Post-traumatic stress disorder among survivors of sexual violence in a hospital in Eastern DR Congo: Predictive factors, protectors, and comorbidities

Study setting and population

This prospective analytical study included all women victims of sexual violence aged 16 years and older who were admitted between September and December 2023 and who agreed to participate by signing an informed consent form. For participants under 18 years of age, consent was additionally obtained from their parent or guardian. Patients presenting with language dissociative disorders at admission were not included, and any patient who did not meet the inclusion criteria defined at admission was excluded.

A comprehensive, non-random volunteer sampling method was used including 312 women. Specifically, all women who met the inclusion criteria and sought care at the center during the study period were invited to participate. This approach aimed to include as many eligible individuals as possible, without random selection. Participation was voluntary, and individuals were informed about the study before consenting. Of the total eligible women, 12 declined to participate, citing reasons such as emotional distress, lack of time, and displacement caused by insecurity and the inaccessibility of residential areas due to the effects of armed conflict.

Procedure

Participants were recruited during their visit to the center. Participants were selected using a comprehensive, non-random volunteer sampling approach. All women who met the inclusion criteria, namely, having experienced sexual violence and seeking care at the center during the study period, were invited to participate. Recruitment was conducted on-site by trained staff who explained the study and obtained informed consent. Participation was entirely voluntary, and individuals were free to decline or withdraw at any time without affecting their access to care. A research team member explained the study’s objectives and obtained informed consent from the participants, before administering the questionnaire in a face-to-face format by trained interviewers in a private space that ensured confidentiality. In addition, participants were informed that they could receive the study’s results and recommendations derived from them once the research was completed.

Data collection spanned 4 months and was conducted in two phases. The first phase, conducted upon admission, involved identifying the participants, gathering their sociodemographic information, and carrying out a psychometric evaluation using the PTSD scale. Simultaneously, the assessment included administering the hospital anxiety and depression scale to detect anxiety-depressive comorbidities associated with PTSD.

Data entry and analysis

Statistical analyses were performed using the Statistical Package for the Social Sciences Version 27. The sociodemographic and clinical characteristics of the patients were summarized using descriptive statistics such as percentages, means, and standard deviations. A bivariate analysis, complemented by a multivariate logistic regression model, was employed to identify potential risk factors associated with anxiety-depressive disorders and to establish links between sociodemographic variables, histories of violence, and the presence of PTSD symptoms (assessed during the first and second measurements). Pearson’s correlation tests and the Chi-square test were used to explore the relationships between continuous and categorical variables. Variables with P < 0.20 in the bivariate analysis were incorporated into the multivariate model, where the associations were expressed by the adjusted odds ratio (aOR) with a 95% confidence interval (CI). The threshold for statistical significance was set at P < 0.05 in the final model.

Ethical considerations

This study was approved by the Medical Ethics Committee of the University of Goma (Approval N° UNIGOM/CEM/0014/2023). Free and informed consent was obtained both orally and in writing from each participant after a brief explanation of the study. Participants were informed of their right to refuse or withdraw at any time without consequences. Personal identifiers such as names and phone numbers were not recorded, and all data were kept confidential and used solely for research purposes.

Given the conflict-affected context marked by ongoing security challenges, participants who showed signs of psychological distress or expressed a need for support received care within the psychiatry unit, where psychosocial support was integrated into their overall treatment.

Peri-traumatic determinants

Regarding physical injuries, 218 patients without any injury exhibited a PTSD rate of 57.8%, compared to 94 patients who had sustained injuries, with a PTSD rate of 42.6% (P = 0.013). The presence of an sexually transmitted infection (STD) did not show a significant difference (P = 0.055), even though 60.2% of individuals testing positive for an STD experienced PTSD, compared to 49.0% of those testing negative. With respect to the number of perpetrators, incidents involving multiple perpetrators – whether 2–5 (n = 130) or more than 5 (n = 36) – demonstrated higher PTSD percentages (61.5% and 61.1%, respectively) compared to cases with a single perpetrator, which showed a PTSD rate of 43.8% (n = 146, P = 0.008). The variables “collective violence,” “public violence,” and “number of rapes” did not reach statistical significance.

In addition, the variable concerning the loss of material goods proved significant (P = 0.019*): 58.9% of those without any loss experienced PTSD, compared to 45.5% among those who suffered a loss, whereas the involvement of a perpetrator in uniform showed no significant association (P = 0.869) [Table 2].

Post-traumatic determinants

For alcohol and tobacco consumption, no statistically significant differences were observed in the occurrence of PTSD. Specifically, 51.5% of non-alcohol users experienced PTSD compared to 57.6% of alcohol users (P = 0.336); regarding tobacco, 52.2% of non-smokers had PTSD versus 73.3% of smokers (P = 0.109). The time interval between receiving care and the rape did not significantly influence the risk of PTSD: 55.3% of individuals with an interval of 72 h or less experienced PTSD compared to 52.9% for those with an interval longer than 72 h (P = 0.786). Furthermore, being pregnant appears to be associated with a higher rate of PTSD, with 63.0% of pregnant women affected versus 50.2% of non-pregnant women – a difference that approached significance (P = 0.055). Finally, stigma was not linked to a notable difference in the occurrence of PTSD, as 51.9% of women who reported experiencing stigma had PTSD compared to 48.1% among those who did not (P = 0.799) [Table 3].

Multivariate logistic regression results highlight that age is a significant determinant of PTSD. In comparison to the reference group (18–64 years), individuals under 18 years have a reduced risk of PTSD (aOR = 0.335, 95% CI: 0.119–0.945, P = 0.039), while those aged 65 and over exhibit an even more marked reduction (aOR = 0.044, 95% CI: 0.005–0.408, P = 0.006) [Table 4].

Educational level also plays a crucial role: Lacking any formal education (“None”) is associated with a significant increase in risk (aOR = 2.441, 95% CI: 1.144–5.209, P = 0.021) when compared to individuals with secondary-level education. Furthermore, the occupation of cultivator shows a statistically significant association with a lower PTSD risk (aOR = 0.318, 95% CI: 0.125–0.810, P = 0.016), whereas the estimates for the “Housewife” category were aberrant.

Tobacco use is linked to a substantially higher risk of developing PTSD (aOR = 4.208, 95% CI: 1.035–17.109, P = 0.045), underscoring how lifestyle behaviors can influence psychological vulnerability [Table 4].

Moreover, pregnancy emerges as a strong risk factor, with pregnant women having a significantly increased risk of PTSD (aOR = 3.497, 95% CI: 1.646–7.430, P = 0.001). Similarly, exposure to a greater number of perpetrators also significantly elevates the risk – participants exposed to 2–5 perpetrators have an aOR of 2.395 (95% CI: 1.274–4.505, P = 0.007), and those exposed to 5 or more perpetrators have an aOR of 3.018 (95% CI: 1.192–7.642, P = 0.020). These results illustrate the complex interplay of demographic and socio-economic factors that influence the risk of developing PTSD [Table 4].

Comorbidities, social support, and self-esteem

Table overview: Comorbidities, social support, and self-esteem

This table presents Pearson correlations between various psychological variables measured at 2 time points (Moment 1 and Moment 2). The strongest relationships are found between PTSD at Moment 1 and PTSD at Moment 2 (r = 0.723; P < 0.001) and between Anxiety 1 and Depression 1 (r = 0.662; P < 0.001), pointing to a strong stability and interconnection of symptoms over time. Moderate correlations were also observed between Anxiety 1 and Anxiety 2 (r = 0.644) and between Depression 1 and Depression 2 (r = 0.617), all significant at P < 0.001, which underscores the persistence of these disorders as time goes on. In addition, satisfaction with social support is negatively correlated with Anxiety 2 (r = –0.156; P = 0.006) and Depression 2 (r = –0.123; P = 0.029), while the availability of support is positively linked with both satisfaction (r = 0.480; P < 0.001) and self-esteem (r = 0.182;P = 0.001). Finally, self-esteem is significantly correlated with the perceived satisfaction of social support (r = 0.466; P < 0.001), but it does not show significant correlations with the levels of PTSD, anxiety, or depression at either Moment 1 or Moment 2. Together, these associations reveal important links between post-traumatic symptoms, affective states, and the perceived role of social support in modulating these symptoms [Table 5].

Sociodemographic determinants

The findings of this study reveal a statistically significant association between age and the occurrence of PTSD (P = 0.000). Notably, individuals aged 65 and older show a reduced risk compared to those under 65. These results are in line with Vaillant-Ciszewics et al., who noted that PTSD is rarely diagnosed among older individuals, often due to underrecognition,^[6] and with Goldstein et al., who reported a higher prevalence of PTSD among younger adults.^[7]

In our study, variables relating to living environment and marital status did not show significant associations with PTSD (P = 0.385 and P = 0.115, respectively). These findings contrast with those of Goldstein et al., who observed lower rates of PTSD in urban areas compared to rural settings,^[7] and with Feki et al., who reported that marital status can be linked to an increased risk of PTSD.^[3]

A significant relationship was also observed with education level (P = 0.018), as individuals without formal education had higher rates of PTSD. This is consistent with research by Goldstein et al., Ju et al., and Feki et al., all of whom highlight low educational attainment as a key vulnerability factor for developing PTSD.^[3,4,7] Finally, occupation emerged as an associated factor (P = 0.024), with higher PTSD rates among saleswomen and cultivators. Feki et al. similarly observed that most individuals with PTSD had been professionally active before the traumatic event.^[3]

Peri-and Post-traumatic determinants

Our data suggest that the absence of physical injuries is statistically associated with a higher prevalence of PTSD (P = 0.013). This may be related to phenomena of reviviscence or post-traumatic flashbacks: Each time the victim sees the scars from their injuries, it reactivates the memory of the event, making the trauma persist over time. This finding aligns with Schmitt et al., who noted that victims without visible injuries might develop more severe PTSD – possibly due to a lack of validation of their psychological suffering.^[8] Moreover, the number of perpetrators emerges as a significant factor, with the risk of PTSD increasing in proportion to the number of aggressors involved. This observation is supported by Dokkedahl et al., who emphasized that repeated exposure to multiple aggressors intensifies the severity and chronicity of PTSD symptoms.^[9]

Conversely, some factors typically examined in the literature did not reach significance in our sample. For instance, neither alcohol nor tobacco consumption showed a clear statistical association with PTSD, contrary to Fu et al., who linked addictive behaviors after trauma to greater PTSD severity.^[10] Similarly, the time interval between the assault and the receipt of care was not a determining factor (P = 0.786), despite studies like Kearney et al. suggesting that a prompt medical and psychological response can mitigate symptom severity.^[11]

Finally, factors such as being pregnant at the time of the assault (P = 0.055), experiencing stigma, the number of rapes endured, and the presence of a uniformed perpetrator were not significantly associated with PTSD in our analysis. These factors may not be significantly associated due to cultural resilience mechanisms, underreporting, or limited statistical power. Stigma and perpetrator identity may have complex, context-specific psychological impacts that are difficult to measure. These results differ from those of Mitchell et al.), who reported higher PTSD rates among pregnant women following sexual assault,^[12] and from Bryant et al., who identified social stigma as a significant predictor of post-traumatic distress.^[13] The lack of significant findings here might be explained by unexamined moderating effects or differences in the cultural, social, and security contexts of our sample.

Multivariate logistic regression

It emerges from this multivariate logistic regression that several factors significantly influence the risk of developing PTSD. Survivors aged between 18 and 64, those with no schooling (aOR: 2.441 [1.144–5.209]; P = 0.021), saleswomen, tobacco users (AOR: 4.208 [1.035–17.109]; P = 0.045), pregnant individuals (AOR: 3.497 [1.646–7.430]; P = 0.001), and those exposed to an increasing number of aggressors – 2–5 aggressors (aOR: 2.395 [1.274–4.505]; P = 0.007) and more than 5 aggressors (AOR: 3.018 [1.192–7.642]; P = 0.020) – have a higher risk of PTSD.

Conversely, certain factors appear to offer protection against PTSD in rape survivors. Being under 18 and being 65 years or older are protective, with aORs of 0.335 (95% CI: 0.119–0.945; P = 0.039) and 0.044 (95% CI: 0.005–0.408; P = 0.006), respectively. In addition, having a secondary level of education, working as a cultivator (aOR: 0.318 [0.125–0.810]; P = 0.016), not smoking, and not being pregnant seem to lower the likelihood of PTSD. In their studies, Feki et al. and Goldstein et al. found that PTSD was significantly linked to low educational levels, being female, substance use, and urban residence.^[3,7]

Comorbidity and the role of social support

Regarding comorbidity, there is a strong linear correlation between PTSD measured at the first and second evaluations (r = 0.723, P < 0.001). Similarly, Anxiety 1 and Depression 1 are significantly correlated with PTSD at the second evaluation (r = 0.407, P < 0.001 and r = 0.304, P < 0.001, respectively), indicating that the findings at the second evaluation vary according to those at the first. In other words, the presence of PTSD, anxiety, and depression initially may contribute to the persistence of PTSD later on.

Moreover, the results of PTSD at the first evaluation vary with those of Anxiety 1 and Depression 1 (r = 0.614, P < 0.001 and r = 0.498, P < 0.001, respectively), and vice versa. The levels of Anxiety 2 and Depression 2 also depend on PTSD at time 1 (r = 0.452, P < 0.001 and r = 0.370, P < 0.001), suggesting that earlier PTSD may foster later anxiety and depression. In addition, both Anxiety 2 and Depression 2 are significantly associated with PTSD at the second evaluation (r = 0.507, P < 0.001 and r = 0.462, P < 0.001). Goldstein et al.^[7] also found that PTSD coexists with mood disorders, anxiety disorders, borderline personality disorder, schizophrenia, and disability.

For social support and self-esteem, our data show that satisfaction with social support is negatively correlated with Anxiety 2 and Depression 2 (r = –0.156, P = 0.006 and r = –0.123, P = 0.029). This means that decreases in satisfaction with social support are associated with increases in Anxiety 2 and Depression 2. Moreover, the availability of social support is positively correlated with both satisfaction (r = 0.480, P < 0.001) and self-esteem (r = 0.466, P < 0.001), and self-esteem is also positively linked with availability (r = 0.182, P = 0.0001).

Self-confidence is a major personal asset in resisting PTSD, and its effectiveness is enhanced by a solid network of social support. This underscores why it is so important to apply self-affirmation techniques within the 30 days following a sexual assault.

Traumatology research highlights that self-esteem is a key individual factor in a person’s resilience against PTSD. High self-esteem not only enables individuals to appreciate their own resources but also helps them develop a more nuanced understanding of the traumatic experience, leading to better emotional regulation.^[14] A dynamic perspective suggests that this sense of personal worth is most effective when accompanied by quality social support, which reinforces and consolidates one’s ability to adapt in the aftermath of trauma.

Social support plays a catalytic role in resilience by facilitating the sharing of experiences, offering a non-judgmental ear, and validating the emotions felt after the assault.^[15] In this context, employing self-affirmation techniques within the first 30 days after the assault could quickly bolster self-esteem. By promoting the expression of personal needs and the establishment of healthy boundaries, self-affirmation serves as an early intervention that may help prevent the consolidation of PTSD.^[16] Such techniques, when integrated into a comprehensive care framework, put the survivor at the center of their psychological recovery, combining inner strength with an external network of support.

This integrated approach – enhancing self-esteem while optimizing social support – appears promising for fostering positive adaptation after a sexual assault. Recovery ultimately depends on an individual’s ability to tap into their own inner resources while also benefiting from a supportive and nurturing environment. Early intervention with self-affirmation techniques during the 1^st weeks following the assault may therefore be a vital lever in preventing the development of chronic PTSD symptoms.

Limitations of the study

This study has several limitations. The use of voluntary sampling may introduce selection bias, excluding the most traumatized or isolated individuals. Self-reported data are subject to recall and underreporting biases, especially for sensitive experiences. Key contextual factors such as family support and access to care were not systematically assessed. In addition, security challenges, mass displacement, and regional instability prevented some victims from reaching the hospital, leading to loss to follow-up and participant dropout. These factors may affect the representativeness and generalizability of the findings.