Uptake of oral-based HIV self-testing among Key Populations taking HIV pre-exposure prophylaxis in central Uganda

Introduction: HIV incidence is higher among key populations such as sex workers, and who HIV self-testing among key populations taking HIV Prophylaxis has not been studied in Uganda yet it could be an alternative testing approach for these population categories. We assessed the uptake of oral based HIV self-testing among key populations enrolled on Pre-Exposure Prophylaxis in Uganda. Methods : A mixed methods cross-sectional study was conducted on 366 key populations attending Pre-Exposure Prophylaxis clinics at MARPI Mulago and Kasensero HC II Uganda, during May-August 2018. Respondents aged 15 years and above with no prior history of HIV self-testing were issued with “Oraquick” HIV rapid diagnostic test (kit) for self-testing. Data were collected using an interviewer administered structured questionnaire and through in depth interviews of selected respondents. A chi square test was performed to determine association between socio demographic, economic factors and propensity to uptake HIV self-testing. Logistic regression was used to determine the predictors of HIV self-testing uptake. Thematic analysis was performed to explore factors that promote and inhibit HIV self-testing uptake. Results : HIV self-testing acceptability was 100% (365/365) with 73% (n=265) of respondents willing to pay for a kit. Of these, 85%(n=227) would pay not more than 1.4USD. Sixty-six percent (n=243) of respondents had high propensity to uptake HIV self-testing. Predictors of uptake were clinic location (rural vs urban) (aOR = 17.63 95% CI 8.44 – 36.81, p<0.001); Key Population category (female sex worker’s vs others) aOR= 4.36 (95% CI: 1.63 – 11.66, p=0.003) and education level ( < primary vs post primary) aOR=0.38 (95% CI: 0.20 – 0.73). Using an oral fluid based kit “Oraquick” was reported to be free of pain, convenient, easy to use and time saving. A multimodal approach to distributing HIV self-testing kits was suggested by respondents. Conclusion self-testing

Human Immunodeficiency Virus (HIV) remains a global health threat. Globally, by the end of the year 2016, there were about 36.7 million people living with HIV and in the same year, approximately one million people died of HIV related illnesses (1). More HIV cases are found in Sub Saharan Africa with approximately 25 million people in the region living with HIV (2). It is estimated that globally, 70% of all people living with HIV know their status, and as the world prepares to end the HIV pandemic by 2030, the remaining undiagnosed proportion of 30% needs to be identified, initiated on treatment, and virally suppressed (1).
In Uganda, HIV is the 3rd leading cause of death following neonatal disorders and diarrhoeal diseases (3). The HIV epidemic is considered generally stable for past 7 years (4). However, despite the improving trends, Uganda is still among the countries in sub Saharan Africa categorized as high "burden" in respect to HIV prevalence (5). Most of the new HIV infections in the country are reported to occur in Key Populations (KP) which include sex workers, fisher folks, long distance truck drivers, uniformed service personnel, men who have sex with men (MSM), and boda-boda taxi-men (6). including HIV acquisition, drug toxicities, and cessation risk of exposure to HIV. It is hoped that HIV Self Testing can be ideal in such a category of people who have to frequently retest for HIV during the follow up period.

Overview of HIV Self-Testing (HIVST)
Human Immunodeficiency Virus (HIV) Self Testing (HIVST) is where a person collects his or her own specimen (oral fluid or blood) and performs an HIV test and interprets the result, often in a private setting, either alone or with someone he or she trusts (7) . The WHO in 2015 recommended the use of HIVST as an additional approach to HIV testing services (8). This declaration was graded as a strong recommendation, moderate quality of evidence and called for more research in the field.
HIVST is believed to be a good strategy for increasing HIV testing uptake (9). A number of factors such as its ability to increase convenience, offer privacy, save time and minimise stigma are attributed to its possible success as observed in a study by Walensky and Bassett (10). Oral HIVST has been associated with additional advantages including; the oral fluid-based self-test being noninvasive, ease of use of the test, and having no fear of pain (11). Nevertheless, like any other approaches, HIV self-testing has been attributed to some barriers such as: unease of use, in case the particular test may be difficult to use, increased costs where the tests may have to be bought by the users, and accessibility to professional support in case one has concerns about the test (10,12).
Other challenges as noted by Figueroa et al. (13) include; lack of counselling where a person may have psychosocial concerns, and accuracy concerns since the test is not performed by a health professional. For effective implementation and optimisation of the outcomes of HIVST, these barriers require to be adequately addressed.

Acceptability of HIVST
Overall, studies conducted worldwide about HIV self-testing have shown high acceptability, with varying levels among population categories, age and gender. A literature review undertaken by Figueroa,Johnson (13) on attitudes and acceptability of HIVST among key populations indicated that in 8 out of 14 included articles, an average acceptability level for HIVST of 67% was found and, overall, acceptability ranged from 21-98%. In a qualitative acceptability study on HIV self-testing among 144 participants, including transgender women, men who have sex with men, and female entertainment workers in Cambodia by Pal and colleagues (12), all 144 participants expressed acceptability of HIVST. Another study conducted in USA, by Katz and others (14), to assess the acceptability of home-based HIVST revealed an acceptability of 84% in men having sex with men.
Research has also shown high acceptability levels of oral HIVST. Kurth and others (15), in their study to determine the accuracy and acceptability of oral HIV testing in Kenya concluded that the test was highly acceptable in the general population, represented by an acceptability of 94%. Similarly, another study from Kenya by Mugo et al. (16) on the uptake and acceptability of an oral HIVST among pharmacy clients showed a high level (84%) of uptake of the test. Another related study conducted in Uganda to determine the accuracy and acceptability of rapid HIV oral testing revealed a rate of acceptability of 87% among 440 clients attending an urban health facility in Kampala (17). As these studies were conducted from a similar setting, they are likely to offer some understanding on the acceptability of oral HIVST in the Ugandan context. Nevertheless, all of these studies focused on the general population, and not on any particular key population. Contextualised data is still required to provide a picture of acceptability of the test in key populations in Uganda.

HIVST among PrEP users
Pre-Exposure Prophylaxis is one of the biomedical interventions put in place for the prevention of HIV. It involves taking ART by a person who has continuous exposure to HIV but is HIV negative at the time of starting on PrEP (6). The aim is to prevent HIV infection in those taking the medicines. It is recommended that clients on PrEP are regularly monitored for HIV acquisition every after three months (1), to detect HIV infection as early as possible and start the client early on long life HAART (6).
In Uganda, over 80% of clients on PrEP belong to 3 categories: female sex workers, MSM, and injectable drug users (IDUs) (6). Such key populations are characterised by poor access to HIV testing due to a number of causes including sigma in the public, lack of key population friendly services, and the nature of their work which does not easily fit into the existing public HIV services delivery structures. Due to the advantages of HIVST listed in the previous section (such as ease of administration, privacy, and time saving), it is likely to be an appropriate method to scale up HIV testing among the key populations, the majority of which use PrEP regularly due to their HIV exposure risk, hence the need to evaluate HIVST among this population.
HIV self-testing has been reported to be highly acceptable among PrEP users. A recent study conducted in Kenya on the feasibility and acceptability of HIV self-testing among 240 PrEP users in Kenya revealed that 98% of the clients accepted to use the self-testing kit (18), representing high acceptability of HIVST among PrEP clients. This study applies well to the Ugandan setting, although the PrEP clinic studied only involved discordant couples, yet in Uganda discordant couples undergoing PrEP represent only about 2% of the entire clients on PrEP.

HIVST KITS distribution approaches/ models
For HIVST kits to reach the intended users, innovative distribution approaches need to be identified and established. Various models of HIVST kit distribution have been evaluated to establish the preferences of the users. Geng and his colleagues (9) studied the direct provision versus facility collection models of HIV self-tests among female sex workers in Uganda. Their findings indicated that HIV self-testing was more likely to be successful if kits were distributed by peers compared to health facility pick up. The direct provision model was more likely the priority due to the privacy it could offer for the sex workers and could be a better choice for other key populations likely to be stigmatised by society.
In summary, literature shows that HIVST is acceptable. What is not well documented however, is the acceptability of HIVST among individuals taking PreP. Our study aims to address this evidence gap.

Study justification
Key populations on PreP in Uganda are required to retest for HIV every after 3 months. The only existing HIV testing method is by use of a health worker provided blood based test either at the facility or in the community during outreaches. Providing a variety of HIV testing kits to the public enables people to choose which kit to use and this can enhance HTS uptake of HIV testing.
Acceptability for HIVST in Uganda has been studied in a few settings as part of determining the diagnostic accuracy of self-testing kits in key populations as well as in the general population. There is no study in Uganda that has determined acceptability of HIV self-testing using an oral based kit among PreP users. Our study aimed to determine the uptake and factors associated with oral based HIV self-testing in KPs receiving PrEP at two of the six PreP clinics in Uganda. The study is hoped to contribute to generation of evidence that can be used to inform policies on HTS programming.

1.
To determine the proportion of Key Populations taking Pre-Exposure Prophylaxis for HIV in central Uganda who accept oral-based HIV Self-Testing.

2.
To establish factors associated with HIVST uptake among Key Populations taking Pre-Exposure Prophylaxis in central Uganda.

3.
To determine the preferred HIV Self-Testing kit distribution approaches/models among Key Populations taking PrEP in central Uganda

Study design
This was a cross sectional study that employed both quantitative and qualitative research methods.

Study setting
The study was conducted at two clinics in Uganda

Study participants
The study targeted all people categorized as key populations who had been enrolled and active on Therefore, a total of 367 participants was considered for the study.

Sampling
For logistical reasons as well as due to the nature of the study population, we employed a consecutive convenience sampling strategy. Potential participants were screened for eligibility as they came for their respective clinic appointments until the desired sample size was achieved for each clinic to make a total of 366 respondents. Most respondents were identified from outreach clinics since both PrEP clinics use targeted outreaches to recruit clients for PrEP as well as to provide drug refills.
Purposive sampling was used for the qualitative study and respondents were drawn from those who had attended the quantitative study, after giving consent to participate in the qualitative interview.

Dependent variables
The primary outcome (dependent) variable for the study was propensity for HIVST uptake.
To estimate uptake of HIV Self testing, we used five parameters to calculate a composite score "propensity for HIV testing uptake". The variables were: acceptability of HIVST, testing kit preference, willingness to pay for the kits, amount to pay for the kit and preferred testing kit distribution approach/model. Uptake was categorised as low or high using a cut off score of 6.

Independent variables i. Socio-demographic Characteristics
Age, gender, education level, occupation, marital status and religion were assessed.

ii. Level of income
Level of income was assessed by asking participants the approximate amount of money each earns per month from all sources, captured as a range of sums in Uganda shillings.

iii. KP category
Respondents were categorized as follows: female sex workers, negative partners in discordant sexual relationships, men who have sex with men, people who use and inject drugs, fisher folks, transgender and adolescent girls and young women (who were in essence part of FSWs). Each respondent was captured only once even if he/she belonged to more than one of the above categories

Data Collection
Quantitative data was collected using an interviewer administered structured questionnaire. The questionnaire was in English language. For clients who did not understand English language, questions were directly translated in the preferred language during the interview session. In-depth individual interviews were used to collect the qualitative data after collection of the quantitative data.
An interview guide with open ended questions was used to interview the participants. Data collection took place from May to August 2018.

Data collection procedures, data sources and measurement
Quantitative data was collected using an interviewer administered structured questionnaire (Appendix 1). The questionnaire was in English language. For those clients who did not understand English language, questions were directly translated in the preferred language during the interview session.
Research assistants were trained for two days by the principle investigator before data collection.
Data was cleaned every day and kept in water proof folders. The data questionnaire was pretested on a population similar to the target population before setting off to the field. Participants for pretesting were selected from MARPI Mulago clinic. These clients were not included in the actual study. Minor edits to the tool were made and majority were mainly spelling and grammatical errors.
Research assistants educated study participants on how to use an Oral HIVST kit (Oraquick) by use of a video and practical demonstration. Leaflets guiding HIVST using the Oraquick in a local language and English were provided. Such information was included in the leaflet insert (pre-packaged) in the testing kit pack. Information giving to the clients included an account on the availability of the standard testing services at the health facility. Clients were first taken through a demonstration on how to perform self-testing. This demonstration involved three steps of preparing for oral fluid sample collection, the actual process of using the kit to test, and finally interpreting the results. Clients were asked if they were willing to perform a self-test and those willing were provided with a kit to do selftesting (accepting). Research assistants supervised self-testing to ensure adherence to standard operating procedure as a quality assurance measure.
After self-testing, respondents were asked to interpret the results and thereafter, interviewed further to answer the study objectives. All respondents were willing to perform a self-test. Those who tested HIV positive on self-testing were subjected to a confirmatory test using the national HIV testing algorithm. Respondents who were HIV positive on confirmatory testing (4 in number) did not continue with the questionnaire.

HIV testing kit type preference
We measured this variable by interviewing participants and asking them questions regarding the preference for routine 3 monthly follow up HIV testing. Options were either using a self-testing kit they had which had been issued to them, using a blood based self-testing kit if introduced in future or using the conventional health facility based-health worker provided HIV test.

HIVST kit distribution model preference
Participants who preferred self-testing were asked how they would wish to access the testing kits through an interviewer guided questionnaire. Responses were graded as choices of "either," or "or "for each distribution model.

Willingness to pay for a test kit
Participants were asked whether or not, they would be able to pay for an HIVST kit if it was at a cost.
In addition, those that answered "yes" were asked how much they would be willing pay and responses were captured in Uganda shillings (UGX) and converted in USD equivalent during analysis.

Addressing potential sources of bias
Our study being a cross sectional study had potential to face non response. During sample size estimation, we included a non-response rate of 20%. We measured variables that did not require long time information recall, hence recall bias was minimised. The proportionate allocation of study subjects between the two PreP clinics minimised selection bias.

Quantitative variables analysis and statistical methods
Quantitative data was analysed using STATA software version 14.2. Acceptability to HIVST was calculated as a percentage of those who performed self-testing out of those who were offered the testing kit. Client characteristics were summarized using descriptive statistics by computing proportions for categorical variables. Continuous variables were summarized by computing the mean, median, mode and range. Age was grouped into 7 categories with a 5-year age range for the first 6 categories and those 45 years and above for the seventh category. This 5year age categorisation aimed at establishing if age would influence acceptability of HIVST Data for specific objectives was analysed as indicated below:

What is the proportion of KPs taking PrEP in central Uganda who would accept HIVST?
We analyzed this objective by computing the proportion of clients who performed HIV self-testing over those who were issued with an HIVST kit. The proportion was computed as follows: Number of respondents who performed a self-test X 100 Number of clients who were issued with a self-testing kit There was no missing data in all data capture tools during this study

What factors are associated with HIVST uptake among KPs taking PrEP in central
Uganda?
In order to answer this study question, five variables (HIVST acceptance, kit preference, willingness to pay for the testing kit, the amount of money one is willing to pay, and the model preferred for kit distribution) were used to derive a composite variable (propensity for HIVST uptake) graded as high or low. This score was used as the dependent variable in estimating HIVST uptake using social demographic and social economic characteristics of the respondents.
Hence, somebody who accepted HIVST, prefers oral fluid-based, is willing to pay, can pay more 5000 and prefers to buy from health facility, the score would be 9. Thus, someone with a score 0 would have the lowest propensity to uptake HIVST and a highest propensity if he/she scores 9. A cut off point of 6 (median score) was used to define high and low propensity to accept HIVST.
Demographic and socioeconomic characteristics were analysed against the propensity score using logistic regression (coded 1 if propensity score is greater than 6 and 0 if otherwise). Factors with p values less than 0.2 at bi-variable analysis were subjected to multivariable analysis producing adjusted estimates. Associations with p values less than 0.05 were considered to be statistically significant. Odds ratio was used as a measure for association.

What are the preferred HIV Self-Testing kit distribution approaches/models among Key
Populations taking PrEP in central Uganda?
We answered this objective by computing the proportion (percentage) of clients who preferred a particular HIVST kit distribution model of choice over others. Such models were: picking kits from a health facility, having kits delivered home by health workers, having kits delivered home by peers or buying from a private facility. We used the following formula to get the proportion for each model.
Number of respondents who prefer a given HIVST kit distribution model X 100 Total number of respondents The qualitative aspect of the study The qualitative study employed an in-depth interview technique to gather data from participants and followed the quantitative data collection. We interviewed 20 participants with effort to ensure variation between KP categories, gender and testing site. Purposive sampling was used to select the participants and the principle of data circulation (where no new insights emerge from the data) was used to guide participant enrolment into the study.
An interview guide with open ended questions was used to interview the participants. The tool was in English but questions were directly translated in the preferred language during the interview session.
To ensure the appropriateness of the data collection instrument, it was piloted on 2 individuals (10 % of the study the qualitative study sample) with similar characteristics to the study participants.
Consent was sought before interviews were conducted and all interviews were audio recorded.
Recorded information was transcribed verbatim and important statements were identified and extracted. The thematic analysis approach was employed for data analysis, where statements/words were grouped according to similarity in meaning and into meaningful labels (Codes). These labels were eventually grouped into themes, and finally into categories. Categories were presented in a narrative form and supported with verbatim quotes from participants.

Participants
A total of three hundred sixty-seven (367) clients were reached for interviews but one client opted out leaving 366 study participants. The client who opted out preferred not to be interviewed for personal reasons. Out of the 366 clients who were willing to perform HIV self-testing, 365 performed the test translating into 99.7% acceptability. Of the 365 clients who performed a self-test, 5 clients tested HIV positive on self-testing. Both the negative (360) and the positive (5) clients were subjected national HIV testing algorithm, and 4 were confirmed HIV positive. All the positive were linked to Anti-retroviral Therapy (ART) clinics for initiation on treatment.

Socio demographic and economic characteristics of participants
Three hundred sixty-six individuals were in the study. The mean age was 28 (SD=7.5) with the youngest participant at 17 years and the oldest at 54 years. Of these, 54.1% (n=198) were from urban settings, 49.2% (n=180) were males, and 48.4% (n=177) had attained up to primary education.
Casual labor was the major source of income contributing to 62.6%. (n=229) as detailed in the table below.

Acceptability of HIV self-testing
Of the 366 participants, 365 (99.7%) were willing to perform self-testing. One client declined to take the kit. All the 365 performed the self-test resulting in 100% acceptability.

Distribution of the propensity score
The median propensity score for HIVST was 6 (interquartile range, IQR: 5, 7) on a scale of 0 to 9 with the latter indicating the highest propensity. The median was used as a cut off to indicate the degree of HIVST uptake (high and low propensity  i.e, the odds of propensity for uptake of HIVST among those with lower than secondary school education were 0.38 times those with secondary education and above. The odds to uptake HIVST did not significantly vary by age, marital status, religion, source of income or monthly income levels at 5% significance level.

Proportion of KPs taking PrEP in central Uganda who prefer specific HIV Self-Testing kits distribution approaches/models.
Majority of respondents (88.4%, n=320) prefer to use an oral fluid based HIVST for routine 3 monthly HIV testing while on PreP, 73.2% (n=265) were willing to pay for an HIV self-testing kit and of those willing to pay, 85.7% (n=227) were willing to pay an amount not exceeding 5000/ Uganda shillings.  Figure 4 shows the distribution of all respondents by model of test kit preference and gender.
Majority of KPs (85.7%, n=227) were willing to pay for kit at fee not exceeding 5,000 UGX (USD 1.4) Figure 4 shows that more male respondents compared to female prefer home delivered kits (37.8% vs 34.6%) followed by picking from health facilities (29.4% vs 28.7), as well as buying from private facilities (9.4% vs 7%). Females preferred, (compared to men) distribution of the kits by peers (25.4% vs 19.4).

Kit distribution model preferences
FSWs and AGYW (who were in essence a sub category of FSWs) would use multiple distribution approaches to access the kits. Discordant couples however prefer picking kits from health facilities (21.7%, n=23), PWUID mostly prefer kits to be delivered to them by health workers (20.5, n=27), and fisher folks prefer kits delivered by peer (30.5%, n=35).

Main findings
Four main themes emerged from the qualitative study which were generally aligned to the study aim.
These were: acceptability of HIVST, choice of an HIV testing approach, preferred kit distribution models and willingness to pay for the test, as presented below.

Acceptability of HIVST
All (20)  The time factor was particularly an important concern for special categories of people, such as sex workers, who work during night and sleep during day hence find day time very valuable. Additionally, some participants felt HIVST was particularly important for special groups who feel stigmatised when they go to health facilities. Particularly, FSWs and MSMs reported this concern and felt society has not accepted them and discriminates them, hence found HIVST more convenient. However, the emotional concerns were expressed by fewer (6) participants compared with those (14) who felt that was not a problem to them. Nevertheless, this concern together with the concern of linkage to care should not be taken for granted and therefore need further exploration, as this study did not have an in-depth evaluation of them, yet these have been reported in literature as significant concerns of the HIVST approach.

Choice of an HIV Testing Kit during routine 3 monthly follow up HIV testing for clients taking PreP
Participants Other participants expressed a fear of continued blood loss during HIV testing procedures that involved blood based testing. People on PrEP were particularly concerned as they frequently underwent HIV testing, on a three monthly basis. Some felt that the continued blood loss could cause them health complications, yet with the oral fluid based test, no such fears could arise. In terms of ease of use, participants found the oral based HIVST easier compared with the other approaches. For example, participants explained that the test procedure was easy for lay persons compared e.g. with the finger prick based HIVST, where the technique of drawing blood could be more difficult. Whereas this concern could result from lack of training in the finger prick approach, some felt even with training, the finger prick would still be more complex as it required one piercing themselves (which many also be feared) as opposed to only paring a test instrument around the mouth, a procedure many compared to brushing of the teeth. Hence, the majority felt the oral fluid based HIVST approach was more suitable to lay people compared with the finger-prick approach.

And sometimes you may use it (the finger prick based test) wrongly since you are not a health
professional. So, it may not treat you well. But this method is so far the best (M01, PUWD).

Willingness to pay for an HIV self-test
Paying for HIVST kits was contested by the majority of participants. Although some were willing to pay, they felt if this approach is to be fully successful, it should be at no or a very reduced cost, in order to benefit the majority of Ugandans who are generally poor. In addition, the majority of KPs are individuals with poor financial status, yet these require frequent testing. Hence to help them, the tests require to be free of charge. There are those who have, meaning it will be the rich to use them (M08, PWUD).
When asked how much they would be willing to pay, the amount ranged between 1000-20,000, indicating variation in opinion on this issue. However, the majority of participants were at the lower side, and many expressed willingness to pay so long as the figure did not exceed 2,000. This finding suggested that in order for the saliva based HIVST approach to be successful, it should be free of charge and if at a cost, this should be very minimal.

Preferred kit distribution models
Various ways were mentioned in regard to the choice of where participants would prefer to receive the oral based HIVST kits. These included: both public and private health facilities, small and large health facilities, community centres and groups, plus commercial sites e.g. shops. Although many models were suggested, the motivation behind the choice was highly associated with the ease of access, cost implications and privacy concerns. For example, some preferred their current health facilities (facility based model), even if these are far from their residencies as they feel to be better understood by their primary service providers. Other participants preferred to receive the kits from their peers (Community Based Model), whom they feel identify with them, while others preferred the kits to be distributed at all levels of health care facilities to facilitate accessibility by all those who need them.
We The above findings suggest that in determining the models of HIVST kit distribution, accessibility by the users, privacy, and cost implications should be considered.

Discussion
Our study findings reveal that HIVST is highly acceptable among clients taking PreP irrespective of their social demographic and economic characteristics. Our acceptability study findings are comparable to other studies; for instance, acceptability of HIVST was 100% among pregnant women in India (19) and in Kenya, acceptability and feasibility of HIVST in clients taking PreP was found to be 98% (18). However unlike our study which involved a variety of MAPRS, the study by Sarkar and colleagues (19) was conducted among pregnant women. Similarly, the Kenya study focussed on only KPs on PREP who were in sero discordant relationships (18). Our study, having catered for a variety of KPs including MSM, sero discordant, FSW, PWUIDS, fisher folks, adolescent young girls and women provides grounded evidence on the acceptability of HIV self-testing among clients taking PreP.
Acceptability results in our quantitative study are further justified by the qualitative findings in which respondents singled out key reasons to justify self-testing as: The procedure being time saving, cost effective (eliminates transport costs), private, and accessible; findings that are similar to those from previous research (20).
Although the majority of respondents preferred the oral fluid-based kit for subsequent 3 monthly HIV testing, a limited number of respondents (8.5%) preferred either the conventional health worker and health facility based testing or a blood based self-testing kit if introduced in future (3%). These findings suggest that HTS providers need to customise HIV testing according to client specific In the qualitative study, respondents gave reasons for preferring the oral-based testing kit over others. These were non-invasiveness of the test procedure, being painless, ease of use, lack of blood loss, and cost effectiveness. These findings did not differ much from an acceptability study conducted in the general population in Uganda (17), and from other studies conducted elsewhere (11,21). It is feared that blood-based methods can cause blood loss and injury to the body, causing cosmetic concerns for some KP groups such as sex workers. A non-invasive (saliva-based) method is therefore likely to be appropriate for such a group.
Despite the high acceptability, concerns were raised about the HIVST model. The commonest and most important to participants was the lack of emotional support, since pre and post-test counselling are not catered for in this model. Such a revelation calls for an establishment of a system that ensures that clients access pre and post-test counselling as they use HIVST.  (24), who reported that although HIVST was highly favoured in many aspects, participants strongly opposed it being sold to them.
The current study indicated that there is no single most popular HIVST kits distribution model that would apply to all categories of respondents. However, the three most preferred distribution models are home delivered by health workers, picking from health facility, and home delivered by peers.
These findings relate with those from other studies which favored home based (25), facility based, and peer facilitated (9) models. It was also noted that special groups such as FSWs and fisher folks preferred the peer distribution approach compared to the rest of the KPs, while some groups such as fisher folks also considered finding the kits at commercial sites such as drug shops for easy access.
However, Geng and colleagues (9), in their study to find the best distribution approach for HIV selftesting kits among FSWs noted that picking kits from the facility was a better method and was associated with linkage to care compared to community distribution of kits using coupons, a finding different from those of our study. Nonetheless, our findings indicate that it is important to have multiple kit distribution models to meet varying distribution needs of different KPs.

Study limitations
Our study findings could have been affected by the following limitations:

1.
A non-probability consecutive convenience sampling method was used to enrol respondents hence not all clients on PreP had an equal chance of being selected.
This could have led to a selection bias. The choice of the sampling method was majorly influenced by the available time and resources to conduct the study, plus the dynamic nature of the study population.

2.
HIV self-testing was supervised by a trained personnel and therefore, the ability to perform a correct self-test and interpret results by respondents could not be assessed yet it is an important factor if these clients have to test at the comfort of their homes subsequently. Additionally, the presence of a supervisor could have increased respondent's acceptability of HIV testing since supervisors provided further procedural guidance during the process of self-testing.  The youngest respond in this study was 17 years old hence approval for consent was not sought from the parent/guardian (No study participant was 16 years old or younger). All respondents provided a written informed consent.

Consent for publication
Not applicable

Availability of data and materials
The datasets used and/or analysed plus data collection instruments used during the study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that there is no competing interest

Funding
This study was solely funded by the principal investigator as part of a Master of Public Health (MPH) course. The principal investigator solely met all the financial and material requirements for the study.
There was no any other external source of funding other than financial resources from the principal investigator. The principal investigator led the design and conduct of the study as well as writing of the manuscript with support from co investigators.

Authors contributions
MJBJr conceptualized and designed the study, led the data acquisition process, analysis and interpretation, and drafted the manuscript. NS participated in the qualitative data collection and analysis, and reviewed the manuscript. JTB and RWM improved the study design, data analysis and interpretation of results, and reviewed the manuscript. All the authors approved the final manuscript and agree to be accountable to it. Distribution of respondents by willingness to pay for an HIV self-testing kit as a percentage