Longitudinal randomized comparison study on the community resiliency model for addressing mental health challenges in survivors and perpetrators of genocide in Rwanda

Abstract

Background

In a post-genocide context, mental health disorders among Rwandan genocide survivors and released perpetrators remain a critical concern. To date, no study has evaluated the effectiveness of the Community Resiliency Model (CRM) skills in addressing the mental health needs of both groups simultaneously. This study assessed the impact of CRM when delivered to a combined group of survivors and perpetrators, compared to groups trained separately.

Methods

A total of 152 participants were recruited from Nyamagabe district, Rwanda. Participants were assigned into three groups including genocide survivors (n = 51), released genocide perpetrators (n = 51), and a combined group of both survivors and perpetrators (n = 50). Data were collected at three points: pre-intervention, immediately post-intervention, and six months post-intervention using validated psychometric scales for anxiety, depression, posttraumatic stress disorder (PTSD), emotional dysregulation, and anger. Repeated measures ANOVA and Bonferroni post hoc tests were used to analyze changes over time. A statistical significance of p < 0.005 and p < 0.001 was applied.

Results

Our findings showed significant reduction of anxiety (F = 20.17, p < 0.001), depression (F = 37.03, p < 0.001), anger (F = 95.97, p < 0.001), and emotional dysregulation (F = 76.68, p < 0.001) across all groups of participants. These positive changes were sustained at 6 months post-intervention for anxiety, depression, anger, and emotional dysregulation. In contrast, PTSD symptoms only showed a slight, non-significant reduction over time (F = 0.59, p = 0.44). Additionally, there were no significant differences in outcomes between groups that received the intervention separately (survivor-only or perpetrator only) and those that received it in mixed survivor-perpetrator groups.

Conclusion

Although the CRM intervention does not replace psychotherapy, it produced lasting and positive effects on mental health symptoms among both genocide survivors and perpetrators, particularly in reducing anxiety, depression, and emotional dysregulation. Importantly, outcomes did not differ whether the intervention was delivered to separate or combined groups. A randomized controlled trial is recommended to further evaluate the long-term effects of CRM on community healing and cohesion.

Background

Exposure to colonization, political instability, civil wars, and mass atrocities including genocide leads to deep and lasting mental health challenges by subjecting populations to prolonged violence, displacement, poverty, and social fragmentation [1–3]. These exposures cause cumulative trauma that severely undermines individual wellbeing, disrupts family cohesion, and weakens the fabric of entire communities [4–6]. Genocide survivors are particularly vulnerable, as they endure extreme and repeated traumatic experiences including exposure to death, killings, torture, sexual violence, and destruction that place them at significantly higher risk of mental health conditions compared to other conflict-affected populations [7, 8]. These conditions not only increase the risk of depression, post-traumatic stress disorders (PTSD), and anxiety disorders but also severely affect important socio-cultural values, family dynamics, and community cohesion Additionally, the resulting poverty and harsh living conditions further exacerbate these mental health issues [9].

A growing body of scientific evidence shows that historical and collective trauma in communities affected by discrimination, wars, or genocide contribute to disparities in mental health outcomes between oppressed and processor groups [2, 3, 10]. For example, in South Africa, after more than 20 years of apartheid, research indicates that the victims of apartheid and their descendants face a significantly higher risk of depression, PTSD, and anxiety disorders compared to the oppressors and children from white families. The Black adolescents experienced a relative risk (RR) of 2.27 for depression and 2.21 for PTSD compared to white adolescent [10]. Consequently, much research and intervention efforts focus on addressing the mental health needs of survivors [8]. However, a notable gap remains because few mental health interventions have been implemented to promote mental health for both groups. For example, trauma-informed interventions are designed to address the intertwined needs of both genocide survivors, and perpetrators especially in contexts where community rebuilding and reconciliation are important examples of effective interventions in other settings include trauma-informed programs that reduced mental health problems in Liberia [11], behavioral interventions have alleviated relationship-related challenges on former child soldiers in Sierra Leone [12], and integrative mental health care has reduced mental health problems related to apartheid in South Africa [13].

Rwanda, a small, densely populated, landlocked country of about 14 million people [14], faced the devastating 1994 genocide against the Tutsis during which over one million lives were lost. In the years that followed, the country faced massive displacement, a surge in poverty, the spread of mental disorders, and the emergence of thousands of orphans, and widows. More than 225,000 people were imprisoned for genocide crimes, destroying over 3 million households [15]. Beyond the killings, many survivors bore long-term injuries and disabilities that not only affect them but also communities and their families. Displacement was widespread, with many fleeing within Rwanda or to neighboring countries like Democratic Republic of Congo, where they faced additional health concerns such as cholera outbreaks and famine that killed more than additional 48,000 Rwandans [16, 17]. Today, a significant portion of population lives on less than 0.60$ per day, far below the World Bank poverty threshold. The complicated history of colonization by Germany and Belgium, followed by political instability and civil wars, ultimately set the stage for the 1994 genocide [18, 19]. Three decades later, the mental health impacts remain profound: over 35% of genocide survivors experience depression and nearly 30% suffer from PTSD [18], a condition following exposure to traumatic events that presents as four main symptoms, including re-experiencing (e.g. flashbacks), avoidance (e.g. avoiding trauma-related issues), negative alterations in mood and cognition (e.g. persistent anger, fear), and alterations in arousal and creativity (e.g. lack of concentration, sleep disturbances) [20]. Notably, PTSD can occur after experiencing trauma directly, witnessing trauma, or through repeated exposures to details of traumatic events [21]. Moreover, about 40% of perpetrators also report significant mental health problems [18, 22, 23] as research documents that many released perpetrators return to their communities with moral injuries, identity crises, emotional dysregulation, and deep emotional pain [24, 25]. These challenges hinder psychosocial reintegration and perpetuate cycles of trauma. It is therefore critical to develop effective mental health interventions that address these psychosocial issues and support individual recovery and community resilience in post-genocide settings [26, 27].

Programs to address the continuing mental health challenges

To address the above-mentioned psychosocial effects of the genocide, the Rwandan government together with local, national and international organizations including non-governmental organizations (NGOs), has implemented several programs and interventions targeting to improve mental health care in the post-genocide period [28]. These efforts included decentralizing mental health services from referral hospitals to health centers, expanding the mental healthcare professionals, and integrating psychology and mental health into academic training programs [29]. Despite these efforts represent major progress; mental health needs remain far greater than the service available, leaving many survivors and perpetrators as well as their descendants without adequate care [22, 24, 30]. Therefore, the rate of psychosocial distresses continues to be high among both survivors and perpetrators [24, 25, 31].

In June 2002, after adjudicating the crimes committed during the genocide through the Rwandan court system, Rwanda implemented an innovative form of justice known as “Gacaca courts,” which used traditional cultural practices to progress toward justice, reconciliation, and healing [24, 32]. The Gacaca courts saw more than 30,000 trials, resulting in the incarceration of many genocide perpetrators. Though this cultural and community-driven system helped more genocide survivors access justice [33], other research showed that the Gacaca courts led to increased mental health problems, as both perpetrators and survivors were required to face each other in court [34]. Beyond justice processes, mental health interventions were implemented by government institutions, universities, and both local and international NGOs, though most programs concentrated on survivors and offered limited support for perpetrators [35–37]. For example, a quasi-experimental study on the Community Resiliency Model (CRM), a set of self-care skills derived from somatic psychotherapy demonstrated reductions in trauma symptoms, mental disorders like depression and PTSD, as well as the improvement in quality of life and resilience among survivors [38, 39]. However, with over 116,000 genocide prisoners released between 2014 and 2024 after the genocide [40], and many once again co-living as neighbors, existing interventions remain insufficient to address the complex, intergenerational impacts of genocide [24, 41]. There is an urgent need for inclusive mental health initiatives that simultaneously support survivors and perpetrators, while also fostering healing, reconciliation, and social cohesion.

To address this gap, we were motivated to implement the CRM among genocide survivors and released perpetrators in vulnerable communities. The CRM intervention is based on a theory of change that involves educating participants about self-regulation, engaging them in skill-based practice, and evaluating the changes in mental health outcomes [42, 43]. CRM is not psychotherapy rather, but a biological-based intervention consisting of six core skills, specifically, Tracking, Resourcing, Grounding, Gesturing, Shift and Stay, and Help Now. In addition, the model incorporates key concepts such as the ‘Resilience Zone,’ ‘Okay Zone,’ ‘High Zone,’ and ‘Low Zone,’ which describe states of well-being and stress responses of individuals [44]. CRM was first implemented in Rwanda in 2015, where participants nicknamed it as Inzira y’Umutuzo (translated as “a pathway to resilience”) based on the positive impact on their psychosocial wellbeing to reflect its application and cultural relevance in daily life. Since then, the staff (e.g. CRM teachers) of Rwanda Resilience and Grounding Organization (RRGO) adapted CRM to the local contexts in collaboration with multidisciplinary team of healthcare professionals (e.g. psychologists, social workers), community members (e.g. genocide survivors, ex-prisoners), university students (e.g. from nurse, clinical psychology, social work), and academicians. They cultural adopted and translated the CRM materials from English to native language (Kinyarwanda), which allow them integrating culturally relevant concepts. From that time, all CRM training sessions are carried out in Kinyarwanda language [38]. The six skills are interdependent; practicing one or two skills may be sufficient to restore well-being, while in other situations, individuals may choose to engage in all six, depending on personal needs and circumstances [43].

Previous research established effectiveness of CRM in Rwanda among genocide survivors [38], as well as in other post-conflict and low-resource settings such as Sierra Leone, where the population experienced a long history of civil war and more recently the Ebola crisis [45]. This model has also been shown to foster psychosocial well-being in other low-resource or non-conflict settings, such as San Bernardino, California, where it improved the resilience, well-being and reduced stress among underserved traumatized individuals and communities [39, 46]. Despite this growing evidence base, there has been no study conducted in Rwanda that evaluates a CRM intervention designed to address the mental health needs of both genocide survivors and released perpetrators simultaneously, particularly in the contexts of post-genocide. To address this critical gap, we conducted a longitudinal study to evaluate the effectiveness of the CRM as a mental health intervention among genocide survivors and released perpetrators. Specifically, this study aimed to determine whether CRM could reduce symptoms of PTSD, emotional dysregulation, anxiety, and depression in both groups. We hypothesized that the CRM intervention could reduce the symptoms of mental health issues like depression, anxiety, PTSD, anger, and emotional dysregulation in both survivors and perpetrators, whether delivered in separate or combined groups. By examining the effectiveness of the CRM in combined versus separate groups impacted by the Rwandan genocide, this research promises to offer valuable insights how to most optimally delivery community-based mental health interventions in post-conflict settings like Rwanda.

Methods

Study design

We implemented a randomized longitudinal comparison study to examine types of delivery outcomes among genocide survivors alone, perpetrators alone and a mixed group of survivors and perpetrators.

Procedures and study settings

This research represents the outcome phase of a larger project that began with a qualitative study conducted in early 2023, which explored the feasibility of, and the specific needs for, mental health and reconciliation interventions among genocide survivors and perpetrators in Rwanda. The initial findings highlighted a strong desire among participants for community-based interventions that would bring together both survivors and perpetrators, reflecting their current reality of living as neighbors who must learn to coexist peacefully. In response, we delivered the CRM intervention between August 20 and September 22, 2023, followed by a 6-month post-intervention evaluation between February 17 and March 4, 2024, to assess its impact on mental health outcomes across survivor-only, perpetrator-only, and mixed groups.

The study was conducted in Nyamagabe district, one of 30 districts of Rwanda, which together with Kigali City forms the country’s administrative structure. Nyamagabe is one of six districts that make up Rwanda’s Southern Province. This district was intentionally selected for the study because it is historically one of the national areas most severely affected by the genocide. This study area has a high concentration of both genocide survivors and perpetrators who have since been reintegrated into the community, making it a critical setting for interventions designed to address cumulative trauma and strengthen social cohesion. Four sectors within Nyamagabe specifically Cyanika, Kamegeri, Tare, and Gasaka were randomly selected from among those most impacted by the genocide, characterized by prolonged violence and large populations of survivors and former perpetrators. This site provided a meaningful and complex environment to evaluate the feasibility, acceptability, and effects of the CRM intervention [28].

Population and participants

Participants were recruited through convenience sampling, with local leaders providing lists of genocide survivors and perpetrators who had completed prison sentences for genocide-related crimes and had reintegrated into their communities. The inclusion criteria included (i) being 18 years or older; (ii) being either a genocide survivor or a perpetrator who had been released through completion of a sentence, confession, forgiveness-seeking, or community service; (iii) residing in one of the targeted sectors of Nyamagabe district; and (iv) being willing and able to provide informed consent to participate in the CRM skills training and related study activities. Individuals were excluded if they had severe untreated mental health conditions (such as psychosis) that would prevent meaningful participation, if they were not residents of the selected sectors, or if they declined to give consent or could not commit to attending the sessions. After the lists of eligible participants were finalized, each person was assigned a number, and a random number generator was used to allocate participants to one of three group formats: survivor-only groups (n = 51), perpetrator-only groups (n = 51), or mixed survivor–perpetrator groups (n = 50). This process ensured that all participants had an equal chance of being assigned to a mixed group or a group based on their status. The research team, supported by local leaders and data enumerators, met with potential participants to explain the purpose of the project, the training process, and the group structures. All those approached agreed to take part in the CRM skills training (Fig. 1).

Fig. 1.

Study participants and their social categories. Notes Samplings and flow chat of participation: PP: Participant perpetrators, PS: Participants Survivors and PPS: Participants both perpetrators and survivors

Description of a community resiliency model intervention

The proposed intervention of this study is CRM skills training; a biologically based approach used for enhancing individual and community resilience by addressing psychosocial disorders. As described in Fig. 2 below, the CRM training offers six wellness skills to restore mental, physical, and spiritual well-being during or after a traumatic experience, educating participants about the autonomic nervous system (ANS) and its responses to stressful situations [47].

Fig. 2.

Community resilience model skills framework

The intervention training employed in this study was delivered to six groups of 20 to 25 participants per group. Participants for each group were residing in the same neighborhood. The training lasted 3 days and each session took 7 h per day. Sessions occurred in safe locations within the participants’ immediate living environments, including schools, churches, offices, and private spaces, or outdoor areas under a tree or in a grassy field. The trainings were delivered by trained local CRM teachers (one lead instructor and three facilitators) certified by the Trauma Resource Institute [43]. Of note, while the core skills (Fig. 2) remained the same, content and deliveries were adapted based on the results from Phase I. For instance, during the training, participants primarily discussed present-day and past problems related to the specific skill. To respect and reflect the Rwandan context practical and locally relevant resources from their daily life related to each skill were used.

While overall no significant tensions emerged within the mixed groups of survivors and perpetrators, on the first day, the survivors sat on one side of the room while the perpetrators sat on the other side. During the session, one survivor participant noted that the seating arrangement seemed divided. A group facilitator asked the group to sit in any way they felt comfortable. All participants agreed that they would mix the seating arrangement to facilitate the session, since they were practicing together. The groups voluntarily changed their seats. From that point onward, they continued sitting together for the remainder of the training.

.

Study variables and materials

The primary outcome of this study was the mean change in mental health symptoms including depression, anxiety, anger, PTSD, and emotional regulation between the baseline and the 6 months post-intervention evaluation. The secondary outcome was to evaluate changes between the 3 comparison groups (survivor-only, perpetrator-only, and mixed groups) over time. All mental health outcomes were assessed using validated, culturally adapted tools that had been previously applied in Rwanda and demonstrated high levels of comprehension, cultural appropriateness, and acceptability among Rwandan respondents [48, 49].

Data collection procedures

CRM skills trainings were delivered in the participants using a native language (Kinyarwanda) by certified and well-trained facilitators from the Rwanda Resilience and Grounding Organization (RRGO). Data collection was carried out independently by trained enumerators who were blinded to participant identities and group assignments, thereby minimizing potential bias. Data were captured electronically using KoboToolbox, a secure platform designed for efficient and reliable collection of quantitative data in field settings. All data collectors completed comprehensive three-day training covering the study objectives, use of the data collection software, ethical principles (including confidentiality and informed consent), and standard operating procedures for interacting with participants. Data collection occurred at three key points: prior to the intervention (baseline), immediately after completion of the training, and at six months post-intervention.

Measures

Socio-demographic information gathered included gender, age, marital status, education level, employment status, religious affiliation, food insecurity (e.g., concerns about running out of food each month), household composition (living alone or with family), perceived safety in their environment, history as a victim or perpetrator, and household access to electricity. Mental health outcomes were assessed using four psychometrically robust and validated instruments:

• Hopkins symptom checklist (HSCL-25) It is a 25-item scale measuring anxiety (10 items) and depression (15 items) on a 4-point Likert scale [50]. This psychometric instrument has been validated in Rwanda and proven reliable for the Rwandan population. A cut-off point of 1.75 in refugee situations and cross-cultural research was often used [51]. In our study, a good internal consistency was found on anxiety subscale (Cronbach’s alpha, α = 0.87), and on depression scale (Cronbach’s alpha, α = 0.92).

• PTSD checklist for DSM-5 (PCL-5) It is a screening measure of PTSD [52]. The items of this instrument are responded using a 5-point Likert scale ranging from not at all (0) to extremely (5) [53]. Referring to the study conducted in Rwanda, a PCL-5 cut-off score range of 31–33 predicts probable PTSD across samples. This instrument was mostly used in Rwanda, validated and indicated good psychometric measures such as validity and reliability [54].

• Difficulties in emotion regulation scale (DERS) A 36-item psychometric instrument that assesses emotional awareness, clarity, acceptance, impulse control, and access to emotion regulation strategies [55]. The items of this measure are scored using a 5-point Likert scale. The DERS has been validated in Rwanda and demonstrated excellent internal consistency in our stud (Cronbach’s Alpha; α = 0.955) [56].

• Dimensions of Anger Reaction-Revised (DAR-5) This five-item psychometric instrument assesses anger frequency, severity, length, aggression, and interference with social functioning [57]. Its items are rated on a 5-point Likert scale with scores ranging from 5 to 25, with higher scores indicating severe symptoms of anger [58]. We validated this tool and found a great internal consistency (Cronbach’s alpha, α = 0.83).

Data analysis procedures

Quantitative data were cleaned, and scales were created following published guidelines. Descriptive data were presented with percentages (%) for categorical data and mean and standard deviation for continuous variables. Assumptions of normality were tested, and sphericity was assessed using Mauchly’s test. Greenhouse-Geisser correction was applied to adjust for violations of sphericity. We used psychometric measurement to determine mental health symptoms such as depression, anxiety, anger, PTSD, and emotional dysregulation. Further, we performed ANOVA one way to compare the significant difference among outcome variable on all participants groups at baseline level. We also used paired measurement ANOVA to determine whether the CRM intervention affected changes in means over time (pre-evaluation, immediate post-evaluation, and 6-month post-evaluation) on mental health outcome variables. Lastly, the significance of variables was confirmed through ANOVA, using a p-value threshold of less than 0.05.

Ethics

The ethics committee of the College of Medicine and Health Sciences at the University of Rwanda approved the study (N0: 106/CMHS IRB/2023). Before this, the authors received ethical approval from the Loma Linda University. Additionally, district local leaders of the study area authorized the data collection. Prior to data collection, the study and its procedures, including confidentiality measures to minimize risks were explained to the participants. All participants were informed that they had the right to withdraw from the study at any time without giving any reason. The consent forms were obtained from all participants. Interviews were conducted confidentially. All procedures were in accordance with the principles outlined by the Helsinki Declaration [59].

Results

Descriptive analysis for the participants

One hundred and fifty-two (n = 152) participants were recruited for this comparison group study. Groups made up of survivors and perpetrators had 51 participants each, and a combined group of survivors and perpetrators had 50 participants. Many of the participants in the group of survivors were females (62.7%), while the perpetrators group was primarily male (76.5%). Most participants attended high school and primary school, while a smaller number (17.8%) attended university. To measure poverty, we asked questions related to food insecurity, i.e., how often participants ran out of food. Food insecurity was found to be higher in the perpetrator-only and combined groups: 60.8% and 52.0%, respectively, compared to survivors, at 35.3%. Farming was the predominant occupation across all groups, representing 80.4% of survivors, 86.3% of perpetrators, and 82.0% of the combined group. Access to electricity varied, with higher proportions among survivors (52.9%) and the combined group (58%) reporting access, compared to genocide perpetrators at 27.5% (Table 1).

Table 1.

Socio-demographic descriptive characteristics of participants

Variables	Only survivors (51) N (%)	Released genocide prisoners (51) N (%)	Combined (50), N (%)
Gender
Female	32 (62.7)	12 (23.5)	18 (38.0)
Male	19 (37.3)	39 (76.5)	32 (64.0)
Age	53.16 (11.79)	59.75 (12.16)	53.96 (13.49)
Marital Status
Widower	23 (45.1)	40 (78.4)	33 (66.0)
Marriage	28 (54.9)	10 (19.6)	17 (34.0)
Single	0	1 (2.0)	0.
Education
Non formal	11 (21.6)	13 (25.5)	16 (32.0)
Primary	4 (7.8)	7 (13.7)	6 (12.0)
Secondary	24 (47.1)	22 (43.1)	24 (48.0)
Technical	6 (11.8)	8 (15.7)	2 (4.0)
University	6 (11.8)	1 (2.0)	2 (4.0)
Worried food
Never true	3 (5.9)	2 (3.9)	3 (6.0)
Sometimes true	37 (72.5)	25 (49.0)	23 (46.0)
Often true	11 (21.6)	24 (47.1)	24 (48.0)
Run out food
Never true	3 (5.9)	1 (2.0)	3 (6.0)
Sometimes true	30 (58.8)	19 (39.3)	21 (42)
Often true	18 (35.3)	31 (60.8)	26 (52.0)
Feel safe
Never true	4 (7.8)	0	2 (4.0)
Sometimes true	17 (33.3)	14 (27.5)	21 (42.0)
Often true	30 (58.8)	37 (72.5)	27 (54.0)
Number of people house
Alone	2 (3.9)	4 (7.8)	4 (8.0)
Three people	14 (27.5)	11 (21.6)	15 (30.0)
Four people	11 (21.6)	14 (27.5)	12 (23.0)
More than 5 people	24 (47.1)	22 (43.1)	19 (38.0)
Employment status
Unemployed	5 (9.8)	2 (3.9)	2 (4.0)
Employed with wage	2 (3.9)	2 (3.9)	5 (10.0)
Farmers	41 (80.4)	44 (86.3)	41 (82.0)
Retired	0	2 (3.9)	0
Self employed	3 (5.9)	1 (2.0)	2 (4.0)
Religion
No religion	1 (2)	0	0
Catholic	19 (37.3)	29 (56.9)	26 (52%)
SDA	2 (3.9)	2 (3.9)	1 (2.0)
Protestant	26 (51.0)	18 (35.3)	19 (38.0)
Others	3 (5.9)	2 (3.9)	4 (8.0)
Electricity
Yes	27 (52.9)	14 (27.5)	29 (58.0)
No	24 (47.1)	37 (72.5)	21 (42.0)

SDA Seventh day adventist

Mental health symptoms change for pre, post and 6-month post intervention

Table 2 presents the mean scores of various mental health symptoms across multiple assessment periods for all participant groups. Initially, the study comprised 152 participants who underwent both pre-intervention and immediate post-intervention evaluations. However, by the 6-month follow-up, data attrition was observed, with eight participants (15.8%) from the only-perpetrators group missing from the dataset. The findings indicate that anxiety exhibited the most substantial reduction, particularly among genocide survivors, whose mean scores declined from 9.88 (SD = 3.54) at baseline to 6.72 (SD = 3.85) at the 6-month post-intervention follow-up. In contrast, the only-perpetrator group demonstrated a more modest decrease; with mean anxiety scores reducing from 7.88 (SD = 3.45) to 7.11 (SD = 2.93), and the combined group exhibited a decline from 8.80 (SD = 3.33) pre-intervention to 7.32 (SD = 3.33) at the 6-month follow-up. Other indicators like depression, emotional dysregulation, and anger showed a downward trend across all time points. However, PTSD symptoms revealed a slight increase within the group of only genocide perpetrators, rising from a mean score of 7.11 (SD = 3.35) pre-intervention to 7.46 (SD = 3.94) at the 6-month evaluation, the only-survivors and combined group decreased at the 6-months post intervention. This table suggests the changes that occurred on mental health symptoms in all groups could be sustained even 6 months post-intervention (Table 2).

Table 2.

Outcome variables over time for groups of survivors, perpetrators and mixed groups of survivors and perpetrators

Indicator	CRM Groups	Assessment	N	Mean	SD
Anxiety	Survivors	Pre intervention	51	9.88	3.54
		Immediate Post intervention	51	6.72	3.85
		Six Month Post intervention	51	6.76	2.56
	Perpetrators	Pre intervention	51	7.88	3.45
		Immediate Post intervention	51	5.81	3.41
		Six Month Post intervention	43	7.11	2.93
	Survivors and perpetrators, mixed groups	Pre intervention	50	8.80	3.33
		Immediate Post intervention	50	7.62	3.52
		Six Month Post intervention	50	7.32	3.32
Depression	Survivors	Pre intervention	51	14.78	5.16
		Immediate Post intervention	51	10.09	3.91
		Six Month Post intervention	51	11.07	3.74
	Perpetrators	Pre intervention	51	12.72	4.43
		Immediate Post intervention	51	9.89	4.04
		Six Month Post intervention	43	10.16	3.44
	Survivors and Perpetrators, mixed groups	Pre intervention	50	13.48	4.15
		Immediate Post intervention	50	11.76	4.37
		Six Month Post intervention	50	10.70	3.69
PTSD	Survivors	Pre intervention	51	7.86	3.40
		Immediate Post intervention	51	5.29	4.30
		Six Month Post intervention	51	7.56	3.60
	Perpetrators	Pre intervention	51	7.11	3.35
		Immediate Post intervention	51	4.60	4.90
		Six Month Post intervention	43	7.46	3.46
	Survivors and Perpetrators, mixed groups	Pre intervention	50	6.96	3.35
		Immediate Post intervention	50	5.82	4.49
		Six Month Post intervention	50	6.44	3.94
Emotional dysregulation	Survivors	Pre intervention	51	54.31	14.73
		Immediate Post intervention	51	43.15	19.11
		Six Month Post intervention	51	40.94	13.59
	Perpetrators	Pre intervention	51	55.65	12.62
		Immediate Post intervention	51	36.65	21.43
		Six Month Post intervention	43	41.67	12.44
	Survivors and Perpetrators, mixed groups	Pre intervention	50	50.36	9.193
		Immediate Post intervention	50	44.14	23.06
		Six Month Post intervention	50	38.42	13.04
Anger	Survivors	Pre intervention	51	13.00	4.31
		Immediate Post intervention	51	9.98	6.54
		Six Month Post intervention	51	8.23	3.96
	Perpetrators	Pre intervention	51	13.81	4.84
		Immediate Post intervention	51	8.39	5.07
		Six Month Post intervention	43	8.83	4.25
	Survivors and Perpetrators, mixed groups	Pre intervention	50	12.08	2.82
		Immediate Post intervention	50	9.78	5.81
		Six Month Post intervention	50	8.24	3.91

a. Based on modified population marginal mean

b. Immediate Post: Immediate post intervention, Pre: Pre intervention survey; Six Month Post: Six months post intervention survey; SD: Standard deviation

One-way ANOVA to explore baseline values on mental health outcome across the 3 groups (genocide survivors, perpetrators, and both groups together)

This study used one-way ANOVA to examine group differences on mental health outcomes at baseline. The results showed no significant differences across the three groups: for depression F (2) = 2.53, p = 0.083); anger (F (2) = 1.07, p = 0.34); PTSD (F (2) = 0.88, p = 0.415; and emotional dysregulation (F(2) = 1.28, p = 0.28). However, significant group differences were seen in baseline anxiety between the groups F (2) = 4.21, p = 0.017, with perpetrators reporting significantly lower scores than the combined group of survivors and perpetrators. These findings suggested that except for anxiety, the groups seemed to have similar mental health challenges (Table 3).

Table 3.

One-way ANOVA for comparative tables for outcome variables at baseline across group of only survivors, only perpetrators and mixed group

Mental health concerns	Sum of square	df	Mean square	F	p–value
Anxiety	98.38	2	49.19	4.21	0.017*
Depression	110.62	2	55.31	2.53	0.083
PTSD	23.58	2	11.79	0.885	0.415
Emotional dysregulation	418.55	2	209.28	1.28	0.280
Anger	38.46	2	19.23	1.07	0.345

PTSD Post-traumatic stress disorder

*Significant at p < 0.05

Effects of the CRM intervention on mental health symptoms over time (6 months) across the 3 groups

Our findings showed that repeated-measure ANOVA was used to evaluate the impact of the CRM intervention over 6 months across the 3 groups. The results showed that outcome variables like anxiety, F(1) = 20.17, p < 0.001); depression F(1) = 37.03, p < 0.001; anger F(1) = 95.97, p < 0.001; and emotional dysregulation F(1) = 76.68, p < 0.001 significantly improved across all groups. PTSD symptoms decreased slightly by 6 months but not in statistically significant ways, with F (1) = 0.59, p = 0.44. In sum, the observed changes in mental health variables over time show sustained significant improvements in mental health symptomatology following the implementation of the CRM intervention, across the groups, reinforcing CRMs efficacy in promoting psychological well-being the diverse participant groups (Table 4).

Table 4.

Repeat-measures ANOVA of outcome variables after six months post intervention, across group of survivors, perpetrators and mixed group of survivors and perpetrators

Measure	Times of evaluation	Sum of squares	Df	Mean square	F	P-value	Partial eta squared
Anxiety	Time	228.941	1	228.94	20.17	< 0.001*	0.125
Depression	Time	674.2	1	674.20	37.03	< 0.001*	0.208
PTSD	Time	6.885	1	6.88	0.59	0.441	0.004
Emotional dysregulation	Time	12277.75	1	12277.75	76.68	< 0.001*	0.352
Anger	Time	1467.110	1	1467.11	95.97	< 0.001*	0.405

PTSD Post-traumatic stress disorder; F: F-test; *statistical significance at p < 0.001

Discussion

The purpose of our study was to evaluate the effectiveness of the CRM skills in reducing symptomatology of mental health issues including PTSD, anger, emotional dysregulation, anxiety, and depression among genocide survivors and perpetrators. Our results revealed that they are equally high mental health symptoms such depression, PTSD, Anger and emotional dysregulation in all participated groups of survivors, perpetrators and mixed groups. Importantly, anxiety symptoms were significantly more pronounced among survivors compared to perpetrators or the combined group. Our findings align with many research conducted on mental health in post-genocide whereby the research revealed they are high prevalence of mental health problem in Rwanda [31, 60]. Our research brought other findings while previous study conducted in 2023 maintained that genocide survivors have a significant mental health struggles compared to the overall population, with 35% experiencing depression compared to 12% in the total population, and 29% suffering from PTSD compared to 6% of the total population [22]. Our results indicated no significant differences in depression, PTSD, anger, and emotional dysregulation between the groups of genocide survivors and perpetrators when compared to the combined group.

Our study modality was that offering a CRM intervention to groups of genocide survivors or groups of genocide perpetrators, or a combined group of survivors and perpetrators, would significantly improve their mental health outcome. In our study results found that the mental health outcome variables like anxiety, depression, Anger and emotional dysregulation have significantly improved at 6 months post-intervention in all participated groups: survivors, perpetrators and mixed group of survivors and perpetrators. Our findings are in line with previous community-based intervention studies conducted in Rwanda among survivors and perpetrators which showed all groups exhibited mental health problems and improved when they received the intervention [61, 62]. Also, a recent study on genocide perpetrators held in Rwandan jails (180 men and 122 women), nearly two-thirds of the sample (62.8%) indicated experiencing somatic symptoms such as tingling or numbness; over half (53.2%) stated they were experiencing nausea; 42.9% reported intestinal disturbances; and 42.9% experienced breathing difficulties [63].

Our results revealed that PTSD symptoms decreased more slowly than other outcomes, and the reduction was not statistically significant at six-month post-intervention assessment. Several contextual determinants may help explain these findings. Notably, the six-month follow up conducted February to March 2024 coincided with the period just before the annual Rwandan genocide commemoration that occurs between April and July, a time when instances of trauma re-experiencing are known to increase across the population [64]. Previous research has documented that genocide survivors are particularly vulnerable to PTSD Acute Exacerbations (PAE) during the commemoration period, intensifying symptoms even months after interventions [65]. This seasonal pattern may have significantly affected our PTSD outcomes in the six-month follow-up. Surprisingly, despite persistent PTSD symptoms, participants demonstrated meaningful improvements in other areas of mental health, including emotional regulation, anxiety, depression, and anger, indicating that the intervention had a broader positive impact beyond PTSD alone.

A recent study by Kayiteshonga on trauma cases during the commemoration period of the 1994 genocide found that 92% of women surveyed experienced trauma-related symptoms, and 47% suffered trauma crises as many as 24 times during the three-month commemoration period. This highlights the severe and recurring psychological distress that many individuals endure during this time [66]. Navigating trauma recovery is complex, and sustained healing often requires more than a single set of training sessions. While we did not offer booster sessions in this study, other evidence suggests that follow-up booster sessions can play a critical role in reinforcing and maintaining the gains achieved through initial interventions. For example, prior CRM studies in African countries such as Sierra Leone [45] and recent, yet unpublished evidence from a CRM intervention in Angola demonstrate that participants who received booster sessions showed greater improvements in mental health symptoms over time.

Moreover, the unexpected finding that mixed groups of genocide survivors and perpetrators showed slightly greater improvement in PTSD symptoms and emotional regulation at the six months follow-up compared to the other groups, is particularly noteworthy and merits further exploration. Another possible explanation for the rebound of PTSD relates to the cultural and contextual factors in Rwanda. In the months leading up to the annual genocide commemoration period, there is a strong cultural tradition of collective remembrance and mutual support. During this time, Rwandans prepare to honor the victims and reflect on the events of the 1994 genocide. This may explain the observed increases in PTSD symptoms at the six-month follow-up, despite the significant reductions observed immediately after the intervention across all groups and future studies should be deigned with this in mind.

Strengths and limitations of the study

This study had several strengths. First, it brought together genocide survivors and perpetrators within the same settings for CRM trainings that are a unique approach that stands out from previous studies. This integration demonstrated that, despite a painful history, community members can come together to engage in collective healing and resilience-building. Second, the use of CRM as a biologically based and accessible approach was an important strength. It is an innovative approach that is simple to learn, cost-free, and can be practiced by individuals without clinical supervision, making it particularly suitable for low-resource settings. Third, the longitudinal study design allowed for an appropriate time frame to evaluate the impact of an intervention on mental health symptoms. Conducting baseline testing and the rigor of randomly assigning participants to an intervention and collecting data for 6 months post-intervention allowed us to assess its long-term impact of an intervention within different group modalities. Fourth, the use of standardized and validated instruments for data collection added rigor and credibility to the findings. Another strength of the study was the balanced sample size across the three groups, which enabled meaningful comparisons of the impact of an intervention. Additionally, having recruited 50 participants or more per group provided a sufficient sample size to support reliable conclusions and informed decision-making.

Despite these strengths, we recognize several limitations of this research. First, the study did not utilize traditional control groups or a fully experimental design, participants were randomly assigned to groups and intervention. The original local recruitment and referral process relied on convenience sampling to identify participants from four sectors, which themselves were randomly selected from the seventeen sectors of Nyamagabe district, one of the areas most affected by the genocide. Participants were randomly assigned to a survivor-only group, a perpetrator-only group, or a mixed group comprising both survivors and perpetrators together. We note that this non-random community recruitment strategy was necessary to secure collaboration from local leadership and community support, given the high levels of distrust and anxiety between these groups. In addition, as this was an unfunded pilot feasibility study aimed at determining the most effective teaching methods, our sample size was limited and ideally would have been larger. The study was also conducted in a single district, whereas Rwanda has over 30 districts plus Kigali City. Thus, the results might not be fully generalized at a national level or reflect regional diversity. Moreover, the study did not explore potential differences in intervention impact between rural and urban settings, which limit understanding of how local contexts might influence outcomes. Although eight participants discontinued before the six-month follow-up, a comparison of their pre- and immediate post-intervention outcomes with those of remaining participants revealed no significant differences. This finding suggests that participant attrition did not substantially affect overall results. Last, we did not include extensive qualitative data, which restricted our ability to understand their contextual and structural factors that could influence an impact of our intervention on mental health.

Future directions

We suggest replicating our study with larger numbers of participants, now that we know that we can safely do so as results are just as positive (if not more so) for the combined groups as they are within groups of survivors and perpetrators. Consideration of this design in future studies may enhance the validity and causal inference of the results and may indicate the predictors that influence the outcomes efficacity of an intervention. Additionally, it is important to investigate mechanisms of change in mixed versus separate groups to optimize intervention delivery. Finally, we suggest expanding the measurement design to 1 year and offering half-day boosters to participants based on their feedback and the data pattern. In addition, future research could incorporate qualitative methods to explore the impact of the intervention in greater depth, as well as utilize Perceived Justice Questionnaires to assess participants’ perceptions of fairness and justice more systematically. Lastly, measuring the impact of the CRM intervention is complex and future studies may consider comparing CRM to other evidence-based community interventions was a CRM delivery modality comparison design.

Conclusion

This paper addresses the critical problem of which community-based mental health delivery modalities mostly effectively benefit communities of genocide survivors and perpetrators, who are often co-living after perpetrators were released from Rwanda Correctional Services. This is not unusual countries and communities experience genocides or civil wars. Our findings demonstrate that providing psycho-educational and trauma-informed support, such as the CRM, to genocide survivors and perpetrators is effective in reducing depression, anxiety, anger, and emotional dysregulation. There are many implications for results from our study. There are urgent needs policymakers to highlight the importance of CRM to combined groups for improved mental health throughout the community. Given our impactful results, the Rwandan Ministry of Health and its partners should consider the investment to more broadly offer CRM to mixed groups of affected populations to combat the continuing post-genocide mental health challenges. This intervention could also be a pivotal point for designing mental health policies that look at both groups for prevalence of mental health problems. Besides, the findings from this research could be useful to other post-conflict settings given it potential to help communities reduce mental health symptoms and ability of CRM skills to be delivered effectively in cost effective ways by non-mental health professionals, especially where there is shortage of mental health professionals. Lastly, the future researchers are suggested longitudinal study designs to explore more additional predictors that may impact the impact the effectiveness of CRM on mental health outcomes among genocide survivors, perpetrators, and their descendants.

Author contributions

SH, EB, ZL, KF, SM made substantial contributions to the conceptualization and design of the study, as well as to the critical revision of the manuscript. SH, EB, and ZL conducted the data analysis, interpreted the findings, and contributed to drafting the initial version of the manuscript. SH and SM facilitated access to resources and supported the data collection process. EB and SH were responsible for the administration and for preparing the visualizations of this work. SH took the lead in coordinating data acquisition. SM provided oversight, supervision, and made significant contributions to the review and editing of the manuscript. All authors reviewed and approved the final version of the manuscript for submission.

Funding

The authors report no funding for this study.

Data availability

The dataset analyzed is available from the corresponding author upon reasonable request.

Declarations

Ethical approval

Ethical clearance for this research was obtained from the Ethics Committee of the College of Medicine and Health Sciences of the University of Rwanda (N0: 106/CMHS IRB/2023). In addition, authorisation to commence data collection was obtained from the Rwanda Governance Board through the respective district authorities, as required by national regulations.

Consent for publication

The consent for publication was obtained from all study participants.

Consent to participate

Before data collection, the researchers informed participants about the study objectives, procedures, potential benefits, and their rights to withdraw at any time without providing any reason. They were also informed that there no negative consequences resulting from such decision. After receiving this information, all participants gave both oral and written consent to participate. Confidentiality was strictly kept, and all data were gathered anonymously.

Competing interests

The authors declare no competing interests.