Background
Sexual violence has been prevalent throughout the armed conflict in eastern Democratic Republic of Congo (DRC) and has been systematically used by various armed groups to destabilize communities [
1,
2]. Exposure to sexual violence – often in the form of gang rape or sexual slavery – places women at high risk of unintended pregnancies. According to a study conducted in eastern DRC in 2010, 40 % of women reported sexual violence and an estimated 17 % of those women reported resultant sexual violence-related pregnancies (SVRPs) [
2]. Unintended pregnancy is associated with risks to maternal health and poor pregnancy outcomes [
3,
4], and in conflict and post-conflict settings such as eastern DRC, access to comprehensive pregnancy care, including pregnancy termination and post-abortion care, is limited [
5‐
7].
Recent studies estimate that between 8 and 13 % of maternal deaths globally are due to unsafe induced abortion [
8,
9]. Regional data suggests that at least 12 % of maternal deaths in central Africa can be attributed to unsafe abortion [
9]. However, stigma and restrictions associated with abortion, may result in misclassification and under reporting of abortion-related morbidity in certain contexts [
8]. In the DRC, termination of pregnancy is highly restricted, and the law is generally interpreted to allow induced abortion only to save the life of the woman [
10‐
12]. The impact of unsafe abortion and complications from an SVRP on the maternal mortality rate in DRC (estimated at 730/1000,000 live births) is unknown [
13].
While there is a growing body of evidence across multiple countries showing that women who experience intimate partner violence have an increased odds of unintended pregnancies and abortion [
14], research specific to pregnancy termination among women with SVRPs, and in particular data on access to termination and methods of termination, is limited. A hospital-based study in Liberia found that of women who presented with an SVRP, 59 % requested a termination upon seeking care [
15]. Another study conducted in Goma, North Kivu Province, DRC, found that 87 % of women who reported an SVRP carried the pregnancy to term; however, 55 % said they were willing to terminate, 39 % said they would terminate if appropriate care was available, and 10 % attempted to terminate the pregnancy, all under unsafe conditions [
16]. Research on SVRPs and pregnancy termination in eastern DRC, a context with a high prevalence of sexual violence, high maternal mortality, and restrictive abortion laws, is important to improving the overall health of women in eastern DRC. Further understanding of women’s experiences and perspectives with regards to pregnancies and pregnancy termination in the context of sexual violence is critical for developing effective, targeted programming [
17]. The data presented in this paper are part of a larger mixed methods study conducted in eastern DRC to assess the experiences of women with SVRPs. This paper presents qualitative data on access to pregnancy termination and describes the methods of pregnancy termination among women with SVRPs in eastern DRC. To our knowledge, this is the first study to examine women’s perceptions of and access to abortion in this context.
Methods
The study was conducted in Bukavu, DRC in October and November 2012. The qualitative data presented in this article were derived from a larger mixed-methods study to understand: 1) the parenting experiences of women raising children born from an SVRP, and 2) the experiences of women who terminated an SVRP. Qualitative data on perceptions and access to termination will be presented in this manuscript.
Inclusion and exclusion criteria
For the larger mixed methods study, participants were eligible for study participation if they self-identified as a survivor of sexual violence since the start of the war (~1996), reported a pregnancy as a result of sexual violence, and were 18 years or older. Two subgroups of women with SVRPs were included in this study: 1) women currently raising children from SVRPs (parenting group), and 2) women who had terminated SVRPs (termination group). Women were included in the parenting subgroup if they reported delivery of a live born infant and were living with and caring for the child at the time of the study. Women were excluded from the parenting group if the pregnancy resulted in a stillborn infant, if the child had since died, or if the child was not currently living with or in the care of the mother. Women were included in the termination subgroup if they self-reported an induced abortion of the SVRP. Women were ineligible to participate in the study if the SVRP aborted spontaneously. The parenting and termination groups were not mutually exclusive, and women with a history of more than one SVRP were eligible to participate in both study groups provided that the inclusion criteria for each was otherwise met. All participants completed a quantitative interview, and a subset described below completed qualitative interviews.
Participant recruitment using RDS
Participants for the larger mixed methods study were initially recruited using respondent-driven sampling (RDS), a peer-recruitment method designed to sample hard-to-reach populations [
18]. RDS methodology was selected because it allows access to a hidden population that would have otherwise been difficult to access except through convenience samples [
19]. This sampling approach offers advantages over convenience samples because it controls for some of the biases that are inherent in the sampling process. Under ideal conditions RDS may approximate a randomized sample [
18]. Full details of the research methodology have been published elsewhere [
19] and are described briefly below.
As is typical for RDS methodology, initial study participants who met study eligibility criteria were purposively selected by local partner organizations and completed a quantitative interview. After completing the interview, each initial participant was asked to recruit up to three additional eligible peers by sharing uniquely numbered recruitment vouchers. Those peers who were eligible for the study were subsequently interviewed and then asked to recruit other potential participants, and so on. A total of 852 participants (764 in the parenting group, 83 in the termination group, and 3 who met criteria for both study groups) completed quantitative interviews. Quantitative data from the parenting and the termination group have been published elsewhere [
20‐
22].
Qualitative study procedures
A subset of study participants were systematically selected and invited to participate in a semi-structured qualitative interview. Due to different rates of recruitment within each study group, every 20th participant in the parenting group and every 5th participant in the termination group were invited to participate in a qualitative interview following completion of the quantitative questionnaire. Two semi-structured interview guides, one specific to each group, were developed for this study by the research team in collaboration with local partner organizations; the interview guides were not previously used in other settings. Interview guides were initially written in English, translated into Kiswahili by a local translator and then back translated by a different translator. Translation differences were resolved by consensus with a third translator. A panel of local collaborators reviewed the translated interview guides for accuracy and was piloted prior to data collection. The interview guide corresponding to the study group (parenting versus termination) was then verbally administered in Kiswahili by female interviewers from eastern DRC, who had completed a comprehensive 6-day training on research ethics and methods. Interviews were semi-structured; participants were asked open-ended questions to obtain further information about pregnancy history, sexual violence and the SVRP, decision-making process regarding continuation or termination of the pregnancy, and concerns for the future. Interviews were conducted in private in the study office located in Bukavu. Responses were hand-recorded in Kiswahili by the interviewer, translated into English and transcribed verbatim into electronic transcripts by a local trained translator. Interviews were not audio recorded to protect participant identities.
Coding and analysis
Thematic content analysis, which was complemented by the constant comparative technique from grounded theory, was used as the analytic approach for data analysis [
23,
24]. The iterative process of concurrent data collection and analysis was not used due to study logistics and lag time in the translation of interview transcripts.
Electronic files containing the English transcripts were uploaded to the qualitative data analysis software Dedoose (Version 5.0.11, Los Angeles, CA) [
25]. Documents were examined consecutively line-by-line in order to identify the thought patterns, feelings and actions that were described by participants in their interviews, while also remaining sensitive to the existing literature base as well as to being purposeful about separating relevant and non-relevant data. Two researchers initially reviewed transcripts in order to determine preliminary coding structures for organizing the data thematically. These researchers then independently coded transcripts and the data were organized into key conceptual themes. New codes were added as themes that did not fit the initial coding structure emerged. Coding inter-rater reliability was measured with a pooled Cohen’s kappa, which demonstrated a very strong agreement at 0.92 [
26]. This paper presents data on the key conceptual themes of access to and perceptions of abortion; other themes will be presented in separate manuscripts.
Positionality of researchers
Our research team is comprised of eight academic physicians and one academic nurse midwife based at U.S. academic institutions, all with significant expertise in DRC and sexual violence. Eight members of the research team worked in the DRC either during the study design, implementation, and dissemination or had prior experience working in the DRC. All members of the research team possessed a clinical and/or research background in sexual violence and/or comprehensive reproductive health in humanitarian settings. The interviews were conducted by Congolese researchers with extensive experience working with sexual violence survivors and reproductive health. Study design, implementation, and data analysis were all informed through close partnerships with three organizations in Bukavu, DRC: Centre d’Assistance Médico-Psychosociale, Acteurs Dans le Developpement et Droit a la Sante pour Tous, and Action Pour la lutte Contre L’ignorance du SIDA.
Ethics approval and consent to participate
Human subjects research approval was obtained from Harvard School of Public Health and a community advisory board in Bukavu provided study oversight. Permission to conduct the study was also obtained from the medical inspector in South Kivu Province. Verbal informed consent was obtained due to concerns about illiteracy and concerns that written consent could identify the participant as a sexual violence survivor. No identifying information was collected. Participants who completed the study were offered a headscarf ($1 USD) as compensation for their time and transportation reimbursement (up to $8 USD) was provided directly to the taxi driver. A trained psychosocial assistant offered counseling services on-site to participants who were distressed or who requested counseling services. All participants received a referral card for medical care and/or mental health counseling.
Discussion
Findings from the study highlight women’s experiences related to termination of SVRPs in eastern DRC. The qualitative data offer perspectives on the reproductive options available and accessible for women with SVRPs, as well as attitudes toward termination of SVRPs. Our findings also point to limited utilization of evidence-based methods of pregnancy termination for women with SVRPs.
The qualitative results reveal that women in this study often used herbs or medicines to induce an abortion obtained through informal health care networks, either from a friend, family member or traditional healer. The majority of women in this study who terminated the pregnancy used non-evidence based methods, most commonly a traditional herb known as
Cimpokolo, known by the scientific name
Phytolacca dodecandra L’Hérit [
27,
28]. The qualitative finding reflect the data from the larger quantitative RDS sample, in which 31 % (26/86) of women cited
Cimpokolo as the most common oral medication or herb used to induce abortion [
22]. There is minimal evidence on the mechanism or efficacy of this traditional herb, although other herbal and plant-based abortifacients have been described in the literature [
29]. Women in this study did not mention using evidence-based methods of safe abortion such as misoprostol, mifepristone, or vacuum aspiration [
30]. This study did not fully evaluate if such evidence-based methods were available to women or if women were aware of these methods. Further understanding of women’s knowledge on methods of termination is warranted.
Few women in the qualitative sample accessed termination services through the formal health care sector. Of the women who terminated an SVRP, only three mentioned that the abortion was conducted at a health facility or with a health care provider. Future research should examine reasons for this. This may be reflective of the fact that in settings such as DRC, with highly restrictive abortion laws, women’s access to safe abortion services, supplies, and medications, as well as skilled providers, are lacking, and may be even further limited due to the country’s post-conflict setting [
31]. A recent evaluation conducted in eastern DRC found that post-abortion care was available in one hospital and in 11 out of the 25 health centers assessed; however, none of the facilities included in this assessment openly provided safe abortion services [
32]. Other research has noted that implementing agencies may omit safe abortion programming due to the politically charged nature and the associated legal complexities associated with funding restrictions for abortion programs [
7,
31]. Alternatively, the findings from this study could indicate that women in this context prefer to access care in the informal health sector for abortion services or are not aware where abortion services may be available. Further research on women’s preferred point of access for termination services and the availability of evidence-based methods for pregnancy termination in both the formal and informal health care sectors is needed.
For the few women in this study who did access care through the formal healthcare sector, it was not evident from the interviews whether the health providers they saw were trained on comprehensive abortion services or whether the facilities had supplies and equipment for providing safe abortion care. In can be inferred from the medications prescribed by the health facilities (injections and quinine) that the providers were not aware of evidence-based methods to terminate pregnancy. Our narratives also revealed a pervasive awareness among the women in this setting of both fear and risk of death from termination. Although the prevalence and types of complications were not directly assessed through the qualitative methodology, the fear of complications suggests that terminations may have been carried out in unsafe conditions, and that complications were common enough so that women were aware of them. More research is needed on the magnitude of unsafe abortion and the prevalence of complications from termination of SVRPs in this context.
The qualitative data reveal the complexity of attitudes towards termination of SVRPs. The data from our study highlight the importance of religious and cultural beliefs that may influence pregnancy termination in this setting. While our study was not designed to investigate providers’ attitudes towards abortion, one woman in our study sample recounted how she was turned away from a health facility when she asked for a termination. Though not a pervasive theme in our study, providers’ attitudes towards abortion have been cited as barriers to access of care in other settings [
33‐
36]. Other studies reveal that even with legalized abortion services, women may still seek unsafe or self-induced means of pregnancy termination as a result of stigma, financial reasons, or confusion around the law, emphasizing that efforts must be focused on improving accessibility, increasing the number of skilled providers, as well as addressing the cultural attitudes toward termination [
37,
38]. A more detailed understanding of attitudes among both women and providers in this context may help to provide tailored services for women with SVRPs.
Several women in our study called on the government to legalize abortion for women with SVRPs. Currently, the legal statute for abortion services in the DRC is complex and abortion is highly restricted [
10‐
12,
39]. The 1982 DRC Penal Code stipulates that abortions are illegal and subject to 5 to 15 years of imprisonment [
40]. The DRC has ratified the
Maputu Protocol, which stipulates that conferring states can legally provide abortion to preserve physical and mental health of the woman, in cases of fetal impairment, and in cases of rape or incest [
41]. However, the women in our study who discussed legal issues called upon the government to allow abortions for SVRPs, suggesting they were unaware that pregnancy termination could be considered legally permissible in cases of rape. Evidence has shown that in countries such as the DRC, restrictive abortion laws contribute to higher rates of unsafe abortion, and therefore higher rates of maternal morbidity and mortality from abortion related complications [
42,
43]. It is possible that fear of legal consequences could contribute to the selection of non-evidence based methods by the women interviewed. Since the qualitative instrument did not specifically ask about legal consequences, it is also possible this association may not have emerged in our data. Further clarification on the legal context of pregnancy termination and sexual violence in the DRC for both women and health care providers is needed, and could potentially allow for increased access to safe abortion services for women who seek termination of SVRPs [
44‐
46].
Limitations
The data presented here are part of a larger study, which used RDS to recruit participants. The qualitative sample reflects a systematically selected convenience sample from those recruited by RDS for the larger study. Therefore, the qualitative results in this paper represent the attitudes and experiences of those interviewed but are not generalizable. Comparison of demographics between the quantitative and the qualitative interviews indicates that married women raising children from SVRPs were under-sampled in the qualitative interviews; other differences in sampling may exist and may result in over or under representation of other groups.
Study inclusion and exclusion criteria may have affected study results. The study did not include women who gave birth to but were not raising a child from an SVRP, and their opinions are not reflected in the current data. Differences in study instruments could have contributed to information that was or was not obtained among different participants.
The questions assessed termination of pregnancies, a sensitive issue to discuss, and responses may be further influenced by a social desirability bias. Although all efforts were made in assuring fidelity to the context during translation and that local terms were translated appropriately, there may be translation and interpretation errors in this data. Additionally as the interviews were not audio-recorded due to the sensitivities of the issues discussed, some data could have been lost due to filtering by the interviewers.
Finally, as with other qualitative research, the responses may be influenced by the interviewers’ own biases and perceptions and the presence of the interviewer alone can impact the response. Efforts were made to train and hire independent interviewers who were well versed in qualitative research methods as well as the sensitivities of the subject matter and cultural norms. Additionally, there may also be interpretation bias by coders and researchers.
Acknowledgements
Funding for this study was made possible by the Eleanor Miles Shore Foundation [GL Dept ID 01025283] and the Harvard Humanitarian Initiative. The authors would also like to thank the women who participated in this study, the team of interviewers, and local partners in DRC who made this study possible.