Factors Associated With Willingness to Use Oral Pre-exposure Prophylaxis (Prep) Among Fishing Communities in Peri-urban Kampala, Uganda

Background: The World Health Organization (WHO) recommends the use of pre-exposure prophylaxis (PrEP) in key populations at elevated risk for exposure to HIV. If used effectively, PrEP can reduce annual HIV incidence to below 0.05%. However, PrEP is not acceptable among all communities that might benet from it. There is, therefore, a need to understand perceptions of PrEP and factors associated with willingness to use PrEP among key populations at risk of HIV, such as members of communities with exceptionally high HIV prevalence. Objective: To examine the perceptions and factors associated with willingness to use oral PrEP among members of shing communities in Uganda, a key population at risk of HIV. Methods: We conducted an explanatory sequential mixed-methods study at Ggaba shing community from February to June 2019. Survey data were collected from a systematic random sample of 283 community members. We carried out bivariate tests of association of willingness to use PrEP with demographic characteristics, HIV risk perception, HIV testing history. We estimated prevalence ratios for willingness to use PrEP. We used backward elimination to build a multivariable modied Poisson regression model to describe factors associated with willingness to use PrEP. We purposively selected 16 participants for focus group discussions to contextualize survey ndings, analysing data inductively and identifying emergent themes related to perceptions of PrEP. Key results: We enrolled 283 participants with a mean age of 31±8 years. Most (80.9%) were male. The majority of participants had tested for HIV in their lifetime, but 64% had not tested in the past 6 months. Self-reported HIV willing PrEP. focus group described pill burden, side-effects, drug safety, and accessibility as potential barriers to PrEP use.


Introduction
Of the 1.7 million new HIV infections that occurred globally in 2019, 62% were among key populations and 28% of these infections occurred in key populations of East and Southern Africa [1]. In Uganda, as in many settings, the HIV burden is mostly concentrated in key populations [2]. In 2019, the average national HIV prevalence rate among adults aged 15-49 years was 5.8% [3], while 15%-40% of adult sher folk were living with HIV [4]. Fisherfolk in Uganda are at elevated risk of exposure to HIV due to high degrees of mobility, poor access to information about HIV, and limited access to HIV prevention resources, including oral pre-exposure prophylaxis (PrEP) [5].
Daily use of oral PrEP is reported to reduce the risk of HIV transmission through sex by 75-99% [6,7].
Based on this evidence and following the 2015 WHO recommendation [8], the Ministry of Health rolled out oral PrEP (300mg of Tenofovir daily or 300mg of Lamivudine daily) to key populations in Uganda, including sherfolk beginning in August 2017 [9]. To date, an estimated 21,000-22,000 members of key populations in Uganda including but not limited to sex workers, long-distance truck drivers, barmaids, and discordant couples have accessed oral PrEP [10]. However, studies show that poor user knowledge, negative community perceptions toward the drug, stigma, cultural beliefs, and low perceived risk of HIV transmission continue to dampen acceptability of oral PrEP leading to poor adherence and reduced effectiveness [11][12][13]. Studies outside Uganda have found levels of awareness of PrEP as low as 29.7% among HIV key populations, and acceptability ranging from 35.4-64.4% [14,15].
Though shing communities are among the key populations targeted for oral PrEP, to date, the majority of these communities have lacked access to PrEP services due to structural barriers and little is known about their receptivity to PrEP [16]. We aimed study to explore the dynamics of PrEP acceptability among sherfolk and to identify the factors necessary to improve PrEP uptake in this key population in order to inform countrywide PrEP rollout in Uganda.

Materials And Methods
We carried out a cross-sectional study using an explanatory sequential mixed-methods design, integrating survey data and focus group discussions collected from 5th February to 4th June 2019.
The study was conducted in Ggaba, a semi-urban landing site on Lake Victoria located in Makindye division of Kampala district, the capital of Uganda. Ggaba has a population of around 17,000 people [17] with the major economic activity being shing. The population is highly transient and has a high HIV transmission rate of 3.39 per 100 person-years at-risk compared to the general population [18,19].
We included adults aged 18 and above who had been residents at the Ggaba landing site for at least 3 months and provided informed consent to participate in the study. We excluded individuals who were too ill to participate in the study.
The study was powered around the primary quantitative research question, which aimed to determine the proportion of individuals accepting oral PrEP. We used a systematic sampling method for the survey. We estimated a population of 1000 adults at the landing site and generated a sampling fraction of four with a random number between one to four used to generate the sampling interval of three. We, therefore, included every fourth adult was included in the survey beginning with the third adult until the sample size was attained.
Trained interviewers administered a semi-structured questionnaire with both open and closed-ended questions using Open Data Toolkit (ODK) in English and/or Luganda, the local language. The questionnaire was adopted from previous literature about acceptability of PrEP use and pretested on 15 random participants who were not included in the study. Acceptability was de ned as self-reported willingness to use PrEP when provided, and was measured using "Yes" and "No".
For the qualitative component, we further explored acceptability and perceptions of oral PrEP use. We used maximum variation purposive sampling to select participants for focus group discussions (FGD). To facilitate exploration of disparate views and generate point-counterpoint discussion and resolutions, we used FGDs for general discussion on the subject. We ensured relative homogeneity in age, sex, and duration of stay at the landing site when recruiting the 16 participants for each of the two FGDs (8 per FGD).
FGD data were collected using a focus group moderator guide, conducted in the local language, audiotaped and later transcribed and translated to English by the PI (BS) as part of standard operating procedures such that they have retained their meaning.

Statistical Analysis
We produced descriptive statistics such as proportions and mean ± standard deviations for all variables.
The primary outcome was willingness to use oral PrEP. Acceptability to use PrEP was measured as a proportion with its 95% con dence interval. A modi ed Poisson regression was used to estimate the prevalence ratios (PR) for bivariate and and adjusted prevalence ratios (aPR) for multivariate analysis after assessing for assumptions of multicollinearity (by estimating VIF and leverage) and outliers were then tested. The adjusted prevalence ratios were assessed for confounding by 10% change in the crude and adjusted prevalence ratios. Analyses were done using Stata Version 15.1/MP. Qualitative data from the FGDs were transcribed and entered into Atlas.ti 8.3 for analysis. The lead author (BS) read all transcripts and used an inductive, iterative approach to analyze their content, applying three cycles of coding according to the study purpose statement. Emerging themes on perceptions about PrEP were grouped and discussed with other authors (MAH, AK) between each coding cycle.

Description of study participants
Of the 283 participants, 80.9% were male and 82.7% were employed ( Table 1). The mean age of participants was 31 ± 8 years. Self-reported HIV prevalence was 6.4% and 64% had not tested in the past 6 months, while 62.4% had concerns about acquiring HIV. Factors associated with willingness to use PrEP A high proportion (80.6%, 95% CI 75.5-85.0) of participants reported that they were willing to use PrEP if it were provided. In bivariate analyses, age (P = 0.001), location (P = 0.047), and education level (P < 0.001) were signi cantly associated with willingness to use PrEP (Table 2). Those who reported regularly testing for HIV were more likely to say they were willing to use PrEP compared to who tested irregularly (92.2% vs 74.0%, P < 0.001). The proportion of men who were willing to use PrEP was higher than that of women (83.0% vs 70.4%, P = 0.078). Having tested for HIV in the last 6 months (P < 0.001), HIV status, perception of HIV risk (P < 0.001), concern about HIV (P < 0.001), condom use (P = 0.002), and awareness of PrEP (P = 0.024) were also associated with willingness to use PrEP. HIV status, HIV infection concern, and condom use were dropped at multivariate due to multicollinearity In the nal multivariable model, three factors were signi cantly associated with willingness to use PrEP among shing communities: higher levels of education, perceived risk for HIV, and history of testing for HIV within the last six months (Table 3). Compared to those with no formal education, participants who had attained tertiary education were nearly twice as likely to say they were willing to use PrEP (aPR 1.  "I was happy when I heard about the drug because it will help us most especially adolescents who are sexually active."(R8/FGD2) However, participants expressed concerns about adverse effects from the drug, which they perceived to be likely. They anticipated that such side effects would stop them from carrying out their normal daily activities. On lady explained that such side effects could decrease motivation to continue using PrEP: "Maybe if we use it makes us dizzy or when it exhausts your energy, you may say that am not sick why I am bothering to take it, but if it has no problem it brings to your body then I just swallow it and I eat my Others were concerned about the long-term effects of a preventive medication. While asking them about the major concerns about the drug, another explained; "… doctor, this is the reason why we may not take it, they told us that when you swallow that drug when you don't have the disease you get cancer inside the stomach especially where that drug sits because it doesn't have what to treat in your body."(R4/FGD3) Fishermen were also concerned about possible changes in lifestyle PrEP might introduce. They explained; "I was told it makes you lose your appetite, loose sleep when you are start taking it."(R8/FGD3) "If they are to stop us from other foods and drinks like alcohol."(R4/FGD3)

Discussion
In this expanatory-sequential study, most sherfolk said they were willing to use oral PrEP. This is consistent with previous studies assessing the acceptability of PrEP among other key populations in East and Southern Africa [21]. In a discrete choice experiment to assess the acceptability and potential uptake of PrEP, others have also found oral PrEP to be highly acceptable in principle among shing communities in Uganda [22].
In our study, PrEP was most likely to be acceptable among those who perceived themselves as being at high risk of exposure to HIV. This link between risk perception and acceptability to use PrEP also emerged in FGDs, where participants described their communities as being "badly off" with regard to HIV. High self-reported HIV risk perception among key population communities [23] could explain the high willingness to use oral PrEP among key populations.
We further found that having tested HIV in the past 6 months was associated with a 13% increase in likelihood of being willing to use PrEP. Similarly, perceiving oneself to be at risk of HIV infection was associated with a doubling of the likelihood of being willing to use PrEP compared to those who did not perceive themselves to be at risk.
In focus group discussions, participants perceived PrEP to be protective and necessary to defend their lives from HIV. This suggests that people in shing communities who perceive themselves as being at risk of getting HIV and have regular HIV checkups are more likely to accept PrEP if provided to them. Indeed, we found regular HIV testing to be a predictor of PrEP acceptance in survey responses.
We also found that respondents who had completed primary education were 61% more likely to express willingness to use PrEP compared to those with no formal education. There was a stepwise relationship between educational attainment and willingness to use PrEP: secondary-level was associated with 72% increase in willingness while the tertiary level was associated with 97% increase in willingness. This is likely because those who were more educated had greater access to information and a better understanding of HIV risks, rendering them less prone to misconceptions from peers [24,25].
Better understanding of oral PrEP may increase willingness to use it. Therefore, interventions to educate the communities on PrEP may be of importance to facilitate acceptance of the drug. This nding is consistent with studies conducted in other key populations where those who were more educated believed that PrEP would reduce HIV risk to their partners and would accept it and recommend it [12,26,27].
Our study has some limitations. The cross-sectional design we used could not be used to analyze behaviour or attitudes over a period of time.
Secondly, we measured only hypothetical acceptability of oral PrEP rather than preferences after the use of the actual tablets.
Our study also has several strengths. First, the study participants were systematically sampled from its shing sites to ensure high representativeness of the community.
Second, we sequentially employed both quantitative and qualitative methods to gain a more thorough understanding of the factors associated with willingness to use oral PrEP among this community.

Conclusions And Recommendations
Oral PrEP was widely acceptable among members of shing communities in peri-urban Kampala. From the FGDs, we also learned that PrEP is considered as a life saver to the shing communities due to their lifestyle. However, they also had concerns about side effects, change of their lifestyles when taking PrEP and interruption in their normal daily activities. Programs for scaling-up PrEP for sherfolk should merge