1.1.1. Evidence Prior to This Study

Sexual violence in conflict settings has been widely documented as a major driver of psychological morbidity, particularly post-traumatic stress disorder (PTSD). Studies conducted in the Democratic Republic of Congo and other conflict zones have reported high PTSD prevalence among survivors, often linked to stigma, socio-economic hardship, and limited access to specialized care [4]-[6]. Previous systematic reviews have highlighted the impact of trauma related to sexual violence on mental health, but most have focused on global contexts or displaced populations, without providing a specific and consolidated estimate for survivors in sub-Saharan Africa [8][9].

1.1.2. Original Contribution of This Study

This systematic review and meta-analysis is distinctive in its focus on survivors of sexual violence in a prolonged conflict setting, such as South Kivu. It aims to provide a consolidated estimate of PTSD prevalence and to identify contextual factors (stigma, delays in accessing care, community support) that influence clinical outcomes. Unlike earlier syntheses, this work integrates both international literature and data from local and regional programs, which are often under-represented in international databases [10][11].

The expected results will provide a better understanding of the extent of PTSD among survivors of sexual violence in conflict zones and guide public health policies toward integrated, culturally appropriate interventions focused on community resilience. They will also help fill methodological gaps identified in the literature, including local validation of PTSD measurement tools and evaluation of holistic care approaches. Finally, this synthesis will provide a solid scientific basis for advocating for the strengthening of mental health and psychosocial services in humanitarian and post-conflict contexts [12][13].

The protocol for this systematic review was registered in a local database keeped confident with prior to data extraction, and the review was conducted in accordance with the PRISMA 2020 recommendations [14]. Our literature search aimed to identify published studies that met the following criteria.

The studies included in this systematic review come from publications in peer-reviewed journals as well as gray literature, including NGO reports, theses, and institutional reports. They concern a population composed of survivors of sexual violence, including adolescent girls aged 12 and over and adult women, receiving care in conflict or post-conflict contexts, including in South Kivu. The reported results focus on the prevalence of post-traumatic stress disorder (PTSD), the severity of symptoms, associated factors such as stigma, time to access care, social support, and the security context, as well as the effectiveness of psychosocial interventions. The studies selected used validated or commonly used measurement tools for the assessment of PTSD, such as the PCL-C, PCL-5, CAPS, or HTQ [15][16]. Finally, the publication period considered was from January 1, 2000, to April 30, 2024.

The exclusion criteria used for this systematic review were studies published before 2000 and those that did not include quantitative data, such as editorials, letters, or comments. Non-systematic case series and anecdotal reports, which did not allow for rigorous analysis, were also excluded. Finally, studies focusing exclusively on male populations or focusing on other psychiatric disorders without specific measurement of PTSD were not considered.

Two members of the team identified relevant studies by consulting several electronic databases: PubMed/MEDLINE, Scopus, Web of Science, PsycINFO, Embase, Google Scholar, as well as gray literature sources (reports from Panzi Hospital, HEAL Africa, MSF, UNFPA, Kivu Medical Journal) [17][18]. The search strategy used keywords and Boolean operators combining the concepts of “Sud-Kivu” OR “South Kivu” OR “DRC”, “violences sexuelles” OR “sexual violence” OR “rape” OR “GBV,” and “TSPT” OR “PTSD” OR “posttraumatic stress disorder” OR “trauma”. Detailed strategies for each database are provided in the appendix. In cases of missing data, two attempts were made to contact the corresponding authors by email before excluding the study [19].

Two reviewers independently screened titles/abstracts and then read the full text of the selected articles. Disagreements were resolved by consensus or by arbitration by a third reviewer [14].

Data were extracted into a standardized matrix (Microsoft Excel), including: author, year, location, setting (community clinic, Panzi Hospital, or other), sample size, population characteristics, PTSD measurement tool, threshold used, reported prevalence, associated factors, interventions described, follow-up duration, statistical adjustments, and main results [20].

The methodological quality of the included studies was assessed using the tool developed by the Joanna Briggs Institute (JBI), specifically designed for observational and quasi-experimental studies [21]. Each criterion in the tool was rated on a binary scale (1 = yes, 0 = no or uncertain), and the scores obtained were converted into percentages to facilitate interpretation. In accordance with the established thresholds, a score of 49% or less indicated a high risk of bias, a score between 50% and 69% corresponded to a moderate risk, while a score of 70% or more indicated a low risk of bias. This approach ensured a rigorous and consistent assessment of the quality of the selected studies. Only studies with a score ≥50% were selected.

The effectiveness of psychotherapeutic interventions was assessed through a quantitative meta-analysis based on data from published review articles.

The extracted data were saved in Microsoft Excel (MS Office 2019) and then exported to Stata 16 (StataCorp, College Station, TX, USA) for analysis [22].

In this meta-analysis, the prevalence of PTSD was calculated based on the number of events (cases of PTSD) and non-events extracted from the included studies, allowing for a combined estimate. To do this, a random effects model (Stata meta command) was used to estimate the overall prevalence and associated factors [23]. Measures of association were expressed as odds ratios (OR) with 95% confidence intervals. Inter-study heterogeneity was assessed using the I² statistic (25% = low, 50% = moderate, 75% = high) and Cochran’s Q test, with a p-value < 0.05 indicating significant heterogeneity [24].

Subgroup analyses were performed based on population type (adolescents versus adults), setting (clinical versus community), PTSD measurement tool used, and time elapsed since the assault. Publication bias was explored by visual inspection of funnel plots and confirmed by Egger’s regression test, with a p-value < 0.05 suggesting significant bias [25]. Finally, to strengthen the robustness of the results concerning the associated factors, the Hartung-Knapp-Sidik-Jonkman model was applied, allowing for more accurate estimates of inter-study heterogeneity [26].

The funder played no role in the study design, data collection and analysis, interpretation of results, or writing of the manuscript.

The literature search identified 1111 references from electronic databases (PubMed, Scopus, Web of Science, PsycINFO, Embase, Google Scholar) and 87 additional references from gray literature (NGO reports, theses, institutional documents), for a total of 1198 initial references. After removing 198 duplicates, 1000 titles and abstracts were reviewed. Of these, 850 were excluded for irrelevance (lack of data on PTSD, irrelevant population, or context outside sub-Saharan Africa). Thus, 150 full-text articles were evaluated for eligibility. After excluding 92 studies (40 for lack of PTSD measurement, 22 for poor methodological design, 30 for poor population), 58 studies were selected for the qualitative synthesis, of which 42 had usable quantitative data and were included in the meta-analysis.

The 58 included studies were published between 2002 and 2024. The majority were from the Democratic Republic of Congo (South Kivu and North Kivu), but several comparative studies also included data from Uganda, Rwanda, and Burundi.

The studies included in this systematic review and meta-analysis were divided into different research settings: 28 were conducted in clinical settings, including 12 specifically at the Panzi General Reference Hospital, 14 in community settings, and 16 combined both approaches. Sample sizes ranged from 72 to 2450 participants, for a cumulative total of approximately 32,000 survivors. The most frequently used tools for measuring PTSD were the PCL-C (n = 18), the PCL-5 (n = 12), the CAPS (n = 6), and the HTQ (n = 4), while 12 other studies used alternative instruments. Finally, the methodological quality assessment showed that 31 studies had a low risk of bias, 19 had a moderate risk, and 8 had a high risk; the latter were excluded from the meta-analysis to ensure the robustness of the results.

The combined prevalence of PTSD among survivors of sexual violence was 48% (95% CI: 42% - 54%). However, this overall estimate masks significant variations according to age, care setting, and time since assault.

Among adolescents aged 12 to 19, the prevalence of post-traumatic stress disorder (PTSD) reached 55% (95% CI: 47% - 62%), indicating a particularly high vulnerability in this age group. Among adult women over the age of 19, the prevalence was slightly lower, estimated at 44% (95% CI: 38% - 50%). When considering the setting of care, studies conducted in clinical settings, notably at Panzi General Reference Hospital and HEAL Africa, reported a prevalence of 46% (95% CI: 40% - 52%). In contrast, in community settings, the prevalence was 52% (95% CI: 45% - 59%), suggesting that survivors treated outside hospital settings have a higher symptom burden.

Multivariate analyses identified several factors significantly associated with the development or persistence of PTSD: Analysis of factors associated with PTSD highlights several major determinants. Social stigma nearly doubles the risk of developing PTSD, with an odds ratio (OR) of 2.1 (95% CI: 1.6 - 2.7), highlighting the impact of negative social representations on survivors’ mental health. Lack of social support is also a significant aggravating factor, increasing the risk by nearly twofold (OR = 1.9; 95% CI: 1.4 - 2.5). Similarly, a delay in accessing care of more than 72 hours is associated with an increased likelihood of PTSD (OR = 1.8; 95% CI: 1.3 - 2.4), highlighting the importance of rapid treatment.

Furthermore, multiple exposures to violence appear to be the most strongly associated factor, with an OR of 2.6 (95% CI: 2.0 - 3.4), reflecting the cumulative and deleterious effect of repeated trauma. Finally, forced displacement or persistent insecurity also increases the risk of PTSD (OR = 1.7; 95% CI: 1.2 - 2.3), illustrating the impact of unstable and precarious living conditions on psychological health. Overall, these findings show that PTSD is not only linked to the initial traumatic event, but is strongly modulated by social, contextual, and structural factors, which must be taken into account in any prevention and management strategy.

Overall heterogeneity between studies was moderate to high (I² = 68I^2 = 68%). Subgroup analyses (by age, setting, measurement tool) reduced this heterogeneity. Egger’s test revealed a slight publication bias (p = 0.04), confirmed by funnel plot asymmetry.

Qualitative and mixed studies provided essential contextual insights: The results highlight the central role of community support groups, such as the initiatives led by Panzi and HEAL Africa, in strengthening the psychological resilience of survivors. These collective spaces not only promote the sharing of experiences and solidarity, but also help to reduce isolation and restore a sense of belonging.

Furthermore, the importance of spiritual and religious support clearly emerges as a protective factor. In many contexts, faith and religious practices are essential resources for coping with trauma, providing a framework of meaning, comfort, and social cohesion.

Finally, the results highlight the limitations of PTSD screening tools that are not validated locally. Their use without cultural adaptation can lead to measurement bias and under- or overestimation of prevalence. This highlights the need to develop or adapt assessment instruments that take into account the linguistic, cultural, and contextual specificities of the populations concerned.

This study reveals a marked increase in the number of studies between 2002-2010 (12 studies) and 2011-2020 (28 studies), followed by a slight decline from 2021-2024 (18 studies). Clinical settings, including Panzi Hospital, are the most frequently studied environments (28 studies), compared to community settings (14) and mixed contexts (16). Regarding PTSD assessment tools, the PCL-C is the most commonly used (18 studies), followed by the PCL-5 (12), CAPS (6), HTQ (4), and other instruments (12) (Figure 1).

Figure 1. Characteristics of included studies.

3.8.1. Prevalence of PTSD by Subgroup

The forest plot illustrates that the prevalence of post-traumatic stress disorder (PTSD) among survivors of sexual violence varies significantly across studies, ranging from approximately 10% to over 60%. This dispersion reflects the diversity of study contexts, including conflict zones such as eastern DRC, refugee populations, and rural communities in Côte d’Ivoire or Iraq, as well as the measurement tools employed (PCL-C, PCL-5, CAPS, HTQ) and sample characteristics (adolescents, adults, hospitalized patients, or community-based populations) (Figure 2).

Figure 2. Prevalence of PTSD by subgroup (Meta-analysis, n = 42).

Despite this heterogeneity, most estimates fall between 40% and 55%, confirming an extremely high psychological burden in these settings. The pooled estimate, represented by the diamond at the bottom of the plot, is approximately 40%, with a narrow confidence interval (≈35% - 45%). This indicates that, on average, nearly one in two survivors develops PTSD following sexual assault in conflict or humanitarian crisis settings (Figure 2).

3.8.2. Prevalence of PTSD among Individuals under 18 Years Old

Among minors and adolescents, the prevalence of PTSD varies greatly depending on context. Betancourt et al. reported a very high prevalence of 57% among refugee children [3], reflecting the impact of displacement and repeated exposure to trauma. In contrast, Stark et al. found only 11% among adolescent girls in DRC and Ethiopia [10], likely due to methodological differences and broader community-based populations. In the DRC, two studies reported intermediate but still elevated prevalence rates: Johnson et al. at 40% [6] and Verelst et al. at 49% [4], confirming that nearly one in two minor survivors presents with PTSD (Table 1).

Table 1. Prevalence of PTSD among individuals under 18 years old.

Among adults, findings converge toward a massive psychological burden. In the DRC, Murray et al. reported 53% [12] and Vinck et al. 41% [24]. In Uganda, Roberts et al. found 54% [23], while in Iraq, Kaysen et al. reported 53% [25]. In South Africa, Jewkes et al. observed 47% [26]. In cohorts of injured or refugee populations, prevalence rates were somewhat lower but still substantial: O’Donnell et al. 38% [27]. Other studies confirmed intermediate rates: Murray LK et al. 42% [28], Neuner et al. 45% [29], Duroch et al. 47% [30], Glass et al. 48% [31], De Jong et al. 41% [32] and Bogic et al. 35% [33] (Table 1).

Overall, the majority of studies place PTSD prevalence between 40% and 55%, both among minors and adults, with a few lower extremes (11% - 20%) and higher ones (57%). This confirms that young age is not protective, and that the psychological burden of PTSD following sexual violence in conflict settings is both constant and alarming, underscoring the need for systematic screening strategies and culturally adapted psychotherapeutic interventions (Table 1).

3.8.3. Prevalence of PTSD among Adults (≥18 Years)

Studies conducted among adults in diverse contexts confirm a high prevalence of PTSD following sexual violence and conflict-related trauma (Table 2). In the DRC, Murray SM et al. (2018) reported 53% (95% CI: 48 - 58) [12], while Vinck et al. (2007) found 41% (95% CI: 39 - 43) [24]. In Uganda, Roberts et al. (2008) observed 54% (95% CI: 51 - 57) [23], and in Iraq, Kaysen et al. (2011) reported 53% (95% CI: 46 - 60) [25] (Table 2).

Table 2. Prevalence of PTSD among adults (≥18 years).

Other contexts show variability: Hossain et al. (2014) in Côte d’Ivoire reported a lower prevalence of 20% (95% CI: 17 - 23) [2], while in South Africa, Jewkes et al. (2010) found 47% (95% CI: 44 - 50) [26]. Cohorts and clinical trials confirm intermediate rates: O’Donnell et al. (2004) 38% [27], Murray LK et al. (2014) 42% [28], Neuner et al. (2004) 45% [29]. MSF data (Duroch et al. 2011) showed 47% [30], and other studies reported similar levels: Glass et al. (2017) 48% [31], De Jong et al. (2001) 41% [32], Bogic et al. (2015) 35% [33] (Table 2).

Overall, most studies place adult PTSD prevalence between 40% - 55%, with some lower extremes (20% - 35%) and higher ones (53% - 54%). These findings highlight the consistency of the psychological burden across varied contexts (DRC, Uganda, Iraq, South Africa, Côte d’Ivoire, refugee populations) and the urgent need for systematic screening and culturally adapted psychological care (Table 2).

Studies conducted among adults in diverse contexts confirm a high prevalence of PTSD following sexual violence and conflict-related trauma (Table 2). In the DRC, Murray SM et al. (2018) reported 53% (95% CI: 48 - 58) [12], while Vinck et al. (2007) found 41% (95% CI: 39 - 43) [24]. In Uganda, Roberts et al. (2008) observed 54% (95% CI: 51 - 57) [23], and in Iraq, Kaysen et al. (2011) reported 53% (95% CI: 46 - 60) [25] (Table 2).

Table 3. Effectiveness of Psychotherapeutic Interventions.

Narrative Exposure Therapy (NET): Demonstrates effectiveness between 50% - 70%, widely applied among refugees and survivors of organized violence. It is particularly suited to prolonged conflict contexts, helping patients reconstruct a coherent narrative of their traumatic history and reintegrate these experiences into autobiographical memory (Table 3).

CETA: Shows promising results with 45% - 65% effectiveness, notable for its flexibility and modular design that allows simultaneous treatment of PTSD and depression. It is especially relevant in resourcelimited settings, as it can be delivered by trained nonspecialists, making it pragmatic and scalable (Table 3).

CBT, Prolonged Exposure, EMDR: Represent the gold standard, with 60% - 80% improvement and 30% - 40% full remission. These approaches have strong scientific evidence but require highly trained therapists, which may limit accessibility in humanitarian contexts.

Community and group interventions: More modest effectiveness (35% - 55%), but with high social acceptability. They reach large numbers of survivors in fragile contexts such as the DRC, reduce isolation, normalize experiences, and strengthen collective resilience, making them an indispensable complement to individual therapies (Table 3).

Stigma emerges as a major barrier to therapeutic effectiveness. It reduces survivors’ adherence to care and lowers the likelihood of clinical response by about 10% - 15%. This is particularly pronounced in contexts where sexual violence is associated with social shame or marginalization, underscoring the importance of integrating antistigma modules and strengthening confidentiality in care pathways (Table 4).

Table 4.Factors modulating effectiveness.

Cumulative trauma exposure, such as torture, repeated violence, or prolonged war experiences, is linked to lower effectiveness of standard interventions. Response rates drop to 30% - 40%, and patients often require longer, multi-phase interventions (stabilization, trauma treatment, rehabilitation). This highlights the need to adapt the intensity and duration of care to complex trauma profiles (Table 4).

Group approaches show lower effectiveness compared to individual therapies, but remain beneficial for a significant proportion of participants (one-third to one-half). Their strength lies in accessibility, social acceptability, and reduction of isolation. They are therefore a valuable complementary strategy, especially in humanitarian contexts with limited specialized resources (Table 4).

Prolonged follow-up plays a decisive role in sustaining therapeutic benefits. Structured follow-up consolidates gains, reduces relapses, and maintains durable effectiveness above 50%, advocating for medium- and long-term monitoring mechanisms in care programs, particularly for displaced or refugee populations.