the suicide that require gendered suicide prevention strategies (Canetto, 2008; Vijayakumar, John, Pirkis, & Whiteford, 2005). Whereas mental disorders are considered important risk factors of suicidal behavior in high-income countries, external stressors are purported to be the most significant in low and middle income countries (Marecek, 1998; Vijayakumar, Nagaraj, Pirkis, & Whiteford, 2005). In low-income countries, men's suicide has been found to be mainly in response to changed masculinity; that is, the inability to live up to the gendered social expectations as providers and being in control of the home (Adinkrah, 2012; Kizza et al., in press). Women's suicidality on the other hand is often precipitated by interpersonal conflict between spouses and family members (Canetto, 2008; Meng, 2002; Ndosi, 2009; Vijayakumar, John, et al., 2005). According to Ndosi (2009), fatal poisoning by women is “immersed in worrisome and painful suffering, which often results from extreme tensions that have built up within interpersonal relationships or families” (p. 160).

Female suicidal behavior tends to be disregarded (Beautris, 2006) because of women's lower rates of suicide relative to men's (Canetto, 2009). However, if we consider morbidity and mortality together, girls and women have the greatest burden of suicidality (Canetto, 2009). Although suicide rates among women have been reported to be much lower in Africa compared to many other parts of the world (Meel, 2009; Mgaya, Kazaura, Outwater, & Kinabo, 2008), newer reports indicate that suicide among women in Africa is on the increase (Dolan, 2009; Meel, 2009). To the best of our knowledge, during the last decade, only one study has investigated suicidality among women in Africa (see Ndosi, 2009). Ndosi specifically focused on women who fatally poisoned themselves in Dar es Salaam, Tanzania.

In Uganda information from the media and from personal communication with the elders and health professionals indicated increasing incidences of suicidal behavior, especially in the war-ravaged Northern Uganda. Thus we set out to explore psychosocial circumstances that surrounded the individual suicides using qualitative psychological autopsy method. For the 20 cases (that included men and women) recruited for the study, we discovered during the data analysis that the dynamics underlying the suicidal acts of men and women differed considerably. Given that the study is qualitative, with an emphasis on in-depth understanding of the individual suicides, it was imperative to treat the suicides for men (see Kizza et al., in press) and women separately. Therefore this article is dedicated to understand the dynamics that underlay women's suicide in Northern Uganda. The article specifically aims at understanding the psychosocial circumstances contributing to suicide and the suicidal process of women in a war-torn Northern Uganda. Apart from contributing to the general body of knowledge in the field of suicidology, the findings will help understand female suicidal behavior in Northern Uganda and to pave way for cultural and gender sensitive suicide prevention strategies in this region.

Overview of the study context

This study was conducted among the Acholi people, a sub-ethnic group of Luo/Nilotics found in Northern Uganda in the Districts of Amuru and Gulu. These two districts are part of the five districts that form the Acholiland. The Acholi is the seventh largest group in Uganda with a population of 4.8 million people (Uganda Bureau of Statistics, 2002). They speak Luo as their native language. This region was for two decades (1986–2006) severely affected by civil conflict between Ugandan Government armed forces and rebel groups (Harlacher, Okot, Abonyo, Balthazard, & Artkinson, 2006). This period was characterized by horrifying, ruthless, and dehumanizing acts, and at the peak of insurgency in 1997, approximately 2 million people in rural areas were forced to live in “protected villages”, the Internally Displaced Persons’ (IDPs) camps (Roberts, Ocaka, Browne, Oyok, & Sondrop, 2008). These camps were densely populated and in dreadful and precarious conditions. People were mainly dependent on humanitarian aid for survival because arable land was inaccessible (Harlacher et al., 2006; Liebling-Kalifani et al. 2008). They lived in constant fear of rebel attacks and abduction and uncertainty of what was to happen next. Life was associated with lack of freedom, poor living conditions, and fear of contracting HIV/AIDS through sexual assaults and other inhuman behaviors (Kizza et al., in press). This state of affairs created a state of desperation, depression, severe alcoholism, and hopelessness which in turn was associated with increasing rates of suicide and suicidal behavior in the population (Dolan, 2009; Kinyanda et al., 2009; Ovuga & Boardman, 2009). During this period there were increasing reports of incidences of suicide (Ochola, 2006), with two suicides per week reported in a given camp (Rodriquez, 2004).

The conflict led to a breakdown of cultural values1 (El-Bushra & Sahl, 2005) and people were living in a permanent state of trauma and depression (Ovuga, Oyok, & Moro, 2008; Roberts et al., 2008). There were limited opportunities for income, especially for men. As a result they were grossly disempowered due to restricted access to the factors of production, i.e., land and cattle; the source of their wealth and power (El-Bushra & Sahl, 2005). Consequently, most men in the camps resorted to excessive consumption of alcohol, gambling, and womanizing (Kizza et al., in press; Liu Institute for Global Issues, Gulu District NGO Forum, & Ker Kwaro Acholi, 2005). Women assumed a greater number of responsibilities in the camp and became the breadwinners, which is normally a man's responsibility. Brewing alcohol was one of the women's sources of income (Olaa, 2001). The long-lasting conflict that changed the psychosocial environment dramatically, and also changed the traditional balance of power relationship between men and women considerably (Dolan, 2009; El-Bushra & Sahl, 2005).

Method

Study design

The aim of this project was to understand the dynamics and circumstances that surrounded the suicidal death of the study subjects. Thus, a qualitative psychological autopsy method was envisaged appropriate for this study. Qualitative research facilitates understanding of people's experiences of particular conditions (Kvale & Brinkmann, 2009; Willig, 2008), whereas a psychological autopsy approach is a major tool in understanding suicide (Pompili, 2011). A psychological autopsy approach helps to reconstruct an individual's life style, thoughts, feelings, personal relationships, personality traits, and behavior manifested prior the suicide (Gavin & Rogers, 2006; Murthy, 2010). In other words it helps reconstruct the lived experiences and the circumstances around the suicide (Murthy, 2010; Pouliot & De Leo, 2006). Thus, a qualitative psychological autopsy approach is capable of yielding rich and complex data about the lives of suicidal individuals (Owens, Lambert, Lloyds, & Donovan, 2008). Basically, data was gathered by interviewing significant others close to the deceased. For valid and reliable data it is recommended that a number of informants are interviewed per given case. However, other sources could also be utilized (e.g., evidence presented at inquest, medical reports, police records, and suicide notes) to triangulate and enrich the information (Batt, Belliver, Delatte, & Spreux-Varoquaux, 2004). Since the study subjects were community samples, we opted for in-depth interviews with relatives and friends that were close to the deceased; no other reliable sources of information were available.

Sampling

In Uganda, it is a criminal offence to attempt suicide (Penal code Act cap.120 under Chapter XX—“Offenses related to murder and suicide”), thus all suicidal deaths are supposed to be reported to the police, although on several occasions this is not done because of fear of stigmatization. In addition, all deaths should be reported to local authorities of the area who gives a clearance letter to bury the body. These sources were used to locate the suicide; one suicide was identified through police records and the two young women through the local leaders of the camps of residence. Furthermore, it was the local leaders who acted as our scouts in locating the families or the initial relative/friend that were recruited as our informants. Prior to the interview the local leaders together with the study coordinator, who also doubled as the interpreter, approached the bereaved. The coordinator then informed them about our study and requested them to participate. The individuals were clearly informed that it was a free choice; there was no coercion whatsoever (more details given under the section on ethical considerations). An appointment was then made for the interviewer to go and conduct the interviews. Other informants within a particular family were located through a snowballing technique whereby one interviewee could lead us to another.

Study subjects and informants

Study subjects

For the whole study project 20 suicides were identified, 17 men and 3 women. As mentioned earlier, the three women (two young adults and an elderly lady) are the focus of this article. Two of the women died by organophosphate/pesticide poisoning and one took a drug overdose. Two of the suicides took place in the house whereas one was away from home in the bush.

Informants

For each of the cases, three to five relatives and friends were interviewed (N=11) and these included: one spouse, three siblings, one child, four in-laws, and two close friends. The inclusion criteria were that only those aged 18 years and above, had been close to the deceased, and had consented were interviewed.

Interviews

The first author, a Ugandan clinical psychologist, conducted all the interviews. Because the interviewer is from a different ethnic background and none of the informants could speak the national language (English), the interviews had to be conducted through an interpreter. Potential interpreters (a male and a female) were first interviewed on their attitudes regarding the topic of the study to make sure that someone with a negative attitude towards people bereaved by suicide was not recruited. The two interpreters selected then underwent some training about conducting the interviews and their role during the data collection process. They were both mental health workers and therefore already accustomed to doing therapeutic interviews which are not so different from qualitative research interviews (Batt et al., 2004).

All informants were individually interviewed. All the interviews were held in the informants’ homes within or outside (for those that had already relocated back to their original home) the camp. The interviews were composed of both narrative and problem focused elements. The narrative part was based on one broad question: “Can you please describe the circumstances that surrounded the suicidal death of your [relation to deceased]?” Through the problem-focused part we ensured that those areas considered important and not covered in the narratives were explored. The questions in the problem-focused section were guided by 16 themes in a psychological autopsy by Shneidman (1993): (1) Information identifying victim. (2) Details on the death. (3) Brief outline of victim's history. (4) Death history of victim's family. (5) Description of the personality and life-style of the victim. (6) Victim's typical patterns of reaction to stress, emotional upsets, and periods of disequilibrium. (7) Any recent upsets (from the last few days to the last 12 months), pressures, tensions, or anticipation of trouble. (8) Role of alcohol or drugs in: (a) overall life-style of victim, and (b) his death. (9) Nature of victim's interpersonal relationships (including those with physicians/mental health clinicians). (10) Fantasies, dreams, thoughts, premonitions, or fear of victim relating to death, accident, or suicide. (11) Changes in victim before death (of habits, hobbies, eating, sexual patterns, and other life routines). (12) Information relating to the “life side” of victim (upswings, successes, plans). (13) Assessment of intention, i.e., the role of the victim in his own demise. (14) Rating of lethality of suicide method. (15) Reaction of informants to victim's death. (16) Comments, special features, etc. The interviews took place 6–18 months after the suicide, which is in keeping with other psychological autopsy studies (Beskow, Runeson, & Åsgard, 1991; Hawton et al., 1998; Henry & Greenfield, 2009). With informed consent, each interview was tape recorded and lasted approximately 1 h.

Ethical considerations

Ethical clearance was obtained from the Regional Research Ethics Committee in Central Norway and from the Uganda National Council for Science and Technology (UNCST). A clearance letter from the UNCST was presented to the Resident District Commissioner (RDC) of Gulu District in Northern Uganda. The endorsed copies of the letter were distributed to the District Security Officer (DSO), Officer-in-Charge at Gulu Central Police Station and Local Council Officers in the camps. As mentioned above, the families of the identified suicide cases were located through the local leaders who acted as our scouts. They are trusted members of the community. Initially these local leaders together with the coordinator approached the potential informants and requested them to participate after clearly explaining to them the study objective. Freedom to refuse to be interviewed and to withdraw from the interview at any point was emphasized. Each participant initially gave a written informed consent. In addition, the interviewer continuously sought process consent for those informants who appeared distressed during the interview (Kavanaugh & Ayres, 1998). At the end of the formal interview each informant was taken through a brief debriefing session and arrangements were made to offer crisis management if required (see Kizza, Hjelmeland, Kinyanda, & Knizek, 2011). To ensure anonymity, any potentially identifying information has been removed in this paper. For example, names of the deceased in the illustrative quotes are withheld and the informants are not identified in any way with a particular deceased.

Data analysis

We used interpretative phenomenological analysis (IPA) as our guiding tool, since our interest in this study was to get an in-depth understanding of the deceased's experiences and the psychosocial circumstances prior to their suicide through the qualitative psychological autopsy interview. The assumption was that the experiences of the study subjects that culminated into suicide were intricately engrained within their social, cultural, and political world that prevailed in Northern Uganda at the time. Smith and colleagues remind us that people are physical and psychological entities, they do things in the world, they reflect on what they do, and such actions have meaningful, existential consequences (Smith, Flowers, & Larkin, 2009). Further, IPA offers detailed, nuanced analyses of particular instances of lived experiences (Smith et al., 2009). Note that IPA is “not a prescriptive approach, but just gives flexible guidelines to which a researcher can adapt depending on the research aims” (Smith & Eatough, 2007 p. 45). Thus our interest was to understand what made our study subjects end their lives. Primarily the focus of our analysis was the psychosocial circumstances that surrounded the suicide and lived experiences of the deceased which were accessed through the informants’ narratives. In this way we modified the IPA, whose focus in normal circumstances is on the interviewees themselves.

A complete English written transcription of the texts was done by the interpreter. Thereafter the first author read and re-read the transcripts of each of the cases several times. This enabled her to obtain a holistic view of the case and to note the key emerging themes that were substantiated with relevant quotes. These were later compared across the three cases. During the whole analytical process the research team comprising two indigenous and two non-indigenous researchers met frequently to review and discuss the interpretations and clustering of the themes. The validity in the research was thus based on triangulation on three levels. First, the psychological autopsy study is based on the assumption that through the interview of several bereaved, it is possible to construct a valid picture of the deceased. Second, a critical screening of the interpreters (the direct interviewers) aimed at reducing interviewer bias. Third, through the constant discussions of emerging themes and clustering of themes a triangulation at the researcher level was aimed at in order to reduce the investigator bias (Yardley, 2008).

Findings and discussion

All the three women in our sample had been very determined to end their lives, as exemplified in the following narratives: (1) “… ‘One time I am going to poison myself to death. I feel the level of my sadness is increasing. Last night we had again some misunderstanding with my husband, I feel I cannot now bear the situation anymore’ …” (a deceased quoted by a friend); (2) “… as we were trying to give her antidotes for poison, she got up and said that she wants to die because of her husband who was not supporting her …” (a brother-in-law); and (3) “… When she left here she said that she should die. She went … to get a rope so that she could commit suicide, saying that ‘why did this boy go, for I'm suffering because of him’” (a sister-in-law). These statements indicate that the deceased were determined to end their lives as a consequence of relationship problems that were going on between them and some significant other(s). At the time, they seemed to have been disappointed, saddened, or raged and frustrated by what they experienced as unfair treatment by some of those they valued. They seemed to have been going through unpleasant and intolerable conditions: “I feel I cannot bear the situation anymore”; and thus searched for a way out and suicide seemed to have been the best option available.

Earlier in their lives the deceased had been through traumatic experiences attributable to the prolonged war in their area that included witnessing and loss of significant others in massacres and being displaced from their homes; which could have made them vulnerable to suicide. However, the decision to end their lives seemed to have been a combination of experiences/events within the last 3 months or so prior to their suicide. These experiences are presented in two broad themes, namely: No control in life and No care.

No control in life

Having no control in one's life was found to be one of the key factors linked to the suicidal acts of the two young women in our study. No control in life was reflected in four subthemes: distribution of labor and power, no right to fight, change of role in marriage, and health of self and children.

Balance of labor and power

The findings revealed that changes in the traditional labor and power balance may have been one of the key factors that significantly contributed to the young women's suicide. The prolonged war in the region ushered in a lot of changes, including changes in the traditional roles of men and women. As earlier mentioned the women/wives on top of their reproductive role had to take over the functions of men/husbands of providing for the family. Thus, they worked hard to ensure the family's survival, since the men's movements outside the camps were restricted and the external food aid was limited. The informant's voices below illustrate this:

Sometimes, it was difficult, especially when you do not have enough gardens and also the food aid was not always enough, so when I had no money, she could go and do casual work in people's homes/gardens and earn money or raw food so we feed on that (husband).

And

// He gave her no care, no support at all in the work whether it's a man's work or a woman's work she was the one to do all that. [] And this used to over burden my sister (elder sister).

Despite the changed productive roles, husbands remained in control of the family's resources which had been hard earned by their wives. However, instead of spending such resources on their families, they were spending them on their leisure activities including befriending and marrying other women or drinking alcohol and coming back home to harass their wives as suggested in two quotes below:

They also had struggled over produce. You know, my parents were hard-working people, and could produce crops yearly for consumption and sale. They could sell jointly and when that issue of the new wife came, every day, my father would want to sell the produce and used the money for alcohol or take it to that new wife for feeding and it annoyed my mother so much (daughter).

In a similar tone:

Well … before the issue of the co-wife, she could complain about the husband not supporting her in anything, he could not give her money for feeding, or go with her to dig or to bring food at home. She was shouldering everything like gardening and domestic work and yet he could just come and picks the produce and sell and uses the money (friend).

Judging from the two quotes above the deceased were hurting. The action of their husbands spending the family resources irresponsibly was unbearable. The reality was that men in Northern Uganda retained the leeway to make financial decisions affecting the family despite the fact that they had no involvement in providing the income (Dolan, 2009; El-Bushra & Sahl, 2005; Otiso, 2006). However, the women in our study did not seem prepared both to take on the role of being the primary breadwinner as well as being subordinate to the husband. As a result, fights could ensue as the women tried to restrict or control their husbands’ expenditures. This is clearly expressed in the following quotes:

In the camp here we have problems of men; most of them sit in the trading center drinking alcohol together with womanizing. These men just come back home to pick the little food there is in the house to take to these women. If we try to resist, they become very violent and start fighting us (sister-in-law).

And

[Initially] before her husband brought the second wife they had a very good relationship; she could work hard and cultivate a lot of crops with good yields. Following the coming of the second wife, she could complain a lot about the husband and on many occasions could seriously fight each other (brother-in-law).

The quotes above depict hurtful feelings and the betrayal felt by the women and they seem to have been caught up in an unbearable situation where they had to take on a lot more responsibility and work without gaining any advantages in return.

Note that Acholi women, like elsewhere in Africa, are socialized to tolerate a high level of conflict in marital and family life as well as tolerance of male authority (Ndosi, 2009). However, a woman's refusal to accept her subordinate position in relation to her husband is likely to cause disputes and even violence (Conner, Ruberstein, Conwell, Seidlitz, & Caine, 2001). This seemed to have been the case with the women in this study, something probably attributable to not having been socialized into cultural submission because of lack of appropriate teachers and role models due to the prolonged unstable environment.

No right to fight

Acholi is a patriarchal society and women/wives are subordinates to men/husbands and therefore are expected to be submissive, never to challenge, or question the man's actions. This means the wife/wives cannot question infidelity and have to accept polygamy as long as it is the husband's wish to marry another woman. However, the women in our sample decided to behave to the contrary. Infidelity and husbands marrying second wives instigated a lot of bitterness and unrest in the homes of these women. In an attempt to influence change, these women could either complain bitterly or refuse to carry out their marital obligations as illustrated by a sister-in-law: “1 day, when the husband wanted bathing water, instead she replied rudely, so the husband grew very annoyed and decided to … box her.” They could also aggressively attack their rivals: “… the husband decided to go in for another woman who had children and [name] … went up to that woman's home and fought” (friend). In the Acholi culture, disputes in the home may call for elders’ intervention during family meetings, but for those regarding extra-marital affairs/infidelity or marrying another wife; the elders are unlikely to decide in favor of the wife since polygamy is culturally acceptable. This is clearly illustrated by the voice below:

… elders said “a man is free to bring as many wives as one can afford. And any woman brought in the home has no right to fight or prevent a husband from loving another.” So [Name] was warned not to fight the other woman again (friend).

This quote implies that the wife attacked the husband's lover, possibly to scare her off and therefore the elders had to intervene. However, their resolutions were in favor of the husband and his lover. Furthermore, the quote seems to suggest that the common attitude was that the wife had no reason to quarrel or fight because it is a cultural norm to have more than one wife. The elders’ concern in this respect was the woman's observance of the code of behavior based on traditional values and practices (Carlson & Mazurana, 2006). Therefore, the deceased had no choice but to accept the situation.

Change of role in marriage

In Acholi culture, children work closely with the parents of the same sex to observe and learn what is culturally expected of them in their adulthood (Liu Institute for Global Issues et al., 2005). This was not the case for our subjects, given the protracted war that constrained the Acholi culture (El-Bushra & Sahl, 2005). This two-decade war hindered the process of transmission of cultural norms from one generation to another (Liu Institute for Global Issues et al., 2005). In Acholi culture, especially in rural areas, men settle within the same homestead as their parents (Harlacher et al., 2006). So when they marry, it's the responsibility of the mother to usher in the daughter(s)-in-law and to orientate her (them) in the affairs of the new home. But in the absence of the mother, the first wife takes over that responsibility. This was the case for the young women in our study. Both of them had lost their mothers-in-law during the war. When their efforts to block the love affairs of their husbands failed, the lovers ended up becoming co-wives as exemplified by this voice: “After the meeting, they resolved that the co-wife should join the husband officially and she was brought and placed under the same roof and [Name] was ordered to look after her and her four fatherless children” (friend). And because the deceased were the acting mothers-in-law, the new wives had to share accommodation during the orientation period as narrated by a sister-in-law:

// the co-wife was brought and they were sharing the same hut, because the mother of their husband died, and she [the deceased] was the older wife [the first to be married], the younger wife had to live with her and she takes care of her as a real mother-in-law.

Although it's a cultural norm, unfortunately the deceased had to take on such a responsibility when they were still very young (late 20s and early 30s): a concern that was voiced by the informants: “… Like my sister's case, she had not yet completed a year with her husband, and was given such a huge responsibility of taking care of the adopted children and their mother, yet her husband was not supportive at all” (sister). This indicates that the deceased possibly had to give up their conjugal roles which they could not have been mentally and emotionally prepared to do after not even 1 year of marriage. In addition, the husbands were unsupportive and took this change of role for granted. In normal situations, by the time the first wife undertakes the role of a mother-in-law, she is much older and would have been in marriage for several years. Even before marriage, in the process of growing up while still with their parents, they would have been socialized into such a responsibility, but this did not happen due to the war.

The marriage life thus became overwhelmingly challenging and unbearable, so at some point the deceased opted to walk out of it. However, because of the dowry and poverty, this was impossible for them as indicated in the following voices: “… she went away, when the issue of the co-wife had just started, but the husband followed her and brought her back saying that he paid his dowry for her. Her guardians [had no option but] to give her back” (friend). A similar voice said: “I had advised her to … go back to her parents … so that the tension/sadness she was experiencing subsides …. She answered that, no! Last time she went … her parents gave her back when the husband followed her” (sister-in-law). These voices suggest that these women were already a property of their husbands and therefore the parents or guardians no longer had any rights over them unless the dowry that was paid was refunded. In the Acholi culture women cannot initiate divorce (Dolan, 2002), and when a woman goes back to her home of origin, the dowry has to be refunded (Burite, 2007). Given the high level of impoverishment, refunding the dowry was impossible. Thus, these women were trapped in a painful and emotionally taxing situation without hope for either change or control over the situation.

Health of self and children

Unfortunately the strategies of rebelling and questioning the man's authority by the deceased seemed to heighten the level of violence in the home leading to physical injuries, not only for the wife but also for the children. For example one of the deceased developed a chronic physical problem because of the constant fights with the husband as narrated by the daughter:

My father could frequently beat her, every time fighting her because of some woman he had an affair with and he insisted until he brought the woman home here. // During the fight, he tended to target her head, and … she began complaining of pus discharge from the ear associated with a lot of pain.

And in one such fight, a child sustained a fracture of the skull as narrated by this informant: “One day, she was bathing the baby and he began fighting her, the baby fell down, sustained a depression of the skull. It was taken to hospital but still has a dent on the head” (daughter). In the above quotes it is demonstrated that the deceased's life and that of their children was at the mercy of their husbands. Furthermore, in agreement with Ndosi (2009) infidelity and betrayals in relationships do not only leave women feeling abandoned, intensely jealous, and unhappy, but also worried about sexually transmitted infections. It was clear from the findings that because of the husbands’ sexual behavior, the women's health was under threat of sexually transmitted infections, particularly HIV infection. Thus, apart from not being in control of their love life, they were also not in control of their own health. The women suspected that they could have contracted HIV infection, which raised a lot of emotional pain, as depicted in the informants’ narrations below:

We talked at length with her about the problem she was facing, how the husband does not give her the money for feeding and how he has gone in for a woman suspected to have HIV. I tried to console her but all in vain. Every day she could consult different people about the [HIV] status of that woman and in the process she got many different stories, which were all pointing towards HIV. The heart of my sister got so broken … What killed my sister was the issue of the co-wife being suspected of having HIV// (sister).

A friend said:

“… she could tell me how this other woman used to move with soldiers, even her last child also was not looking to be healthy. So she had a fear that the co-wife had infected them with the current problem [HIV/AIDS] but she feared to go to hospital for the [HIV] test”.

And,

He [the husband] could pick and drop women at his own will and this annoyed her … and made [Name] to worry so much about their sero-status (sister-in-law).

The above voices suggest that the women highly suspected that they had been infected with HIV in addition to the other stresses and tension they were experiencing. We note here that according to cultural norms the husbands have the right to enjoy their sexuality as they want and they can demand (unprotected) sex from their wives and thus transfer to them the diseases they might have picked up. As pointed out by Silberschmidt, (2001) it is hard for married women in Africa to negotiate use of condoms; as condom use in Uganda is greatly influenced by culture and power relations (Kabonesa, 2011; Kagolo & Nakayima, 2011) and as a result these women could control neither their own health nor the health of their children. As Dolan (2009) noted, “many suicides in the post conflict areas occurred either as a reaction by others to exercise control over the deceased or as a failure by the person committing suicide to achieve control” (p. 210). This seemed to have been fundamental in the suicide of our study subjects, because they seemed also to be without control of their own health/life. Noteworthy is that apart from being victims of physical injuries, the children were also at risk of HIV transmission from mother to child, something that could also have been worrisome to the deceased.

No care

This was either in form of intra-generational or inter-generational feelings of abandonment, depending on the deceased's specific stage of life as elaborated further below.

Intra-generational abandonment

Intra-generational feeling of abandonment was a factor also noted among the young women's suicides. This coupled with having no control in life seemed to have contributed to their suicide. These women were young in marriage and probably had certain expectations in life as married women. However, infidelity and polygamy by the husbands led to feeling unloved and neglected as implied in these voices: “… she was not happy about the way her husband was treating her. [That] he never loved her, he loved the second wife and he was not supporting her in many ways” (brother-in-law); and, “that I don't love her, I don't give money …; … because I was not loving her anymore, I did not want to support her …” (husband). This seems to suggest that these women were experiencing deprivation of something significant in their life—“love and care”. Finding out that their husbands were involved in extra-marital affairs and even going to the extent of marrying other women, possibly signified decline in the husbands’ love and care as the following quote seems to suggest: “Umh … at the beginning there was no problem because her husband loved her so much; only when he began loving the second wife that was the beginning of all the problems …” (sister-in-law). “[Name] was staying well with her husband and later on the husband decided to bring her co-wife and she became very angry that the husband was not caring for her …” (brother-in-law). These quotes suggest that the actual family context with a co-wife angered the deceased. To them it was an indication of declined love and care/attention from their husbands, something they could not tolerate.

Inter-generational abandonment

This form of abandonment characterized the elderly woman's suicide. Old age in this particular context, makes one dependent on significant others. If significant others who are supposed to support the elderly are either absent or fail their duty, the elderly find themselves in an impossible situation as explicated by this quote: “People with children are very lucky. If my children did not die, I would also equally be well off; but now am suffering” (a deceased quoted by sister-in-law). This seems to point to a person envious of those with children and agonized by the loss of her children. All of her biological children and their father had died of a sudden death and/or under mysterious circumstances. She had then invested all her love and care in her only surviving full blood relative, possibly as a way of coping with that loss, as implied by this quote “… relied on this child to forget them [the deceased children]” (half-sister). This relative, however, decided to leave her and move on with his life without her knowledge. The departure of this person frustrated her and was too painful as indicated by this relative: “… she was very disappointed and troubled and thinking of // who left her and did not know where he was. This would hurt her …. She was very disturbed; whenever she sat alone she would cry …” (half-sister). Efforts to convince this person to come back failed and this perturbed her and made her lose hope in life. She felt that her life was not worth living anymore, given that she had lost all that she valued in life. Thus she became overwhelmed with loneliness as illustrated in this narrative: “The major cause of her death was that she was lonely and had lost all her ‘valuables’ …” (half-brother). Children are the social insurance agencies of their aging parents (Dyer, 2007; Kizza et al., in press). Having lost all her children, probably the source for her social security, the deceased could have been overcome by grief and a feeling of loneliness, and to continue living was envisioned as hopeless. Social isolation and loneliness among the elderly are known contributors to suicide (Cattell, 2000).

Human beings have an overwhelming need for acceptance and belongingness (Fiske & Fiske, 2007). According to Leary, Koch, and Hechenbleikner (2001), any perceived decline or lowered relational value in the form of rejection, neglect, isolation, or abandonment, may give rise to various emotions such as sadness, hurt feelings, loneliness, jealousy, anger, embarrassment, amongst others. From what is stated above in the quotes it is apparent that these women exhibited some of these emotional reactions; reactions that are known precursors to suicide or suicide related behaviors (Leary et al., 2001). It is noteworthy that none of these deceased was reported to have suffered from overt mental illness. However, they were said to have been visibly sad, unhappy, and very distressed/perturbed by the experiences they were undergoing. This was probably in agreement with what Pompili (2011) said that “suicide can occur with no psychiatric disorder when profound distress and psychological pain become unbearable and when suicide is seen as the perfect solution” (p. 10).

General discussion

We have shown in the findings that the women in our study were determined to end their lives following the disappointing and/or unpleasant and intolerable interpersonal relationships with those they considered significant in their lives. Such disappointments centered on their inability to have control in life and the perceived lost love and care from those they valued. We do note here that whereas for the young women it was both no control in life and no care that could have been responsible for the development and progression to suicidality, only lost love and care seemed fundamental to the old woman. Furthermore, the analysis pointed to hurt feelings, anger, fear, and feelings of betrayal as emotions underlying the lost control in life, and feelings of rejection, neglect, isolation, loneliness, jealousy, or abandonment being associated with the lost love and care. We contend that power struggle between husbands and wives brought about by the changes in gender roles following the protracted armed conflict/war, was crucial in the young women's perception of lost control in life in Northern Uganda. As noted in our earlier article (Kizza et al., in press), this conflict led most men, especially in the rural area, to feel disempowered and disenfranchised after having lost their wealth and power; that is, their indicator of masculinity had been significantly transformed. In contrast, women became empowered financially and continued to produce food for the family. Some men decided to kill themselves as they could not stand the humiliation of losing their masculinity, whereas others ‘drowned’ their sorrows in alcohol. Because of alcohol they became aggressive and impulsive towards their spouses (Kizza, Hjelmeland, Kinyanda, & Knizek, 2012). According to Silberschmidt (2001), disempowered men tend to use aggressive means to exert control over their wives. Further, Crocker, Lee, and Park (2004) contend that some men in search of their self-esteem may resort to aggression towards their spouse. Through aggression they try to maintain closeness and to control their wives’ behavior. Such circumstances were evident in the suicidal process of the young women in our study. Men's drinking in the camps was also noted by Dolan (2009) to have been a factor in women's attempted suicide and suicidal death.

As earlier mentioned, Acholi society is deeply rooted in a cultural context where women have to adhere to strict codes of behavior based on traditional, patriarchal values and practices (Carlson & Mazurana, 2006). The women have no right to love and no right to complain about the man's behavior in this regard, because a woman is more or less a property of the man. Any attempt to violate such codes is viewed as threatening the patriarchal and traditional power relationship, not only within the household but also within the clan and the Acholi culture as a whole. It is by these means that many women are pushed into silent defeat and helplessness, which makes them vulnerable to unbearable suffering (Ndosi, 2009). Nonetheless some women may decide to transgress from the cultural norm in an effort to gain control of their lives as the women in our study seemed to have done. It is interesting to note that the few studies that have attempted to understand suicide among women, particularly in low-income countries, have tended to regard suicide among young women as an act of revenge towards their spouse, in-laws, or family members (Marecek, 1998; Meng, 2002). But this was not the case with the women in our study much as the circumstances surrounding their suicide rotated around severed spousal relationships. Their goal of suicide was to escape the continued suffering after failing in their endeavor to gain control over their lives and to win back the lost love and care. Normally early in life, women are socialized or indoctrinated into accepting the subordinate position and to be submissive to their husbands (Liu Institute for Global Issues et al., 2005). However, the women who were born or matured in the era of the armed conflict in the region missed out on this socialization process. In addition, soon after the war, empowerment campaigns about women's rights were embarked upon by the Government and Non-Government Organizations (NGOs; Carlson & Mazurana, 2006). This way these women could have gained courage to challenge such cultural norms, thus conflicting with their husbands and elders.

As for the elderly, the primary goal to continue living is the continuity of the lineage and having children is the very reason for survival. This is very often maintained through positive trans-generational interpersonal relationship. According to Conwell and Thampson (2008) the elderly people who die by suicide often have complex problems among which are functional decline, social isolation, and age-related existential challenges, like being the only surviving individual. Unfortunately because of the prolonged war and the associated poor living conditions, some old people in Northern Uganda found themselves as the only surviving individual in the family (Kizza et al., in press). This seemed to be the case with the old woman in our sample. A situation that depicted defeat, no escape because of the unavailable (or perceived lack of) social support and thus she was trapped within a terrible past and blind future with no social insurance. Suicide was therefore seen as the only available alternative to their continued suffering.

The above findings are best explained within the assumptions of the “cry of pain model”, an entrapment model of suicidality (Williams, 2001; Williams & Pollock, 2001). This model asserts that suicidal behavior is a reaction or a response (“the cry”) to a situational stressor that depicts defeat, inescapability, and no possibility of being rescued (Williams, 2001; Williams & Pollock, 2001). This mediated by lack of social support and availability of lethal means, suicide is then the likely outcome. In the findings above we have exemplified all those stressors among our study subjects that depicted defeat and inescapability. Other than the frustrations illuminated above, the women lacked the social support to lean on. For the young women the unquestionable support available would probably have been from their parents/guardians. Unfortunately the only way these relatives could help them out, was to support the couples to divorce and come back home, but it meant a dowry refund. However, as explained above, this was impossible because of the high levels of impoverishment. Besides, the means for self-inflicted death were readily available, despite the restrictions of the selling of organophosphates to women. According to the informants and elders, the sale of organophosphate pesticide to women was prohibited because attempted suicide among young women had become almost an epidemic (Kizza et al., in press). Nevertheless, if one was determined she would be able to access it by any number of means, and for the one who opted to die by drug overdose the drugs were available over the counter.

It is noteworthy that there are gender and age specific similarities and differences in the factors relating to suicide in this war-torn area. Women and men in their late adulthood are noted to be similar as regards lineage continuity, lack of social insurance and/or age-related existential challenges in fostering their suicidal behavior (Kizza et al., in press). These were noted to have aroused feelings of social isolation and loneliness in both women and men. For the young adults, the key issue for young men's suicidality was living up to their expectations, which they could not (Kizza et al., in press) and for the young women it was an issue of not wanting to live up to social expectations.

Strengths and limitations

We acknowledge that the sample in this study is small. However, according to the informants and elders in the community, the cases discussed here are typical of a number of other female suicides as well as attempted suicides in this region in the study period. We therefore believe these findings are of significance in understanding suicide among women in a post-conflict Northern Uganda and possibly those in similar contexts.

We also acknowledge the inherent limitations of psychological autopsy studies, like reporting biases by the informants and problems of memory recall. However, we dealt with them by triangulating at different levels as elaborated in the method section. Another important limitation was the use of interpreters in the interviewing process because the informants could only speak their local language. To allow a smooth flow of the interview and to maintain the informants’ morale, the interpreter was just summarizing the responses to the interviewer to just allow her to ask follow-up questions, a procedure recommended by Pitchforth and Van Teijligen (2005). However, this was constraining in terms of further probing or seeking clarification. Likewise, involvement of interpreters can add an additional layer of bias within the research process (Pitchforth & Van Teijlingen, 2005). To minimize these, we recruited the interpreters with a mental health background as they already had experience in conducting clinical interviews which are not so different from qualitative interviews. The interpreters were also briefed about the study topic and trained in conducting the interviews in accordance with the interview guide. The initial interviews were also used as pilot interviews to point out potential flaws. The subsequent retraining yielded very rich descriptions that was informative enough to make a meaningful interpretation of the data.

Direction for further research

As we recommended in our study of men's suicides in this area (Kizza et al., in press) there is a need to observe the occurrences of suicide and further examine whether the above factors are still prevalent among women now that the Acholi are supposed to have settled back in their homes after the closing of the IDP camps. It would also be of interest to have this study replicated in other areas in the country that have never been affected by war. These findings together would go a long way to foster theory development relevant to a Ugandan context.

Implications for suicide prevention

In this study we have demonstrated that women’ suicides in post-conflict Northern Uganda may partly have been related to men's pursuit of their changed masculinity. In the process of doing so, men caused these women physical and emotional harm the pain of which could only be alleviated by suicide. A divorce may have worked for the younger ones and a social security system for the older one. But none of these options were available. As women are empowered to know and fight for their rights, men should also be helped to regain their self-esteem associated with their changed masculinity. Women need to become empowered and also appreciated for their contribution to the household as much as men are. This has to be done in concerted effort by the entire social services sector including cultural institutions and NGOs. Furthermore, as also recommended in Kizza et al. (in press), both men and women in this region, including the cultural leaders, should be helped to appreciate that the conditions which prevailed in their part of the country in the two decades of conflict tremendously affected their traditional gender roles. This also calls for adjustment of the balance of power relationship between men and women to become commensurate with the changed gender roles.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.