Study design, population and setting
This was a mixed methods cross sectional study which employed both quantitative and qualitative data collection methods. Between June 2012- June 2017, we enrolled MRPs aged ≥18 years old into the study. The MRPs were enrolled through FSWs attending routine three monthly visits at the Good Health for Women Project (GHWP) clinic. The GHWP clinic is in a peri-urban suburb in the south of Kampala, the capital city of Uganda. The clinic was established in 2008 to study the epidemiology of HIV and STIs and to implement HIV/STI prevention among FSWs. HIV prevalence among FSWs in the cohort is 37% (7).
The eligibility criteria for participation in the study included: 1) being an MRP of a FSWs aged > 18 years old; 2) agrees to test for HIV; 3) MRPs who indicated that their relationship with the referring FSW had lasted for at least six months and if he had had sexual intercourse with a FSW within a month. We excluded non- consenting and mentally impaired participants.
Recruitment and study procedures
Participants were recruited through an on-going open cohort of FSWs established since 2008. Female sex workers are recruited by field workers and peer educators who approach them at their “hot spots”. Female sex workers attend three monthly clinic visits for routine HIV counselling and testing, free condoms and other services at each visit. During their follow-up visits, enrolled FSWs are screened for having an eligible MRP on an individual basis. The FSWs who may report having many partners, are asked to identify only one of their sexual partners with whom they have had sexual contact within the last one month, and their relationship has lasted for at least six months. The FSWs having eligible MRPs and willing to bring their MRPs inform them about the study using the “Information sheet” for referral. Potential MRPs are screened to determine their eligibility and to verify that they are the MRPs of FSWs. The MRPs who meet the eligibility criteria are asked if they would be interested to take part in a research study, including drawing blood for HIV testing and a 30-minute face-to-face interview about their sexual behaviours and other related information. The MRPs who were interested and eligible were enrolled into the study. Both HIV-negative and -positive MRPs were recruited into the study. Eligible and interested men were consented and counselled for serial rapid HIV testing (Determine, Statpak, Unigold) on a serum sample. HIV-positive men were assessed and enrolled into HIV care and 3 monthly follow up. Participants were also screened for STIs and those with STIs were offered free treatment according to Ministry of Health (MoH) clinical guidelines . Counselling on STI / HIV prevention was done at each visit and condoms provided.
As part of routine clinical care, blood samples were collected by an experienced and trained health worker for assessing HIV sero-positivity. Participants underwent confidential pre-test counselling and a free anonymous rapid HIV test. HIV diagnosis was made from 2 rapid tests in parallel (Determine, Statpak, Unigold [was used as a tiebreaker in case of discrepant results]) as per national guidelines . Participants received HIV post-test counselling and HIV test results on the initial visit. The HIV test results were linked to participants by their unique study identification number.
Data collection and study measures
After obtaining a written informed consent, eligible MRPs were consecutively enrolled and face-to-face semi-structured interviews conducted. Counsellors collected data on socio-demographic characteristics, sexual behaviour and clinical characteristics. The primary outcome of this study was HIV sero-positivity assessed as either positive or negative at enrolment. Socio-demographic measures included age, marital status, education level, alcohol use, and illicit drug use. Alcohol use was assessed by using a standardized WHO Alcohol Use Disorders Identification Test (AUDIT) . Alcohol use was classified into three categories i.e. harmless or low risk drinkers: score 1-7, harmful or high-risk drinkers: score 8-19 and alcohol dependent: score 20+. The sexual behaviour characteristics included number of sexual partners, paying food or money for sex and condom use at last sexual intercourse with FSW. The clinical characteristics included presence of STI symptoms in the last 12 months. Presence of STI symptoms was determined if an MRP self-reported having had any symptoms suggesting STIs in the last 12 months including urethral discharge, genital ulcer, and pain during urination.
Quantitative data analysis
Data were double entered in OpenClinica (OC), cleaned, and exported to Stata15.0 (StataCorp, College Station, TX, USA) for analysis. We resolved discrepancies by checking the source documents for clarification. Categorical demographic and clinical characteristics were summarized by counts and percentages. Continuous variables were summarized by means and standard deviations or medians and interquartile ranges. The proportion with HIV sero-positivity was analysed by the different demographic, sexual behaviour and clinical characteristics. Factors for which the association attained statistical significance on log likelihood ratio test (LRT) of p<0.20 were selected for the multivariable logistic regression model. Logistic regression models were fitted to identify factors associated with HIV sero-positivity at unadjusted analysis. Factors were retained in the final multivariable logistics regression model if their inclusion did not make the fit of the model significantly worse at the 5% level on a likelihood ratio test (LRT).
Sample selection and recruitment
Informed by a study conducted in Kenya  and based on our experience of conducting qualitative studies, a sample of 34 MRPs was selected, a number which was anticipated to allow us achieve saturation. The participants were part of the quantitative component and were consecutively recruited into this study. Potential participants were identified using the clinic register compiled during the period June 2012- June 2017, as part of the main GHWP MRP cohort. The interviewer then compiled a list and the selection of each participant was based on choosing a range of people representing participants of the different age, educational level, duration in the cohort, and geographical location. Purposive sampling was used to recruit as varied a sample as possible in order to gather a wide range of responses. Purposive sampling was used to ensure maximum variation of the sample and experiences. MRPs were approached by the social scientist, given information and invited to participate during their clinic visit. MRPs in the community were contacted by telephone to ascertain their willingness to be interviewed. Those who responded positively, were invited to participate in the interviews either at the participants’ homes, or in a private place at GHWP clinic, depending on the preference of the MRP.
Data collection procedure
Face-to-face in-depth interviews were conducted by an experienced social scientist in the participants’ preferred language (English or Luganda). The interviews lasted between 45 – 60 minutes and were not audiotaped as was the preference with most of the participants. Instead, detailed notes were written out immediately after each interview and these formed the basis of the analysis. Such a practice has been found to be effective data . An in-depth interview guide was used to elicit background information on participants’ knowledge of HIV and its prevention as well as sexual risk practices. The following topic areas were explored flexibly throughout each interview; knowledge on HIV and the prevention strategies; money, love, sex and HIV risk; multiple sexual relationships; HIV and condom use; alcohol and illicit drug use. The interviews were conducted at times and in quiet places that were convenient to the participants either at the clinic or away from the clinic with written informed consent from the participant.
Qualitative data analysis
The data was analysed thematically using a Framework Analysis approach . Two social scientists went through the scripts to identify common themes which were linked to the research questions. They then coded the data around these themes and transferred these by copying from text and pasting in an excel table in order to identify patterns, associations, and even concepts in relation to the main topics discussed with the MRPs.