Study design, study site and study population: This data was extracted from the clients’ service utilisation component of a larger data set generated to provide baseline information on FSW, MSM and PWID for programmatic purposes. Data was generated using a sequential mix method approach wherein findings from the quantitative data were explored further using qualitative data. Data were collected from a sample of FSW, MSM and PWID resident in four of the 36 states in Nigeria, namely: Enugu, Nassarawa, Benue, and Akwa-Ibom. The study sites were limited to these states because of the need to use the data generated from the study to plan and implement HIV prevention programmes for the target populations in these states as part of a targeted intervention for key populations funded by the Global Fund for AIDS, Tuberculosis and Malaria named the “Enhancing Key Population Intervention in Nigeria through Capacity Development (EKPIN)” project. The study adopted the client-centred access to health care framework developed by Levesque and colleagues [19], with the experience of care dimension of the WHO’S quality of care framework for maternal and newborn health care integrated to address the quality of care issues [20].
All study participants were 18 years of age or older and provided written consent for study participation. The extracted data consisted of both quantitative and qualitative relating to clients’ satisfaction with use of services offered by the public health sector and peer-led service organisations.
Quality Assurance: Selection of field workers was based on competence and experience. They were also trained before data collection commenced, and rigorously supervised during the data collection process. During the project, all field workers uploaded the quantitative data collected and transmitted audio-recordings of the interviews daily. The study coordinator reviewed documents and sought clarification and rectifications promptly. The study Co-PI also reviewed samples of the data.
Ethical Considerations: The Institute of Public Health Institutional Review Board gave ethical clearance for the conduct of the study (IPHOAU/12/268). Study participants were duly informed about the objectives of the study, risk and benefits, voluntary nature of study participation and freedom to withdraw from the study at any time. Written consent was obtained from all participants. No participant identifier was collected. The research outcomes was shared with 61 representatives of the MSW, FSW and PWID community, as well as policy makers and HIV programmers for result validation.
Data collection were carried out in locations considered as safe and private by community members. In all the states, a peer-led NGO setting was used for the data collection except when the respondent identified for the survey or in-depth interview specified alternative spaces that they preferred. Privacy was ensured throughout the study, and all participants were assured of confidentiality. All study participants signed an informed consent form, were provided with an educational material on HIV prevention and treatment and received N2, 000 ($11.50) for transport reimbursement along with condoms and lubricants.
Quantitative study
This aspect of the study sought to identify the proportion of FSW, MSM and PWID who had accessed HIV prevention services from the public health facilities and peer-led initiatives and their level of satisfaction with these services.
Sample size: A standard sample size calculation approach for cross-sectional study was used to determine the sample size for the study using WINPEPI software [21]: a minimum sample of 500 study participants was required (125 participants per state). A minimum of 50 FSW, MSM and PWID were to be recruited from each state respectively.
Participants’ recruitment and data collection: Study participants were recruited between April and June, 2015, from a minimum of two different parts of each states that had been mapped to have clusters of MSM, FSW and PWID. Decisions on the site for study participants was made based on the report of the key populations size estimation report [5]. The target was to recruit 25 FSW, 25 MSM and 25 PWID per each of the two sites in each of the four states.
A semi-structured questionnaire was administered by interviewers to study participants. The study instruments were developed in English. Questions were adapted from the instruments used for the National AIDS and Reproductive Health Survey. Key words and phrases, especially sensitive ones, were translated to the languages of each selected community and collected in a list generated during the training of interviewers. Interviewers used this material as a reference when in the field. A similar technique had been successfully used for past national reproductive health survey [22-26], and the study by Folayan et al [27].
First, study participants were asked if they had use public or peer-led facilities to access HIV prevention services. Those who had used services where then asked about their level of satisfaction with services received using a 5 point likert-like scale: 1 (very dissatisfied), 2(satisfied), 3(neutral), 4 (satisfied), and 5 (very satisfied). For the purpose of analysis, responses were collapsed into 3 categories: ‘satisfied’ and ‘very satisfied’ were categorised into satisfied, and ‘dissatisfied’ and ‘very dissatisfied’ were categorised as dissatisfied. Respondents identified how satisfied they were for three service quality indicators: (a) extent to which service providers listened to their problems and concerns; (b) extent to which service providers assured them of confidentiality and privacy; and (c) extent to which service providers respected their rights as a service recipient.
Next, respondents who had accessed HIV prevention services in public and peer-led organisation were asked to identify potential barriers to accessing these services. The variables specified were distance to services, cost, confidentiality, waiting time, hours of operations, staff attitude. Respondents also had the option of identifying if they perceived no barrier, and to list other potential barriers to accessing services beyond those specified in the questionnaire. Finally, respondents were asked about their willingness to continue to seek HIV prevention care in the public or peer-led facilities, orwould they rather seek for services elsewhere?
Data collection procedure: Four field workers and focal persons for MSM, FSW and PWID had a three-day training on the study objectives, methodology including ethical approach to data collection, and community entry procedures. Field workers recruited were individuals with prior experience conducting national reproductive health surveys and who understood the local language and the target community in each site. The field workers and focal persons pilot-tested the instrument and made recommendations for revisions prior to its printing. For each site, a local community guide was designated to facilitate community mobilisation and community entry processes. These included sensitisation of community members about the study objectives and methodology, making initial contacts with non-governmental organisations (NGOs) working with key populations in the state, and facilitating the recruitment process (locating of key populations hot spots and booking study-related appointments). A one-day training was conducted for the local community guides. The study in each state was overseen by an experienced researcher who serves as the state supervisor.
Study participants were recruited through snowball technique, specifically exponential non-discriminatory snowball sampling [28]. “Seeds” were recruited from members of the target groups, who then recruit other members from their personal networks. The seeds refer to the first contacts made for the snowballing. They were identified by state contacts engaged with the planning of the study. An initial three seeds were recruited, one for each group of key population. Once a seed completed filling the questionnaire, (s)he was asked to recruit three peers into the study. A seed was limited to recruiting a maximum of three peers to the study. This procedure was expected to help ensure a broad array of study participants had the opportunity to recruit for the study thereby promoting a wide spectrum of community members’ engagement with the study.
Data analyses: Univariate analysis was carried out to determine the proportion of respondents who have received HIV and sexual and reproductive health services before from public sector or a peer-led service, their level of satisfaction, and perspective on barrier to services.. Bivariate analysis was used to compare findings between the different key population groups using chi-square test and Fishers exact test where appropriate. Analyses were conducted using SPSS. Statistical significance was specified at a p value less than 0.05.
Quantitative study
This aspect of the study further explored perceived barriers and challenges to accessing HIV services from public and peer-led HIV service settings by key populations. Forty in-depth interviews (10 per state) were conducted with leaders of key population communities and 16 (4 per state) focus group discussions with key populations. The sessions were audio-recorded and hand-written notes taken. Basic socio-demographic data was collected for the participants, and the research personnel developed a detailed debriefing note and a summary report immediately after each interview or discussion session.
Participants’ recruitment and data collection: Participants in the in-depth interviews (IDIs) with identified community leaders who are highly knowledgeable on issues of key populations’ access to HIV prevention, treatment and care services. The community leaders in each state were identified through the national key populations secretariat. They were individuals who represent the communities in key strategic national and state policy and program meetings, and who the community recognise as their political representatives. The interviews solicited the participants’ perspectives about structural and other forms of barriers to key populations’ access to HIV prevention service. The IDIs were conducted using a semi-structured guide that was developed based on the study objectives, information from extant literature, and input from representatives of key populations.
With respect to the focus group discussions (FGD) in each state, one session was held separately for each group (FSW, MSM and PWID) and one joint session for the groups. Each session had between 6-10 participants. The sessions started with an ice breaker, a short briefing on the EKPIN project, and participants offered the opportunity to ask questions about the project and the research. A vignette was used to facilitate the discussions: vignette was used to allow participants to project their views rather than personalise the discussion. For each of the sessions, the discussants were guided to develop the profile and identity of the individual to be discussed in the vignette. This way, the names and sociodemographic profile of the individuals used for each vignette of the vignette was made appropriate for the focal key population group and the study location. The same issues were, however explored, in all the FGD sessions.
Data analyses: The qualitative data generated from the IDIs and FGDs included handwritten notes (the brief field notes, summary notes and debriefing reports), and transcripts from audio recordings. In-depth analysis was conducted using Atlas.ti software. The transcripts were analysed using the Grounded Theory approach. Three designated team members worked in collaboration with two expert qualitative data analysts to develop a codebook and guidelines suitable for use with the software package Atlas.ti. A selected set of texts were double-coded by two analysts to establish intra-coder and inter-coder reliability – a quality control measure that was automatically generated by Atlas.ti. Following this quality check, the coding team discussed the coding discrepancies and resolved them by consensus during several face-to-face sessions. The process continued until the inter-coder reliability was at least 80%. Afterwards, the remaining text was coded by one analyst with regular discussions among all the analysts for the purpose of standardisation and reliability.