Background
Sexual violence, whether perpetrated by militarized actors or non-combatants, is prevalent in eastern Democratic Republic of Congo (DRC) and impacts the psychosocial well-being and mental health of survivors [
1‐
4]. In eastern DRC, up to one in five sexual violence survivors has reported a sexual violence-related pregnancy (SVRP) [
3,
5,
6]. Women with SVRPs may be at greater risk of complex psychosocial sequelae [
5,
7‐
10]; however, few studies focus on this subgroup of sexual violence survivors.
In previous studies on sexual violence in eastern DRC, survivors have reported symptoms of depression, post-traumatic stress disorder (PTSD), anxiety, and suicidality [
3,
11,
12]. Survivors have also described social stigmatization, exclusion, and rejection as a result of sexual violence [
7,
8,
13]. Social stigmatization has been noted to be a mediating factor for mental health outcomes among survivors [
14,
15], highlighting the influence of social factors on mental health. Previous research has demonstrated that social support is important in post-trauma adjustment after sexual violence [
16,
17]; as such, negative social reactions may impede post-traumatic growth [
18]. While negative social reactions toward sexual violence survivors, such as stigmatization, have been shown to influence mental health and severity of trauma-related symptoms [
19‐
21], there are limited data on psychosocial outcomes among women with SVRPs.
Previous quantitative studies of mental health and psychosocial consequences provide a platform from which to appreciate the complexities of assessing mental health on a global level [
22,
23]. However, there are noted limitations of these assessments including limited cultural validity and constraints on discerning and inclusion of other psychosocial factors that may influence mental health. Furthermore, mental health assessments conducted post-trauma are often focused on the traumatic event, such as sexual violence, and fewer assess how the social environment influences psychological well-being following the trauma [
19,
20]. Qualitative data can provide further understanding of individuals’ experiences and inform mental health and psychosocial support interventions for sexual violence survivors in conflict settings and women with SVRPs [
12,
24‐
26].
Finally, further understanding of the experiences of women with SVRPs may also inform broader reproductive health programs and policies. Reproductive options for women with SVRPs are limited in DRC, where pregnancy termination, even in the context of sexual violence, is criminalized [
27]. There is a growing focus on children born from SVRPs [
28] and understanding the experiences of their mothers may add further dimension to the dialogue on this issue. As part of a larger mixed methods study of women with SVRPs in eastern DRC [
29], a qualitative study was conducted to describe their emotional responses, needs, and psychosocial outcomes.
Methods
Study design
A mixed methods study was conducted in October–November 2012 in Bukavu, DRC. The study applied quantitative and qualitative methods to assess outcomes among adult women (18 years and older) who self-identified as survivors of sexual violence since the start of the war (~ 1996) and conceived pregnancies as a result of sexual violence. There were two study groups: (1) women currently raising children from SVRPs (parenting group) and (2) women who had terminated SVRPs (termination group). Recruitment and participation in study groups was not mutually exclusive, such that women in the parenting group could recruit women who had terminated a pregnancy and vice versa, and also to recognize that some participants may have had more than one lifetime SVRP and had different outcomes from those SVRPs. Previous publications from this study present quantitative data on mental health and parenting outcomes [
30,
31], as well as quantitative and qualitative data on pregnancy termination [
31,
32].
Sampling
Women were recruited to participate in the study using respondent-driven sampling (RDS), a method designed to sample hard-to-reach populations [
33]. Respondent-driven sampling uses a peer recruitment sampling system that allows for the estimation of sample weights to correct for biases associated with traditional chain referral sampling [
33]. Further details on the use of RDS in this study were previously published [
29].
Study procedures
Initial study participants who met study eligibility criteria were identified through partner organizations in Bukavu and interviewed. Upon completing interviews, each initial participant received uniquely numbered coupons to recruit up to three peers who met study criteria. Peers presented to the study office where study eligibility was assessed, interviews were conducted, and each participant in turn received uniquely numbered coupons to recruit up to three other peers.
A proportion of participants were invited to participate in a semi-structured interview after completing the quantitative survey. We anticipated lower recruitment and enrollment in the termination group of the larger mixed methods study compared to the parenting group because of the sensitivities related to pregnancy termination in DRC; therefore, every 5th participant in the termination group and every 20th participant in the parenting were invited to complete an interview. The interviews were conducted after the quantitative survey and no one dropped out or refused to participate in a qualitative interview. All participants received a referral card for medical and/or psychosocial care. Transportation was reimbursed and participants received a headscarf as a token of appreciation (~ $1 USD).
Interviewer selection and training
Interviewers were recruited from the local area and were selected based on previous experience working with sexual violence survivors and/or conducting research on sensitive subjects such as sexual violence. Three researchers (SB, JS, KA) conducted an intensive, on-site 6-day interviewer training on research aims, methodology, recruitment, interviewing techniques, and ethics. In addition, daily debriefings were held with the interviewers during data collection. Recognizing that interviewers may have certain attitudes or beliefs about women with children from SVRPs or about women who terminated an SVRP, potential sources of biases and the importance of neutrality were emphasized during the training and debriefings. Interviewers did not have a relationship with study participants prior to study commencement.
Interview procedures
Verbal consent was obtained from participants due to varied literacy levels among the study population and to minimize potential risks to participants from written documentation of their participation in the study. Interviews were conducted in private rooms at the study office in Bukavu, with only the interviewer and participant present. In keeping with gender-based violence research standards [
34], children over the age of two were not present. For the qualitative study, interviewers verbally administered qualitative survey instruments specific to each study group; qualitative interviews lasted approximately 30–45 min. The questions focused on experiences related to the SVRP and not on the sexual violence incident itself, although open responses were encouraged. Participants were also asked to comment on their concerns in general and what would be helpful for women in DRC, women with SVRPs, and children born from SVRPs. The instruments were written in English, translated into Kiswahili, and back translated to English by a different translator. Translation differences were resolved by consensus with a third translator and local collaborators verified accuracy and clarity of the translated questions.
The interviewers recorded responses by hand in Kiswahili during the interview, but no other notes were collected to minimize risk to participants. Following the interview, a trained interpreter in DRC translated the handwritten responses into English and created electronic files. To preserve participant privacy, no identifying information was collected, and participants were not contacted or re-interviewed following the completion of the initial interview. The study terminated in November 2012 due to security reasons.
Ethical considerations
The institutional review board at the Harvard School of Public Health provided human subjects research approval. The medical inspector in South Kivu Province granted permission to conduct the study. A community advisory board in Bukavu provided study oversight. No identifying information was collected and the study name and documentation did not disclose the nature of the study. The interviewer training included dedicated sessions on a trauma-informed approach to interviewing and how to recognize symptoms of psychological distress. A psychosocial assistant was present in the study office at all times, and there was a referral procedure for the study if additional medical and/or psychosocial resources were needed. The manually recorded interviews were stored securely in the study site and the electronic files were password protected.
Data analysis
The electronic files were uploaded to the qualitative data analysis software, Dedoose (Version 5.0.11, Los Angeles, CA). Thematic content analysis was applied to identify themes and patterns in the text data [
35,
36]. Two researchers (CM, SB) conducted an initial review of the transcripts to determine preliminary codes for thematic organization of the data. These preliminary codes were then further refined by two researchers (JS, SB). Data were independently coded by JS and SB and organized into key conceptual themes. Coding inter-rater reliability was measured with a pooled Cohen’s kappa (0.92) [
37].
The overall coding framework for the qualitative study comprised the following categories (Table
1): psychosocial responses, disclosure about the pregnancy, social and contextual determinants of decision-making related to the SVRP, and access to abortion services. Within the psychosocial responses category, there were the following codes, each with multiple subcodes: emotional responses, strategies for responding to a situation, resilience, pride and respect, request for services and aid, and a call for peace. Throughout the process, researchers triangulated the qualitative data with the previously analyzed quantitative data, the existing literature, and peer debriefing. Preliminary findings were presented to study partner organizations in-person in Bukavu.
Table 1
Coding framework for qualitative study and psychosocial analysis
Psychosocial responses | Emotional responses |
Disclosure of the pregnancy | Strategies for responding to a situation |
Decision-making related to the SVRP | Resilience |
Access to abortion services | Pride and respect |
Request for services and aid |
Call for peace |
Discussion
The study provides qualitative data on psychosocial outcomes for women with SVRPs in eastern DRC. The data highlight a myriad of psychological sequelae and influences on psychological well-being, as women navigated complex, and often unstable, social environments after sexual violence and SVRPs. The interviews provided further details on the extent of loss of social supports, including abandonment and rejection by the spouses and families. Among participants, there was prominent mention of feeling excluded from society—an in/out group social identity. Prominent and poignant were the findings related to many respondents’ desire for agency and advocacy to advance the social conditions for women and survivors of violence in eastern DRC.
The qualitative data on emotional responses provide further context to the quantitative results from this study, which assessed symptom criteria for depression, post-traumatic stress disorder, anxiety, and suicidality, finding high prevalence of these mental health disorders among women with SVRPs [
30,
31,
38]. Similar results were seen in other quantitative assessments of mental health outcomes among sexual violence survivors [
3,
12]. The qualitative data provide additional information on concurrent social conditions and factors that are not readily captured by quantitative mental health assessments.
Although the qualitative study was not designed to look at differences between the parenting and termination groups, a number of observations arose. In regard to mental health, suicidal ideation or suicide attempts were described in the parenting group interviews, but not in the termination group interviews. Given the small sample sizes and the lack of design intended to look at differences in mental health outcomes between the two groups, it is difficult to interpret this finding. It is possible that termination of an SVRP provided relief and contributed to overall healing after a traumatic event, as noted in a study of rape-related pregnancies in the United States [
39.] The quantitative analysis of mental health outcomes from this study did not directly compare the study groups; however, a similar proportion of women in each study group met screening criteria for depression and post-traumatic stress disorder, while a lower proportion in the termination group met screening criteria for anxiety or reported suicidality [
30,
31].
A second observation of differences between the study groups in the qualitative study relates to description of marital status in the interviews. Many women in the termination group described being married in the interviews with the spouse’s reaction noted to play an important role in decision-making around the SVRP, whereas women in the parenting group commonly described being abandoned or separated in the interviews. It is important to note that married women were inadvertently under-sampled in the qualitative survey and it is impossible to know in all cases whether the marital separations resulted from sexual violence, the SVRP, a combination of the two, or from other unrelated factors. It is conceivable, however, that having terminated the SVRPs, some women may have been able to continue their marriages. These findings merit further exploration in future studies.
As in other studies of sexual violence survivors in DRC [
7,
8,
13,
40], social stigmatization was commonly described in the interviews. This study’s previously published quantitative data showed that social stigmatization was a mediating factor of mental health outcomes and parenting relationships between women and their children born from SVRPs [
30,
38], similar to the mediating role of stigma on mental health among sexual violence survivors described in other studies [
14,
20]. Most of the interviews described enacted social stigma; however, internalized shame and social isolation were also depicted. There is a growing body of literature on the influences of stigma, negative social attitudes, and low levels of social integration and cohesion on trauma symptom severity and mental health following violence [
19,
41‐
44]. Our study does not allow for conclusions about whether women who terminated SVRPs experienced more or less stigmatization within their communities since stigma from sexual violence, stigma from the SVRP, and stigma around pregnancy termination were not easily distinguishable in the interviews. Stigma related to pregnancy termination is described in other studies, but may also co-exist with a sense of reproductive agency among women [
45]. Further research is needed on attitudes toward termination of SVRPs in this context.
The impact of stigma on mental health ultimately needs to be considered in the context of stressors and trauma related to conflict, poverty, and broader social conditions [
14,
46]. Research shows that trauma is associated with poorer psychosocial functioning, independent of mental health disorders [
47], and that multi-dimensional psychosocial care should be considered within the broader context of trauma-informed care [
48]. Recent evidence also highlights the impacts of daily stressors on mental health, and suggests a sequenced approach to interventions, such that daily stressors are first addressed then followed by more specialized interventions to promote mental health [
46]. According to Maslow’s hierarchy of needs [
49], it is necessary to have basic needs met in order to advance psychological well-being. The interviews emphasized the importance of addressing the manifest needs related to shelter, food, education, and health care; suggesting that not only meeting those needs, but recognizing their significance, is crucial so that survivors are afforded an opportunity to address the trauma they have experienced.
The findings from this study have several important implications for programs and policies in DRC. Respondents identified the need for programs to prevent stigmatization of survivors of sexual violence and women with SVRPs and their children, including programs targeting men; similar to a previously reported program for sexual violence survivors in DRC [
50]. Respondents also called for greater social support programs for women with SVRPs and their children, for restoring respect and dignity of women, and for social justice to end sexual violence and impunity. Other research notes that gender inequality results in inferior social status for women in terms of education, economics, justice, and personal autonomy; which is thought to contribute to the prevalence of sexual violence despite peace accords and sexual violence legislation in DRC [
51,
52]. Future programs to promote accepting attitudes toward survivors of violence and women with SVRPs would need to address the cultural and social norms that support sexual violence as noted in other studies [
53‐
56] and target inequitable gender norms [
43].
Finally, while this study did not directly probe for post-traumatic growth, correlates are found in the interviews. Women reported a desire for agency and advocacy on behalf of themselves and other survivors of sexual violence and SVRPs; exemplifying dimensions of post-traumatic growth [
18,
57]. The qualitative data also provide accounts of individual strength and resilience, which offers an opportunity for discourse on strategies for promoting psychosocial well-being for violence survivors. As an example, a psychosocial program among war-affected Darfuri women in Sudan integrated cultural conceptualizations, such as gender roles and religious expectations, and aimed to operationalize resilience and other psychosocial adaptation skills to meet the complexity of women’s experiences [
58]. Further evidence suggests that there are cognitive, behavioral, and existential resilience factors that can be cultivated and should be included in interventions following exposure to trauma [
59]. Psychosocial factors associated with resilience include optimism, cognitive flexibility, active coping skills, and maintaining a supportive social network [
59]; many of these factors were evidenced in the interviews.
Limitations
The study sample represents a systematic convenience sample of participants recruited using RDS; thus, the results may not be generalizable. While RDS is useful for sampling hard-to-reach populations, stigma and shame related to sexual violence may have prevented some individuals from participating in the study and women with alternative social experiences may not have been included in this sample. Given the cross-sectional study design, there are limitations in the ability to assess processes of recovery and psychological adaptation. The study was not designed to conduct follow-up interviews of participants, which would have allowed for further exploration of identified themes.
Furthermore, even though the interviews focused on the SVRP, it is not possible to know if the responses are related to the SVRP, to sexual violence, or to both. In some instances, women were asked about their experiences years after the event and the SVRP and there is the possibility of recall bias or limitations in the ability to articulate complex emotional responses years later. There are other potential sources of biases, including interviewer bias and a social desirability bias. Due to the sensitive nature of the subject, it is possible that the interviewers’ background and beliefs influenced their approach to the interviews, along with the participants’ willingness to open up to the interviewers. Additionally, biases may have been introduced through the processes of translation, coding and analysis; however, the translator and researchers had prior expertise and training. The interviews were not audio-recorded due to ethical concerns and the findings rely on notes taken at the time of the interview. While interviewers were trained on interviewing techniques, the lack of contextual data, such as non-verbal and verbal cues, may limit the depth of the data and data analysis.
Authors’ contributions
JS contributed to study design and implementation, data collection, analysis, interpretation, and wrote the initial draft of the manuscript. CM contributed to qualitative data analysis and interpretation, and reviewed and revised the manuscript. SR contributed to study design, field testing of survey instruments, data analysis and interpretation, and reviewed and revised the manuscript. PK contributed to study design, data interpretation, and reviewed and revised the manuscript. AG contributed to study design, field testing of survey instruments, data interpretation, and reviewed and revised the manuscript. KA contributed to study implementation and data collection and reviewed and revised the manuscript. GB contributed to the study design, data analysis and interpretation, and reviewed and revised the manuscript. MO contributed to the study design, data analysis and interpretation, and reviewed and revised the manuscript. MV contributed to study design and reviewed and revised the manuscript. SB conceived of the initial idea for the study, contributed to study design and implementation, field testing of survey instruments, data collection, analysis, interpretation, and reviewed and revised the manuscript. All authors read and approved the final manuscript.