Attitudes and Acceptability on HIV Self-testing Among Key Populations: A Literature Review

HIV self-testing (HIVST) is a potential strategy to overcome disparities in access to and uptake of HIV testing, particularly among key populations (KP). A literature review was conducted on the acceptability, values and preferences among KP. Data was analyzed by country income World Bank classification, type of specimen collection, level of support offered and other qualitative aspects. Most studies identified were from high-income countries and among men who have sex with men (MSM) who found HIVST to be acceptable. In general, MSM were interested in HIVST because of its convenient and private nature. However, they had concerns about the lack of counseling, possible user error and accuracy. Data on the values and preferences of other KP groups regarding HIVST is limited. This should be a research priority, as HIVST is likely to become more widely available, including in resource-limited settings. Electronic supplementary material The online version of this article (doi:10.1007/s10461-015-1097-8) contains supplementary material, which is available to authorized users.

and 1 in 10 do not have access to prevention services, including condoms [7]. Regional reports suggest that across 35 countries in sub-Saharan Africa only 60 % of sex workers have received an HIV test in past 12 months, although this may be an over estimation because of nonrepresentative convenience sampling in many instances [8].
In the USA, an estimated 49 % of PWID have received an HIV test in past 12 months [9]. Reaching UNAIDS' ''90 90 90'' targets, 90 % of people with HIV knowing their status, 90 % linked to anti-retroviral therapy (ART) and 90 % virally suppressed [10] will not be possible without increased efforts to improve access to and uptake of HIV testing among KP.
HIV self-testing (HIVST) is an emerging approach with the potential to be high impact, low cost and empowering for those who may not otherwise test, particularly among KP. In order to suit a local context, HIVST may be delivered in multiple ways which vary as to type of support, range of access and site of sale or distribution. Although HIVST does not provide a HIV diagnosis, and all reactive self-test results must be confirmed according to national testing algorithms [11], it may stimulate demand for and increase uptake of HIV testing and counseling among KP, who may be more reluctant to or unable to seek existing services.
Several countries have already introduced or are considering the introduction of HIVST as part of national strategic plans, testing strategies and policy and regulatory frameworks [11][12][13][14]. At this time, however no optimal approach has been identified, particularly to reach KP [11].
Potential benefits of HIVST among KP identified in the literature include: the possibility to increase access to HIV testing [15,16], reduce sexual risk behavior [17], and that it may lead to cost-savings in the context of pre-exposure prophylaxis (PrEP) implementation projects [18]. However there are concerns about linkage to further HIV testing and diagnosis, prevention, care and treatment as appropriate to a client's HIV status, particularly in legally constrained settings, social and emotional harm following HIVST, use for ''point-of-sex testing''(where individuals use HIVST to ''screen'' potential sex partners), risk of sexual disinhibition, or substitution of highly accurate facility-based HIV testing among high incidence populations [19,20]. Additionally, there are concerns about the potential for coercion to test, for example for SW being forced to test by brothel owners and clients [21,22].
While there are several systematic reviews highlighting the high acceptability of HIVST [23][24][25], none focus on KP values and preferences. In July 2014, the World Health Organization (WHO) issued the first consolidated guidelines on HIV prevention, diagnosis and treatment for the five KP groups [26]. This guidance in particular, calls for service delivery approaches that are acceptable and appealing to KP and that will also reduce disparities in coverage and access to HIV services [26]. Based on promising evidence, a changing policy environment, and renewed global emphasis to reach KP and global targets that aim to close the testing gap [10], this review focuses on the acceptability, values and preferences of KP on HIVST.

Methods
From April to July 2014 we performed a systematic search to identify evidence on acceptability, values and preferences regarding HIVST among KP (defined as MSM, SW, transgender people, PWID and people in prison). We searched five electronic databases (PubMed, PopLine, Scopus, EMBASE and PsycINFO) and five major HIV/ AIDS conference databases (British HIV/AIDS Association, Conference on Retroviruses and Opportunistic Infections, European AIDS Society Conference, International AIDS Society and US National HIV Prevention Conference) for publications between January 1995 and July 2014. Abstracts were included if full-texts were not available. Gray literature was identified through a comprehensive Google search. References were also manually searched to identify other sources. Experts and authors of pertinent studies were contacted for any further references and clarifications (Fig. 1). The search was conducted according to the PRISMA checklist (see Electronic Supplementary Material).
Search terms included ((HIV OR HIV seropositivity OR HIV infections) AND ((self test*) OR (home*test*) OR (rapid*test*))). The search was restricted to human subjects. No language or geographic limitations were placed on the search. Two reviewers screened studies. The first reviewer read study titles and abstracts meeting the inclusion criteria. The second reviewer evaluated the screening criteria and approved selected studies. Disagreements between reviewers were resolved through discussion and consensus. Studies were only included if they used original data, included at least one of the five KP groups, used qualitative and/or quantitative methods that evaluated any aspect on HIVST values and preferences. All other articles were excluded. Studies examining home specimen collection kits were excluded, because participants did not interpret their test result (Fig. 1). Literature was summarized qualitatively according to study design and methodology, location, resource and population.

Analysis
Documents were analyzed manually through describing their content. Using Microsoft Excel, a systematic framework and extraction tool was developed, to obtain particular information on HIVST values and preferences. After data was extracted it was coded by country income according to the World Bank [27], the educational level (college, high school, elementary or less), the type of specimen collection (oral fluid-based, blood-based, or not specified), KP group (MSM, SW, PWID, transgender people, or people in prison) and the type of support provided (supervised, unsupervised, or not specified).
Values and preferences were defined as participants' views on HIVST, concerns about HIVST, willingness to pay or buy a HIV self-test, a test kit either specifically packaged for HIVST or a rapid diagnostic test (RDT) distributed or used for HIVST, and other qualitative values and preferences reported by participants. In addition, we examined the acceptability of HIVST, defined as the willingness to take a test in the future or as an increased frequency of testing with a HIV home-test. Reported acceptability was then categorized as high (C67 %), moderate (66-34 %) or low (B33 %).
Approaches to HIVST were defined in accordance to the 2014 WHO and UNAIDS technical update on HIVST [28]. Supervised approaches were defined as those which involved direct support from a health worker or a volunteer before or after individuals tested him or herself. Unsupervised approaches were defined as situations when HIVST offered without requiring direct support, but could include the provision of information about where or how to access support services. Studies with no information or comparing types of approaches or specimen collection were analyzed separately. The studies reviewed included both those where participants were able to perform home tests, and those which did not include self-tests but explored survey participants' values and preferences.
We examined the process of linkage within HIVST for studies where HIVST was performed and where HIVST was not performed by participants answering a questionnaire about HIVST. We primarily analyzed linkage in any study reporting linkage from HIVST to further HIV testing, to receiving a HIV diagnosis in a facility, and/or to enrolment in HIV prevention, care or treatment services. As a secondary analysis we also examined studies which reported on participants' ''intention to link'' following a reactive HIV self-test result.

Quality Assessment
A quality critique of quantitative data from cross-sectional (Electronic Supplementary Tables S1, S2) and cohort studies (Electronic Supplementary Table S3) was performed using the STROBE checklist [29]. Reports were critiqued using the STROBE checklist as they were reporting outcomes of a cross-sectional study [30,31]. For a conference abstract reporting a randomized control trial [16] (Electronic Supplementary Table S4) we used the CONSORT guidelines [32]. Qualitative studies [17,31,[33][34][35] were evaluated with a guide for critically appraising qualitative research [36]. Due to lack of standardized reporting of primary and secondary outcomes, and heterogeneity of data on values and preferences, a metaanalysis was not conducted.

Values and Preferences for HIVST
Twenty-three studies assessed key population values and preferences on HIVST (Tables 2, 3).

Benefits of HIVST
Findings about benefits were variously documented in 18 articles, including: (a) Convenience, (b) Privacy, (c) Painless, and (d) Easiness to Use.
Privacy was more frequently reported as a benefit of HIVST in studies using an unsupervised approach The summary score for quality critique represents the number of criteria reported over the total number of criteria n/a n/a Serious adverse self testing events n/a n/a n/a n/a n/a n/a n/a n/a Linkage to care MSM:50 % would seek post-test counseling and confirmation of results* FSW: 75 % would go to a health facility/VCT for confirmation* n/a n/a n/a n/a 2 HIV reactive tests: [1] search confirmatory testing and care immediately [2] search confirmatory testing and care after 2 months n/a n/a  Median price (IQR) $ 30 (n/a) n/a n/a n/a 24 % would pay $ 50 n/a n/a (n = 5/6) [15,30,38,45,46] compared to those using a supervised approach (n = 2/6) [35,37]. Although approach was not reported 71 % of MSM in Brazil, reported that HIVST would offer more privacy than HIV testing facilities [43]. In general, the benefits for HIVST described by participants across studies, remain similar; even when analyzed by country income, type of KP, participant education level, type of specimen collection, having performed an HIVST and type of approach.

Preferences for HIVST Attributes
Twelve articles provided information on KP preferences [17, 31, 33-35, 37, 41-44, 50, 51]. Preferences for test type of sample collection (oral fluid-based or fingerstick/whole blood-based) (n = 7/12), distribution (n = 7/12), instructions (n = 2/12), the availability to link to counseling (n = 4/12), and how they would like to use the test (n = 6/ 12) were reported. Preferences for HIVST attributes varied across country income setting, type of approach, having performed a self-test for HIV and type of specimen collection. However, in general, participants reported preferring HIVST with an oral fluid-based HIV RDT (n = 4/12), to blood-based HIV RDT (n = 3/12) [33][34][35]43]. Five studies from Kenya, Singapore, USA and Australia reported MSM and FSW generally prefer HIVST to be available over-the-counter [17,31,37,42,44], three of which participants have performed an HIVST [17,42,44], and two studies from Australia and China, reported that MSM preferred HIVST to be available through the Internet, in neither of the two MSM participants have performed an HIVST [37,41]. MSM participants in Australia, desire HIVST to be available over-the-counter, but specifically with proper instructions for use on how to perform a HIV RDT and interpret the test result [37].
Three studies reported participants prefer having counseling available [37,42,44]. However, one study in Hong Kong SAR China among MSM reported that 16.2 % of participants prefer HIVST without counseling [51].

Linkage to Care
Six studies reported on some aspect of linkage to care from HIVST, of which the majority were in HIC settings [16,30,31,33,50,51]. Two studies, Katz et al. [16] and Mayer et al. [50] reported actual linkage and enrolment in care following HIVST. Katz et al. [16] reported two participants with reactive self-test results who were diagnosed HIV positive: one participant searched immediately for additional HIV testing and care and the other waited two months before seeking further HIV testing and care [16].
The remainder of the studies reported on ''intention to link'' following HIVST. In studies from HIC settings, the majority of participants reported that if they received a reactive HIV self-test result they would seek for additional

0%
20% 40% 60% 80% 100% Chakravarty et al [49] Marley et al [35] Ochako et al [31] Gray et al [34] Xun et al [47] Chen et al [38] Bavinton et al [15] Katz et al [16] Carballo-Diéguez et al [33] De la Fuente et al [39] Wong et al [51] Lippman et al [43] By Gray et al [34] Chen et al [38] Bavinton et al [15] Carballo-Diéguez et al [33] Marley et al [35] Katz et al [16] De la Fuente et al [39] UNSUPERVISED SUPERVISED Unsupervised Supervised * Two studies were not included in the above chart as they did not include sample information: Greacen et al (40) and Bavinton et al (48).  [30,33,51]. A study in LIC setting reported that 50 % of MSM would seek post-test counseling and confirmation of results and 75 % of FSW stated that they would go to a health facility for confirmation, after self-testing for HIV [31]. Overall, no differences were found when analyzed by test type of specimen collection, educational level, having performed an HIVST and type of approach.

Adverse Events Resulting from HIVST
There was little information on adverse events reported in reviewed studies. In this review, one study among MSM in the USA, who had performed an oral fluid-based HIV RDT, reported that complicated situations could lead to verbal confrontations or violence among participants who self-tested or proposed self-testing with a sex partner. Also they reported that special circumstances, such as infidelity, could lead to coercively test a partner, a potentially more adverse event [33]. No other serious adverse events were identified.

Quality of Studies
Quality of studies varied. In general, studies did not report sufficient information about qualitative methods and data collection tool, there was also a lack of compliance on how they assessed and measured the different values and preferences. Qualitative data were sparse and an incomplete reporting of data in abstracts and reports limited the evaluation of quality. This lack of clear evaluation of values and preferences limited our understanding of collected data.

Discussion
Twenty-three studies reporting acceptability and other values and preferences of KP regarding HIVST were identified. Values and preferences were largely consistent. This may be because many of the included studies had some similar study characteristics. For instance, the majority of included studies were from HIC settings (n = 18/23), among participants with high educational level (n = 11/23), using oral fluid-based RDT (n = 13/23), using unsupervised approaches (n = 9/23), and were almost exclusively among MSM (n = 23/23). Very few studies in this review included FSW, PWID, transgender people (n = 5/23). Evidence for high acceptability was evident among MSM in HIC settings using oral specimen collection. This aligns with existing literature on HIVST, which suggest Marley et al [35] Xun et al [47] De la Fuente et al [39] Katz et al [16] Carballo-Diéguez et al [33] Did Not Perform HIVST Performed HIVST Wong et al [51] Gray et al [34] Chen et al [38] Bavinton et al [15] Bavinton et al [48] Greacen et al [40] Lippman et al [43] Ochako et al [31] Men who have sex with Men Female Sex Workers LOW MODERATE HIGH Acceptability Fig. 3 HIV self-testing experience among studies evaluating acceptability AIDS Behav (2015) 19:1949-1965 1961 users (including the general population) may prefer oral fluid-based HIV RDT to fingerstick/whole blood-based HIV RDT because they are reportedly easier to perform and are perceived to be less painful [52,53]. Out of all studies reviewed, Chakravarty et al. reported the lowest acceptability of HIVST. However this study only reported acceptability among HIV-negative MSM who were aware of HIVST and reported that they were ''extremely likely'' to self-test for HIV. Since the study did not report on other levels of acceptability, such as ''somewhat likely'', ''likely'' or ''very likely'', we could not infer whether this is reflective of actual acceptability of HIVST among MSM [49].
Research is still ongoing and there are emerging reports from KwaZuluNatal, South Africa which suggest that fingerstick/whole blood-based HIV RDT can also be easy to perform and accurate, when accompanied with clear instructions, packaging and appropriate test system design [54]. In April 2015, two fingerstick/whole blood-based RDTs recently satisfied the legislative requirements in the European Econonomic Area: the BioSure HIV Self Test (BioSure Ltd, UK), sold online at £29.95 [58] and the autotestVIH (Aaz Labs, France) will be sold in pharmacies around 23-28 euros [59]; as an additional option for people to now their HIV serostatus. Various other products are under development and could be adapted for HIVST, including painless or integrated lancets, simplified sampling systems, integrated buffer delivery systems and shorter minimum and maximum reading time [11].
Some studies report that participants desire access to counseling [37,42,44], while a study in Hong Kong SAR China with MSM, reported that 16 % preferred HIVST because of the ''lack of counseling'' [51]. Ways to provide information about or how to link to counseling services, as part of HIVST, should therefore be considered including: face-to-face through community health workers, internetbased, SMS or mobile phones, or computer-based programs.
Studies with unsupervised or an unknown approach to HIVST frequently reported concerns on user error and poor accuracy. These concerns could potentially be overcome by providing links to support and counseling services and clear instructions for use. There might be a small controversy with the benefit of privacy and the concern of an increased user error, depending on the approach, in our findings MSM were not strongly positioned that HIVST has to be performed strictly by a professional [37,42,44,51]. In particular, KP may need more information on how user error can be reduced, accuracy rates and the need for confirmation; especially if HIVST is unsupervised.
Willingness to pay was difficult to compare across all studies, as there were different price points and some used overlapping intervals. Overall willingness to pay was higher in HIC settings [16,37,39,42,[44][45][46] compared to MIC settings [35,41,47] or LIC settings [31], and lower in supervised HIVST [16,37,42,44] than for unsupervised HIVST [41,47]. This may be because supervised HIVST is viewed as similar to current HIV testing services, which are often free of charge. KP may also be willing to pay more for unsupervised HIVST because it offers greater privacy; which was a key benefit and value of HIVST, reported by KP.
All studies in the USA (reporting willingness to pay between US$1 to CUS$50) were conducted using oral fluid-based HIV RDT [16,45,46], and prior to the US Food and Drug Administration approval of the OraQuick Ò In-Home HIV Test [55]. Currently, this product retails direct to consumers for US$40 [56]. The studies reviewed suggest that reluctance to pay was only reported in studies were participants have performed an HIVST, also concerns about the cost of HIVST, were both in MIC and HIC settings. Thus, for HIVST to have higher uptake, it will likely need to be subsidized or free of charge to clients. So far a lowest price has been negotiated, for research purposes the professional use version of this test is available in Kenya for approximately US$11 [18] and in Malawi for US$3 [57].
Evidence on linkage to care and treatment among KP is limited and requires further research. Two studies among MSM in the USA reported actual linkage to HIV testing and diagnosis and enrollment in HIV care and treatment [16,50]. Three studies reported that more than 80 % of participants with a potential or an actual HIV positive test result would seek confirmatory HIV testing and care [16,30,51]. Proactive approaches to support the unique needs of KP may be considered and adapted, for example a study in Malawi among general population offering home (ART) assessment found a three-fold increase in linkage to ART, compared to facility-based HIV testing [60]. It is essential that users with a reactive HIV self-test result first link to further testing and receive an HIV diagnosis; and that users also link to HIV prevention, care and treatment services, as appropriate to their HIV status, in a timely manner. Special attention should be paid to additional risks for KP, including young and adolescent KP. In highly criminalized settings KP may be more vulnerable to delay or not to seek HIV services. Without such support for safe linkage to HIV services, HIVST may be of limited benefit to KP in such settings.
We found no clear evidence to support adverse events as a result of HIVST, such as adverse emotional reactions to positive tests, inter-partner violence, coerced/forced testing, psycho-social or mental health issues, and suicide or self-harm. This is in line with a recent literature review which states that very few studies report harm across various self-tests, including HIV; however it does note that monitoring and reporting systems for harmful outcomes are rare [61].

Limitations
The majority of studies that met inclusion criteria were among MSM and in HIC settings. Only two studies provided data on user preferences among MSM and FSW [31,35]. Our search was for KP, however due to the nature of self-testing, people in prison or closed settings, would not be eligible for HIVST. Almost all studies were observational and used a cross-sectional research design. Only one study in this review was a randomized control trial. We cannot therefore rule out selection bias, including sample representativeness and non-response rate. The inclusion criteria for this review were overly inclusive to capture all or any values and preferences on HIVST among KP. Therefore, study designs, characteristics and sample sizes were heterogeneous, and results may not be generalizable.
Most studies had incomplete reporting of data items and low compliance with the STROBE reporting checklist.

Conclusion
MSM in HIC settings find oral fluid-based HIVST to be highly acceptable, using a supervised or an unsupervised approach. However, concerns about counseling, user error and poor accuracy remain. Data on social harm and adverse events resulting from HIVST was not reported. To better understand user concerns, as well as risk of adverse event and potential social harm, rigorous monitoring and reporting systems should be implemented, so that program managers and policy-makers can consider the potential risks and benefits for introducing HIVST among KP.
The convenience and private nature of HIVST is reportedly advantageous to MSM, and may also be so for other KP, including SW, PWID and transgender people in HIC, MIC or LIC settings. However, information among KP, other than MSM, and in low-or middle-income settings is limited.
Key population values and preferences around HIVST should be considered by researchers, policy-makers and program managers, as HIVST may be an additional approach to reverse inequities in access to HIV testing for KP who have a low access to HIV services, but carry much of the HIV burden globally. Taking into account our study results, more data from diverse settings and among non-MSM key population groups is needed to better understand the potential impact of self-testing as part of the global HIV response.