Ethical Approval
The Institutional Review Boards of the Albert Einstein College of Medicine, Bronx, New York and the National Research Ethics Committee in Cameroon granted ethical approval for the study.
Study Setting
The study setting is the Regional Hospital, located in the coastal town of Limbe in Southwest Cameroon. The hospital, built around the 1940s was accorded a provincial status in 1972, thereby becoming the principal referral hospital for the southwest region (10).. Limbe Regional Hospital (LHR) is also called the “Mile One Hospital” because of its location about one mile from the Atlantic Ocean (10).. Since its establishment, the hospital has experienced several structural and organizational problems that have limited its capacity in providing adequate health services for the population it serves.
Research design and study population
This study utilized an exploratory-descriptive qualitative approach and was conducted as part of a study assessing cervical HPV infection and neoplastic disease in women living with HIV (WLHIV) all embedded within the Central Africa International Epidemiology Database to Evaluate AIDS (CA-IeDEA) project in Cameroon (11).. The study population consisted of women living with HIV (WLHIV) and those not living with HIV (HIV-negative) who met the following eligibility criteria: aged > = 25, ever or currently sexually active, not pregnant at enrollment into the study, never screened for or diagnosed with cervical cancer, was able to provide both self-collected and provider collected biological samples for HPV testing and was able to understand and sign the informed consent. In addition, male spouses or partners of enrolled were recruited to explore men’s perspectives and attitudes toward cervical cancer screening.
Sample selection
We utilized a two-stage sampling strategy to select the women and men who participated in focus group discussions (FGDs) and in-depth interviews (IDIs). In the first stage, we utilized the master list of 877 women represented in the parent study to identify and generate a list of 585 WLHIV and 292 HIV-negative women. The second stage was the systematic sampling of 36 WLHIV who participated in 3 FGD sessions (12 in each FGD) and 4 WLHIV who participated in IDIs. The process was repeated to select the same number for FGDs and IDIs among HIV-negative women.
The selection of FGD discussants and IDI interviewees was done after the study nurses contacted potential participants, provided information about the study and invited those who were interested to participate in the FGD or IDI. Recruitment of male FGD participants was done by asking the women in each focus group to invite their spouses or significant others to participate in the study. A list of all male spouses who agreed to participate was developed and each spouse was contacted by the study team to participate in the men’s focus group discussions. The first 10 male spouses in each category who responded were then invited to join each of the discussion groups.
Data collection
Focus group discussions and in-depth interviews were held between May and August 2018. The rationale for using FGDs and IDIs was to better understand and describe women’s knowledge, attitudes, and practices regarding cancers in general, cervical cancer, HPV infection, screening as well as behavioral and structural facilitators and barriers to cervical cancer prevention. Additional information was obtained to assess and compare perceptions and preferences for self-versus health provider-collected biological specimens to understand women’s preferences given peculiar contextual factors that facilitate or inhibit access to cervical cancer screening for women at risk. Focus groups with male spouses and partners explored similar issues, including their knowledge and attitudes about cancer and cervical cancer screening, as well as if and how they provide support to wives and significant others in cervical cancer prevention, treatment and care.
Overall, we organized 6 women and 2 men’s focus groups, and 8 individual in-depth interviews. These consisted of 3 focus groups for women living with HIV (WLHIV) and 3 focus groups for HIV negative women by the following age categories: 25–35 years, 36–45 years and > = 46 years. Two additional focus groups, (1 for partners of WLHIV and 1 for partners of HIV negative women) were organized for male spouses. Each focus group consisted of 12 people grouped together based on homogeneous characteristics such as age, literacy, socio-economic status and known HIV status category. In addition, 8 in-depth interviews (4 in each category) were held with women study participants.
The focus groups and in-depth interviews were held at an easily accessible and unanimously agreed location within the premises of Limbe Regional Hospital that afforded anonymity to respondents as well as the confidentiality of the information provided by participants. The language used for the interviews and focus group discussion was "pidgin”,, a colloquial form of the English language that is widely spoken in the area. A team of 3 trained research assistants (a moderator, note-taker and observer) facilitated each focus group session, while a team of two, (a moderator and note-taker) facilitated the in-depth interviews. Each focus group lasted 90 minutes on average while in-depth interviews lasted an average of 60 minutes. Each participant in the focus group and in-depth interview received an incentive of approximately $5 (USD) to cover the cost of transportation and other logistics of participation. Prior to starting the group discussion, each participant was required to complete the informed consent process.
Measures
To facilitate data collection, a semi-structured guide was developed by the research team. The guide was organized around a series of specific themes, each with several open-ended questions. The themes that we explored included (i) knowledge, attitudes and behaviors regarding cervical cancer and association with HPV infection, (ii) individual/interpersonal, socio-economic and cultural influences on cancer screening, including stigma (iii) perceptions of the need for and utilization of cervical and reproductive cancer screening, (iv) user experiences regarding self-collected versus provider collected methods for biological samples, (v) men’s perceptions, attitudes and support for cervical cancer prevention among women, and (vi) macro-contextual factors, including cost, accessibility and availability of screening services for women in the study areas.
In addition, participants also discussed social-contextual factors that influence beliefs, attitudes, perceptions of personal risk and community actions that can contribute to cancer prevention including combining screening for non-communicable diseases such as cervical cancer with HIV testing, role of men in cancer prevention as well as effective strategies/mediums for disseminating cancer prevention educational materials. This design of the guide around these specific measures provided valuable insights in the context of cultural and normative factors influencing perceptions, behaviors and the degree to which participants consider access to preventive care as an important component of cervical cancer prevention.
Data processing
Several steps were taken to process the data generated from the FGDs and IDI. First, at the end of each discussion session and interview, the team of the moderator, note-taker and observer for each FGDs held daily debriefing meetings primarily to reconcile notes from each team member with the audio recording of the FGD. This step required each team member to listen to the audio recording and cross-check the audio with the notes taken during the interview to ensure consistency and to fill in any gaps in the notes. Upon reconciliation, audio recordings and field notes were labeled with information regarding the location of the interview, basic information about the participant(s), pertinent interview information including, date, starting/ending time of the interview and names of the moderator, note-taker and observer (in case of FGDs). The process contributed to the quality of data generated and ensured consistency in the way the data was managed.
Second, the group that facilitated the discussions and interviews held daily debriefing meetings with the larger research team to give and receive feedback in terms of what they did well and needed to improve on for subsequent interviews. Prior to these briefing sessions, other members of the research team who did not participate in the interview listened to the audio recordings so they can provide feedback to the team that facilitated the discussion session or interview. This process helped ensure continuous quality improvement in the data that was obtained as well as in identifying important lines of inquiry that need further exploration if they emerge during subsequent interviews.
Thirdly, audio recordings were transcribed in "pidgin”,, the language of communication during the interviews. Translation into English was then done to ensure that those not familiar with pidgin were able to understand the information obtained from the discussion and interviews. Like the process used in transcribing, translation of transcripts from pidgin to English was done by members of the research team who were not part of the group that facilitated the interview/discussion. The transcripts were independently verified, checked for completeness and scanned to ensure all personal identifiers had been deleted.
Analytical approach
The socio-ecological model (9) was the analytical framework that guided the analysis. The analytical process began with data immersion, an essential step in the analytical plan. Immersion involved team members doing the following: (i) listened to audio recording of each interview (ii) read the field notes taken by interview team, (iii) read original verbatim transcripts and translations to ensure consistency with all the data sources and ensure familiarity with the data prior to identifying themes and developing a codebook. This facilitated the identification/validation of apriori themes that were initially developed by the lead author.
The development of the codebook followed a three-stage iterative and continuous process. The first stage involved an inductive-deductive analysis of the transcripts, which involved iteratively reviewing, interpreting and discussing verbatim texts of participants’ ideas, opinions, and experiences about cervical cancer. This allowed us to identify substantive themes that emerged from the focus group discussions and in-depth interviews. The second stage entailed a much more detailed analysis of the text, which resulted in identifying and developing codes that were organized based on the themes. The codes that emerged from this process were subject to discussion and review by the study team before they were applied to the transcripts. Disagreements regarding applied codes or whether codes fit into a theme of interest were discussed by the study team to arrive at a consensus. Finally, we mapped the results by each domain unto an ecological framework and organized the results by their relationship to micro, meso, and macro-structural factors regarding cervical cancer in the study setting.