In April 1994, Rwanda was immersed in a brutal wave of organized violence that left an estimated one million people dead in a period of only three months. Civil war, genocide against the Tutsi minority group and violent reprisal attacks until 1998 horrified its inhabitants. The thoroughly planned and state-monitored genocidal violence was specifically marked by the extensive participation of the local population: neighbors went after neighbors by means of guns, machetes or sticks during house to house searches, at roadblocks or at central congregation points. Looting, destroying property and genocidal acts including murder and sexual violence were common [1]. Overall, more than 10% of the country’s 7.8 million population and approximately 75% of the Tutsi ethnic minority were killed and a huge number of people ended up widowed or orphaned. In the direct aftermath of genocide, two million people took refuge in the neighboring countries. Many of them did not reenter Rwanda prior to 1996, when the refugee camps began breaking down and people felt encouraged and/or coerced to return. In many cases, a return to Rwanda was followed by immediate incarceration. The release of these prisoners did not begin before 2002 when Gacaca, a judicial initiative based on a traditional Rwandan mechanism of local conflict resolution, was implemented to confront the estimated 1.2 million cases [2]. Until the present day, genocide survivors and those who participated in the genocide continue to live next door to each other.

In the aftermath of 1994, genocide survivors showed high rates of mental health and psychosocial problems due to the inconceivable, dehumanized brutality that the majority of them had been exposed or witness to. Entire family systems as well as the general social fabric that formerly provided support were destroyed due to losses of family members and growing mistrust and fear following the genocide. A great majority of the survivors were female and woman-headed households proved to be especially vulnerable, suffering from the effects of economic deprivation, which included a lack of food, housing and money for the education of their children [3]. Apart from general population samples, studies analyzing the mental health situation in Rwanda following the genocide have mainly focused on groups of widows and orphans or children living in child-headed households. An elevated level of depressive and anxious symptoms as well as PTSD was found in each of these groups [4–7]. On the other hand, little to nothing is known about the mental health situation of former prisoners in Rwanda, many of which spent several years in refugee camps after 1994. It is assumed that former prisoners – that is, accused perpetrators and their respective families – also present with mental health problems, whether due to their participation in [8, 9] or exposure to violence, genocide and their refugee status. Concrete data on this is, however, currently lacking.

To our knowledge, no existing study has used data from a local population sample investigating both survivors and former prisoners as well as their respective descendants living together in the same communities. We opted for a comparison of these two, as they constitute important stakeholders within the post-conflict Rwandan society and assumingly present with different mental health and socioeconomic profiles. This also assumed that a neutral, unaffected group would be near impossible to find owing to the pervasive effects of violence on the Rwandan population as a whole. As the primary targets of genocidal acts were men and boys, genocide survivors proved to be mainly women, whereas genocide-related crimes were mainly committed by men [10]. A gender imbalance was therefore assumed in the examined sample of parents. We furthermore hypothesized that families of genocide survivors had been exposed to more traumatic events and that they were more affected in terms of mental and physical health. With regard to social and economic factors, we assumed that they live under worse conditions than the group of former prisoners and their descendants.

The present study examined mental health problems and psychosocial conditions in Rwandan families 16 years following the 1994 genocide. Its main aim was to investigate the impact of war, genocide and other potentially traumatic experiences on genocide survivors on the one hand and former prisoners on the other hand, as well as the respective descendants of both groups. The study also examined correlates of PTSD and the prediction of the PTSD symptom severity. In general, survivors and their descendants reported more traumatic events and proved to be more affected than families of former prisoners. Posttraumatic stress reactions were especially elevated in adult survivors who had experienced a high number of traumatic events, had poor physical health and were lacking in social integration.

Not surprisingly, survivors and their descendants, as the primary targets of the 1994 atrocities, showed the highest exposure to traumatic stressors with twelve and ten different event types, thus reflecting their exposure to genocide-related violence. These findings are in line with other studies conducted in Rwanda [4, 5, 9]. The average number of event types reported by former prisoners and their descendants ranged from six to nine events and was also mainly linked to the period of genocide, although both generations in this group emphasized its aftermath more frequently than the families of survivors. Former prisoners especially pointed to physical attacks experienced in refugee camps in the eastern Congo or related to the imprisonment upon their return. Their descendants often became witnesses of these imprisonments and the circumstances under which they took place. In this way, the past and recent political situations in Rwanda, which were marked by various episodes of persecution, attack, massacre, and forced displacement, were also directly reflected in the number of events reported before and after 1994. This furthermore highlighted the repetitive and cumulative nature of trauma in Rwanda and the Great Lakes Region, which is not only limited to genocide.

In the present study, 25% of the genocide survivors and 22% of the former prisoners were diagnosed with PTSD. In Rwanda, studies reported 25%-29% of PTSD in non-specified adult populations [6, 22, 23], 41%-51% in widows and genocide survivors [6, 7] and 37% within the Southern province of Rwanda [24]. Our sample therefore showed a lower level of distress than previously reported data on Rwanda, while also presenting a high trauma load. In addition, it is in discordance with data collected in the Southern province. These differences might be due to recovery over time [25] but might also be linked to differences in exposure to genocide within the same province. As stated by Straus [26], Gitarama manifested less “anti-Tutsi violence” in comparison to other Southern cities such as Butare or Gikongoro, and, according to des Forges [27], the nearby Kabgayi church offered special protection to a great number of Tutsi in the area. While in the present study genocide survivors and former prisoners significantly differed in their PTSD severity scores, this was not the case with syndromal PTSD. This was due to the fact that both groups manifested the same level of intrusions – the B criterion of the DSM-IV diagnosis. While this elevated level of intrusions seems unsurprising in genocide survivors, it needs further explanation with regard to former prisoners. The prisoners examined in the present study had spent about eight years in prison: some of them were incarcerated in the direct aftermath of the genocide and others upon their return from refugee camps, as only a few did not leave their home district in 1994. For those prisoners, to be put in prison might have felt like the point of no return as the accused did not necessarily expect to ever leave prison or at least not until the implementation of Gacaca jurisdiction in 2002. Adverse experiences throughout their prison time such as malnutrition, lack of appropriate health care, overcrowded detention conditions or physical harassment and attacks might have added new traumatic events to the already existing fear network [28], and feelings of hopelessness and helplessness over the years might have fostered and maintained intrusive symptomatology. Acute fear, one might argue, characterized their time in prison as well as the arrival in their respective communities as they did not know what to expect and how they would be perceived by others in this changed political environment.

Comparing rates of symptoms of anxiety and depression between survivors and former prisoners, the former showed a significantly higher level of distress. With regard to the HSCL score, 37% of survivors and 22% of former prisoners fulfilled the criteria for anxiety disorder, and according to the Bolton algorithm 30% versus 15% for depression. When relying on these results, our findings are consistent with earlier reported studies on survivors and widows or on the general Rwandan population [6, 7, 20, 23, 29], even though no comparable data are available for former prisoners. A high level of disagreement was nonetheless found when comparing prevalence rates gained either by using a cut-off score or following DSM-IV symptom criteria, as suggested by Bolton and others [20], to screen for depression. In a recently published study, Ertl et al. [16] critically discussed the unevaluated adjustment of a cut-off score developed in a different context – that is, for example, not appropriate to the given East African situation. Therefore, in future research, the HSCL might be better applied to investigate symptom severity instead of prevalence rates based on a specific cut-off value.

Suicidal tendencies were found in 25% of all survivors and in 7% of the group of former prisoners, occurring more often among women. Rates showed to be lower compared with an earlier reported study [7], but still displayed a considerable level of distress in a society where suicidal tendencies had not previously been commonly reported and were rejected by the majority of our sample as, to quote a participant, “an inappropriate way to solve problems for any Rwandan believing in God”. Alcohol consumption occurred twice as much in the group of former prisoners (15%) than in the survivors and was especially linked to males. This had not received much attention in earlier reported studies on Rwanda and is not easily admitted to by Rwandans, whose sense of disclosure prohibits openly talking about sensitive topics, even while the consumption of locally produced alcohol is a common phenomenon in rural areas and symbolic for good neighborhood relationships [30]. To further differentiate between general alcohol consumption and clinically relevant problems associated to alcohol and other substances, a recent representative study by the Rwandan Ministry of Health [23] examined a general population sample and found rates of drug and alcohol abuse ranging from 3%-6% and alcohol addiction of 5%-7%. Even while our data do not allow for any causal attribution, one might argue that these two specific features, suicidality and alcohol abuse, point to possible reactions and mechanisms for dealing with loss and trauma. Family dynamics might therefore be affected by these issues, especially if the broader family and community support is broken. Furthermore, it is possible that the experience of war and violence can lead to an elevated level of family violence, which often turns out to be moderated by alcohol abuse in a parent. Catani et al. [31] demonstrated an association between the father’s alcohol intake and maltreatment reported by his children in a Sri Lankan sample of children affected by long-lasting conflict. Although systematically collected data on domestic violence in Rwanda are scarce, another recent study reported that alcoholism ranges under the first three causes of aggressive and violent behavior towards intimate partners or children in this country [32]. Therefore, psychological disorders within the local population following experiences such as war and genocide, including alcohol and drug abuse, need to be considered in community-based interventions, which tend to diminish the risk of further violence on following generations.

Overall, the present study demonstrated a high degree of co-morbidity between diverse disorders, as postulated earlier [33–35]. Altogether, one quarter of all adult survivors suffered from PTSD, clinically relevant depression and/or anxiety, reflecting the serious mental health situation as well as the long-term consequences of massive violence even 16 years following the genocide [36–38].

With regard to the group of descendants, our study revealed that 16% of the descendants of survivors compared to only 1% of the descendants of former prisoners (and none of those born after 1994) fulfilled the DSM-IV criteria for the diagnosis of PTSD. In a study on a general sample of Rwandan youth interviewed during the direct aftermath of the genocide, Neugebauer et al. [39] reported a PTSD rate of 62%. Recent research on vulnerable groups such as orphans showed lower PTSD rates, between 24% and 34% [5, 7, 40]. With regard to those born before 1994, our sample of descendants of survivors manifested a similar level of PTSD to those reported by these last studies. Among these descendants, a particularly high trauma load was found and 50% showed to be half-orphaned. Their specifically vulnerable and life-threatening situation in 1994 and afterwards was strongly shaped by their families’ experiences. Due to persecution and death, parental protection throughout the period of violence was often missing. In the aftermath of the genocide, their families had to cope with severe circumstances and descendants often took over great responsibilities, which often continue today and might explain the ongoing sequelae of distress as depressive and anxiety symptoms [7]. The group of descendants of former prisoners within the present sample, however, differed even more from the youth described in Neugebauer’s study, though concrete comparable data is missing. Throughout the genocide, descendants of former prisoners did not necessarily flee with their families, but rather stayed at home or were individually sent to other remaining family members. When they had to take refuge with their families who were moving to the western parts of Rwanda, they often went in groups and the mainly Hutu background of their mothers offered them special protection in comparison to the descendants of survivors. Even if they had witnessed war and genocidal violence, they had never been specifically targeted, as the primary aim of the genocide perpetrators was to eliminate the group of Tutsi and their families [24]. Finally, information on what was going on in Rwanda in 1994 was scarce. A lack of cognitive understanding of the dimension of the events might therefore also have modulated the affect regulation and in turn have added a protective factor for those children [41]. Our data furthermore suggest that younger children born after 1994 did not specifically suffer from PTSD or other mental disorders. Further research is needed to better understand potential transgenerational effects of genocide on those children who did not live through the genocide in comparison to their older siblings [42]. Apart from these family issues, when referring to previous studies the broader social climate is in question as well, thereby demonstrating a clear association between mental health problems such as PTSD and feelings of hatred and revenge in the aftermath of conflict [22, 43]. These possible adverse implications also need to be considered while developing initiatives to foster reconciliation and mutual understanding.

Another key finding of the present study was that the number of event types as well as physical health and social integration explained the biggest part of the variance of posttraumatic stress symptoms in the parent generation. The presently observed dose-effect of the number of traumatic event types on the PTSD symptom severity score – highlighting the impact of cumulative stress on mental health – has already been widely discussed [44, 45]. The relationship between PTSD and lower self-reported physical health and other health problems has also been reported in previous studies [46, 47]. In a recently published study by Schaal et al. [7], both physical illness and trauma exposure were the two main predictors for PTSD symptom severity in Rwandan widows and orphans, while social factors were not further differentiated. The authors discussed this association as a possible difficulty of survivors with PTSD for developing effective coping mechanisms to deal with somatic and chronic health problems or, conversely, that the latter might affect them in such a way that they are no longer able to take care of themselves. Our findings demonstrated chronic pain to be the main physical complaint. With regard to the hypothesis that both syndromes mutually maintain each other as, for instance, acute pain proves to be mediated by symptoms of arousal and vice versa [48, 49], this offers an alternative explanation of why PTSD continues to occur at this level, even 16 years after the genocide. Recovery without any treatment or only basic medical-oriented services seem to be reserved to only a fraction of the population. While no evidence for direct prediction of PTSD symptom severity using economic factors was found, physical illness seems to function as a mediator between both, as its correlates with social as well as economic factors demonstrate. The direct consequences of genocidal violence such as HIV infection, chronic pain or disability, which were especially present in the group of survivors, have an immediate impact on the economic growth of a family in an already poor environment [50]. The respondents mainly worked as peasants, as is common for the rural Rwandan population. Therefore, as reported by an interviewed female survivor, when a widow was a victim of sexual violence during the genocide and, due to continued bleeding and other associated physical ailments, was no longer able to perform hard farm labor, her family’s economic status was subsequently and negatively impacted. Even if this kind of survivor were willing to receive psychological support, the challenges of the distance from home to center, money for travel and privacy from neighbors highlight the delicate interconnectedness of economic and social issues in this region.

The present study revealed that lower levels of social integration and activity were associated with elevated levels of PTSD symptoms, indicating a mutual maintenance effect. Interestingly, women in our sample had a lower social status than men, whereas no difference was found either between survivors and former prisoners or between widowed and non-widowed persons. One possible explanation for these results might be that especially vulnerable groups such as widows or genocide survivors infected with HIV tend to stick together and support each other, which is not necessarily the case with people currently suffering from PTSD [51]. According to our findings, genocide survivors had lost an average of 14 family members. In Rwanda, family and community support is crucial for the well-being as well as for the reputation and prosperity of individuals. Depressive moods and feelings of hopelessness in survivors, therefore, were often linked to this missing support, as reported by respondents. As remarriage is socially not necessarily tolerated, widows have difficulties taking care of their family in an appropriate way, especially when they are not further integrated in the community [52].

In comparison to this, former prisoners showed a better economic profile, but surprisingly no higher level of social activity and integration. In our study, while released prisoners rarely mentioned problems of reintegration following their release, their other family members such as their wives and children did. As Richters et al. [50] demonstrated, sources for ongoing conflicts might be found within families in which the father is extensively absent due to imprisonment. In such circumstances, the man returning home from prison may find that his wife has brought another man home or has even had children with other men. Additionally, there has been sharp economic decline and property loss – all factors leading to the father’s realization that he has limited authority in the new family system facing him following imprisonment [53]. Alcohol abuse and the aggressive behavior of released prisoners toward family members or, on the contrary, social withdrawal might also affect family dynamics. As several researchers have demonstrated, mistrust in the communities started to grow again after the first massive release of prisoners in 2003 in Rwanda, many of whom have since been presumed innocent or assigned to minor offenses. Alliances that had previously been made between, for example, genocide widows and wives of prisoners, were thereby once again put into question [54, 55]. A large portion of our sample of former prisoners was released within about the same period of time and therefore took part “in the government confession program” that provided a reduced sentence for perpetrators who admitted guilt and remorse [56, 57]. Released prisoners expressed being especially grateful to the government as well as for the introduction of the Gacaca tribunals, as most of them were released due to this new judicial initiative. At the same time, the majority felt that their own suffering due to imprisonment did not receive any recognition. Therefore, one can argue that social reintegration as a so-called aim of governmental-driven directions on how to behave when going back to their families and communities [56, 58] might not necessarily be experienced by the individual former prisoner. This also includes the notion of being innocent and a victim of “someone else’s war” [59], as, in the present study, released prisoners frequently reported physical attack or incarceration as their worst experiences according to the PSS-I. These results are in line with data from a recently published study on incarcerated, accused perpetrators in Rwandan prisons [9]. Only 13% reported their participation in murder as their most stressful event; Schaal et al. thus argue whether high rates of mental health problems in this population are in fact due to the causes or the consequences of imprisonment.

These results highlight the ambiguous and complex nature of victimhood in post-conflict societies as well as the need for further empirical evidence and lead to a first limitation of the study: comparing survivors with former prisoners seems critical with regard to its restriction in conceptualization. As emphasized by numerous authors, dichotomous categories are inappropriate when describing and reflecting complex circumstances of life in and after wartimes [12]. A dynamic view on participation is required whenever a better understanding of the role and participation in periods of violence is in question. People do not often fit to one category alone, such as victim, perpetrator or bystander, and can change throughout time in their concrete behavior. This change and assumption of different positions within the same period of time could, therefore, influence the impact on their mental health situation. Following Bar-On [60] there might be an association between the use of interchangeable roles and reduced moral responsibility in a person, which together could have a disburdening effect. Therefore, one individual who protected and rescued a nearby Tutsi neighbor (who in turn survived the genocide) while having also participated in roadblocks and manhunts for the Tutsi might be a conceivable example of how one person could incorporate many roles at the same time. Our findings might not necessarily hold for the whole Rwandan population. As already demonstrated, results from different regions might vary and our data can therefore only be seen as representative for central Gitarama. Still, they offer initial insight into a local population consisting of families of survivors and former prisoners living next door to each other who had never before been examined together and compared with each other. Another limitation lies in our definition of family. In Rwandan families, older children in particular do not necessarily grow up near their parents or siblings, but often refer to the broader family context including uncles, aunts, or even grandparents as their parents or to cousins as their siblings as well. The present study does not provide further information about the previous childhood circumstances of descendants in order to draw further conclusions on potential protective factors in the aftermath of violence. Finally, conducting research in a post-conflict setting demanded a critical reflection upon data validity. Due to political restriction and oppression, speaking out loudly is not common in Rwanda. The effect of introducing local interviewers also needs to be taken into consideration. While local researchers normally benefit from their close relationships to the cultural background in question, the specific historically shaped relationships between Rwandans might sometimes foster an even stronger mistrust between Rwandans than toward foreigners.

16 years following the Rwandan genocide, survivors and their families continue to present with considerable rates of PTSD and substantial depressive and anxiety symptoms. The data revealed a strong association between health problems and psychosocial factors such as social integration. A high level of PTSD in the group of former prisoners – and thus the likely genocide perpetrators – demonstrated that psychological suffering affected the population at large, although the nature of traumatic stressors may have varied. By challenging the question of who “owns” trauma discourses in a post-conflict society such as Rwanda, mental health services should take the needs of the entire population into consideration when offering care – an idea that is still not mainstream in Rwanda. Finally, descendants of genocide survivors showed a higher risk for mental health problems than descendants of former prisoners. A high trauma load as well as missing family integration and support characterizes their specific vulnerable situation. Interventions and initiatives that stimulate reconciliation processes between future generations should take these particularities into consideration.