Overview of research methods

Data for this analysis come from a mixed-methods study of abortion in Kinshasa and Kongo Central, DRC conducted in 2021–2022. The study was conducted as part of the Performance Monitoring for Action (PMA) project, which implements population-based surveys to monitor key family planning and reproductive health indicators in nine countries throughout sub-Saharan Africa and south Asia. Additional information about the PMA project, including the methodology, questionnaires, and data, is available at www.pmadata.org/data/survey-methodology. In the DRC, PMA is led by the Kinshasa School of Public Health with technical support provided by the Johns Hopkins Bloomberg School of Public Health. Ethical approval for this study was obtained from the Institutional Review Board at the Johns Hopkins Bloomberg School of Public Health (#14590) and the Comité d’Éthique at the Kinshasa School of Public Health (#ESP/CE/159B/2021).

The initial phase of the PMA DRC abortion study included a cross-sectional survey using a multi-stage cluster random sampling design to identify a representative sample of women aged 15–49 years living in Kinshasa. A total of 2,326 participants provided informed consent and completed a face-to-face survey that collected information on their sociodemographic profile, reproductive history, contraceptive practices, and abortion history. Women in Kinshasa who completed the quantitative survey [6] and reported an abortion in the last 10 years were eligible to participate in the qualitative follow-up interview. This study uses the data from these qualitative interviews.

Altogether, 93 women reported having an abortion between 2015 and 2021 in the quantitative survey and agreed to be contacted for a follow-up interview about their abortion experience. Women were contacted by telephone according to a script approved by the ethics committee. From the 93 eligible participants, interviewers first attempted to contact those who had the most recent abortions, continuing to contact women until reaching the goal of 50 completed interviews. In total, interviewers attempted to contact 65 women, among whom 52 women provided written informed consent to participate in an in-depth interview about their abortion experience and allow a digital recording of the interview. The interviews were conducted at a day, time, and place of the woman’s choosing.

Nine qualitative interviewers–all of whom resided in Kinshasa and had prior data collection experience via the PMA project–conducted the interviews after completing a four-day training about research ethics, informed consent, qualitative research approaches, and interviewing techniques. The interviews, conducted in Lingala, lasted 30–75 minutes. Interviews followed a structured interview guide, including open-ended questions and probes, developed by a multidisciplinary team of researchers in Kinshasa and the United States. The guide was pre-tested during the interviewer training and addressed the following topics: 1) local abortion terminology used by women, 2) information-seeking practices for abortion options, 3) decision-making processes, and 4) abortion care trajectories. Interviewers also asked women to reflect on the abortion laws in the country and share any recommendations for improving the quality of care they received during their abortion experiences. The interviewers took notes to complement the audio recordings of each interview. All transcripts were translated verbatim into French by two transcribers, with random checks for quality of translation. Confidentiality and anonymity were guaranteed during each participant’s informed consent process, and participants were not identified by name or any other identifier in study materials. Audio recordings were immediately downloaded onto a password-protected computer, then erased from the recording device. The interviewers’ notes were archived in secure cupboards in locked offices on the premises of the local DRC team. No personal information about respondents was included in the interviewers’ notes. All women were reimbursed $2 US for time spent in the interview and to cover transportation costs. Data were protected in locked, password-protected computer files and were de-identified for analysis. Only the project team had access to the raw data. The recordings will be archived for five years, after which they will be destroyed.

Analysis

The study team used Atlas.ti 22 for data management, coding, and analysis. First, members of the study team reviewed all interview transcripts in French and generated memos for each transcript, recording any salient ideas, questions, or important notes about the transcript and/or participant. Next, a team of six researchers (three based in the DRC and three based in the US) developed a common codebook using open coding techniques—i.e., identifying potential codes from three transcripts, followed by discussions among the research team members to identify and generate a draft codebook. The coding approach followed a paradigm developed by Strauss and Corbin (1990) to identify terminology and definitions for coding subcategories (e.g., phenomenon, causal conditions, strategies, consequences, context, and intervention conditions related to abortion information seeking) [21]. Next, we conducted axial coding—i.e., the development of a coding structure and a final codebook that was used for the remaining transcripts. Once the final codebook was established, each researcher coded approximately 10 transcripts, and then each coded transcript was reviewed by another team member to ensure consistency among coders and completeness of codes. During the coding phase, memos were recorded to explore any emerging research questions or ideas. Coders also noted any inconsistencies in participants’ accounts of their abortion experience so that discordant dimensions of care could be explored.

After coding was finalized, we conducted an inductive thematic analysis and developed a coding report in Atlas.ti to export a complete list of coded quotes with the relevant codes for our analysis. We then developed a matrix to identify and describe the major themes and subthemes. The matrix included relevant quotes illustrating each theme or subtheme, including participant characteristics (e.g., age, marital status, parity) for each quote. The researchers also reviewed the transcripts directly, focusing on sections of the interviews that asked respondents how they sought information about their abortion options and whom they used in the process, to identify potential quotes, subthemes, and themes that had not been captured in the original coding report. Inductive thematic saturation techniques were used to achieve saturation of themes and sub-themes. Research team members focused on identifying newly emergent codes and themes, instead of the completeness of the existing codes and themes, at each research team meeting until no new codes or themes emerged [22]. Rigor was maintained throughout the research process through weekly research team meetings to discuss challenges, or questions regarding the interpretation of quotes, and ideas related to conceptual linkages between key themes, which helped generate the final thematic matrix.

Participant characteristics

More than half of the women interviewed were younger than 30 years old and had never been married (Table 1). Nearly 7 out of 10 women had completed secondary school. Thirty-six (69.2%) women had at least one child. Over half of women received a procedural abortion (51.9%), the next most common method being pills other than medication abortion pills (34.6%). Most women visited private facilities (44.2%) or pharmacies (42.3%) to obtain their abortion.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of qualitative sample of reproductive-aged women in Kinshasa, DRC who had a recent abortion.

https://doi.org/10.1371/journal.pgph.0002383.t001

Thematic analysis revealed the intersection of two main themes related to care-seeking behaviors and quality of interactions. The first and largest theme on care-seeking behaviors highlighted the highly selective and narrow information search process women engaged in, involving no others or very few individuals (e.g., partners, close female networks, or health care providers) that the pregnant woman chose strategically. The second theme on the quality of these interactions revealed the heterogeneous and often stigmatizing nature of these conversations, including attempts at deterrence and information of varying completeness and accuracy.

Care-seeking as a selective process

Women’s search for abortion care information mostly depended on their sense of social safety, as abortion remained widely stigmatized and potentially punishable by law. Confidentiality guided their paths, favoring self-reliance if women had prior knowledge, extending to partners if they were supportive or activating trusted female networks (family, friends, or neighbors) if women had no option but to seek external support.

An autonomous course

Respondents who already knew about methods and places of abortion care—either through their education or medical training, their circle of acquaintances, or because they had had a prior abortion—said they did not feel compelled to solicit information from their social network, often managing the situation on their own to avoid the risk of disclosure.

As one 33-year-old woman shared about her ability to navigate her abortion options due to her profession, "Yes, I already knew some methods … It’s because I, myself, am a medical professional [health care provider] … As soon as I arrived at the pharmacy, and I had bought the medication …I had written my prescription myself and the pharmacist had served me." (Divorced woman, no children).

Other women shared how their prior experiences with abortion shaped their knowledge and pursuit of specific abortion methods. A 28-year-old woman stated, "Yes (I knew the methods of abortion) since I had already done it once. I know that they give a medicine to put vaginally and another to inject in the buttock. Another method, they use instruments that open the cervix and then circle the blood." (Separated woman, one child).

Women in these scenarios were not obligated to disclose their abortion to anyone in order to learn about methods and sources for seeking care. This limited potential social sanctions, sometimes at the expense of more accurate and complete abortion information from other sources.

Activation of the social network

Many respondents, however, possessed no prior knowledge about abortion, requiring them to “activate” their personal networks. Women were very selective about who they approached, often limiting their interactions to a trusted individual, whether a partner, sister, or friend, to maintain secrecy throughout their abortion care trajectory. The nature of these interactions resulted in some women being dependent on the knowledge of only one confidant to protect their reputation.

Quality of interactions

The nature and quality of the information women received varied across sources and was as much about the decision to abort as it was about the methods and risks involved. Some exchanges were aimed at discouraging women from terminating the pregnancy, while other conversations focused on the methods and safety of the procedures, warning of the risks and perceived negative consequences of abortion, including inaccurate information.

Attempts at deterrence

In seeking advice on where and how to obtain an abortion from those around them, many women were discouraged from terminating the pregnancy. Sometimes this took the form of a partner’s refusal or a family member’s fear of tarnishing the family’s honor due to the stigma of abortion in the community.

As one adolescent shared, "On the same day of the abortion at home, my mother was already prejudiced and indirectly forbade me to be able to have an abortion by saying, ‘At home, here, none of my children will be able to have an abortion, and the one who will dare to do it will die.’ Afterwards, my older brother, who lives in Gombe, called me and said, ‘I just had a dream that you are having an abortion, while I never have an abortion and I don’t want one of my sisters to have one.’ I was so afraid." (18-year-old woman, never married, no children).

Some women expressed how they were discouraged from terminating pregnancies by their partners. In some cases, women acted independently in seeking the abortion, despite this attempt at deterrence; as another adolescent described, "My partner had refused to have an abortion, but I had done it unilaterally. I could see my partner’s state of mind was not good …But I didn’t tell my partner because if and only if my partner knew about it, he didn’t want me to have an abortion." (19-year-old woman, never married, no children).

Some providers only offered medical advice without facilitating the abortion for fear of legal reprisal. However, several women seemed to overcome providers’ reluctance by involving their male partners or offering providers bribes for their services. For example, one 20-year-old woman shared her experience of negotiating with the provider alongside her male partner, saying, "[My boyfriend and I] went to a pharmacy to buy a medicine, whose name I forget, but the pharmacist refused to serve us because it is only for married people and boys. I bribed the pharmacist…and he served me but telling me that if this product has side effects on you it will have nothing to do with me." (20-year-old woman, never married, no children). Others were less successful in their efforts to persuade providers, as one young woman described, “Like when my boyfriend and I went to see the first doctor, at first, he didn’t agree. He advised against it because he felt that after an abortion, there are always negative repercussions that could even cause him to stop or even close his hospital. He really advised against it and said: ‘It’s better to have a miscarriage than an abortion’.” (18-year-old woman, never married, no children).

Heterogeneous information on procedures

The informant network was a key determinant of the quality of the information that women obtained. While providers generally advised recommended methods, this was often not the case with social network or community sources, which relayed information on non-recommended methods, such as commonly used drugs or traditional or home remedies that frequently led to complications.

Media was rarely mentioned as a source of abortion information (and the internet or social media were never mentioned), providing only sporadic and incomplete exposure to abortion methods, as one participant shared, "I had pretty poor knowledge about abortion, but on TV I had once heard about curettage and abortion pills." (40-year-old woman, living with partner, three children).

Others who relied on their friends or community network similarly received incomplete and inaccurate information about abortion options, often resulting in use of non-recommended methods. One woman described recommendations she received to consume an herbal remedy, stating, "Yes, I already knew that you can take the ‘tangawisi’, and it can remove the pregnancy. From what is said in the street by everyone, a pregnant woman cannot drink the tangawisi juice, ‘esopaka zemi’ (i.e., this juice has ‘abortive virtues’). From such information, I had known that if I drink the tangawisi, my pregnancy will go away.” (29-year-old woman, never married, two children). Scenarios like these were common when women sought information from members of their community, avoiding care and clinical guidance from healthcare providers.

In contrast, conversations with healthcare providers generally focused on the abortion method deemed appropriate according to the gestational age of the pregnancy. These interactions also included information about the safety of the procedure by explaining possible complications. However, the complications were often overestimated or misrepresented compared to the risks of the WHO-recommended methods. The risks appeared to be amplified as a mechanism for dissuading women or reflected the reality of a context characterized by suboptimal quality of care.

As two respondents shared their positive care experience with a provider: as one woman shared, "He [the provider] had explained everything to me, the pros and cons of abortion." (27-year-old woman, never married, one child), and another, "When he [the doctor] came to the house I asked him questions, I asked him if we should go through curettage or other methods, he told me that we would buy the medicines and that I would take them. He told me that I was going to have excruciating pain for 3 or 4 days because it was a forced thing…he explained everything to me. He told me about all the complications that could occur, such as damaging the woman’s cervix, even getting to the point of cutting the tubes, not having any more children. He made me see all the possible consequences." (21-year-old woman, never married, one child). Similarly, one participant described how her interaction with the health provider emphasized potential long-term consequences of an abortion for her fertility, which were not grounded in evidence, "He [the nurse] told me that I can be at risk. It can have complications such as infertility." (23-year-old woman, living with partner, one child).

Overall, women relied heavily on information circulating in their communities about their methods and sources for seeking an abortion. While women shared their experiences of successfully obtaining information from the formal health sector, guidance from providers was generally of low quality, as many evidence-based aspects of safe abortion care were not shared with women or were communicated inaccurately.