Introduction

Screening plays a major role in current HIV prevention strategies [1–5]. However, in France, a large number of people at risk of acquiring HIV are not being tested, or are being tested too late [6–8], with serious consequences for the people involved [7] and a significantly increased risk of onward transmission to others [1]. Indeed, despite the fact that over five million tests are performed in France every year, it is estimated that some 30,000 people, i.e. 19% of people infected with HIV, are unaware of their HIV status [9]. Amongst these, one third would be men who have sex with men (MSM), one third heterosexual migrants from sub-Saharan Africa and the remaining third, heterosexuals born in France. Thus, facilitating access to repeated screening opportunities for the more vulnerable groups with regard to HIV remains a major public health priority. Given this epidemiological context, the French Ministry of Health, based on reports by the National AIDS Council [10] and the National Ethics Committee [11], recently ruled in favor of making HIV self-tests available to the general public. The first such test went on sale on September 15, 2015.

Some highly at risk populations are clearly interested in accessing self-tests. In an online survey performed in French on HIV self-tests, 87% of the 5,908 non HIV positive MSM previously unaware of the existence of these tests expressed interest in accessing them if they were authorized [12,13]. These men were more likely to be younger, to live in smaller towns and to live with their parents or their wives and children. They also were more likely to engage in unsafe sex with casual male partners, to live their sex lives with men in absolute secrecy and yet not to have been tested for HIV recently. When asked why they were interested in accessing self-tests, respondents gave the same three main reasons already identified in the USA for this same population group: convenience, rapidity and anonymity [14]. Men not interested in accessing self-tests gave the following reasons for their lack of interest: satisfaction with current screening methods, doubts about reliability, not wanting to be alone when discovering results and fear of incorrect use [12].

Interest for self-tests is not just limited to at-risk population groups such as MSM. The general population is also interested [15–17]. In France, around 70% of respondents in a 2010 survey declared that self-tests would facilitate testing, with the younger respondents being more likely to prefer self-testing [18]. A review of the international literature by Krause et al. (2013) concluded that self-testing is highly acceptable among all populations studied, whether in resource-limited contexts (Kenya, Malawi) or in higher income countries (USA, Spain, Singapore) [19]. Participants underlined the importance of anonymity, but at the same time expressed worries about accessing counseling and care [19]. A systematic review on evidence from supervised and unsupervised strategies found, for both strategies, high levels of acceptability (74%-96%), preference for this type of test (61%-91%), and interest in doing the test with sexual partners (80%-97%) [20]. However, for coherent public health policy on making HIV self-tests publicly available, the authors of the latter study highlighted the lack of research exploring people’s motivations to use the self-test, their fears, needs, perceptions, preferences and priorities. Finally, a third more recent literature review on attitudes and acceptability of HIV self-testing for key populations concluded that most studies addressing this question have taken place in high-income countries with men who have sex with men], underlining the need for research concerning other key populations [21].

The present study aimed to address these issues using a qualitative survey with HIV screening experts working with vulnerable or less vulnerable populations with regard to HIV, and with the principle objective of identifying the information and support needs of the different groups of potential self-test users in France.

Methods

From February to May 2014, experts working in eight parallel groups participated in a three-round Delphi process conducted on the Internet. The Delphi process collects and synthesizes expert opinion on a given issue in the area of their expertise. Participants are chosen for their expertise, and are ensured anonymity with respect to the recommendations they might make. Each expert then gives a score to all the recommendations made by the members of their group. In light of the replies of the other participants, they are then asked to revise their initial answers with the aim of bringing the group towards a consensus, or presenting clear sets of opposing arguments should points of view be discordant [22]. Three rounds are generally viewed as being sufficient for achieving a high level of agreement [23].

The experts were chosen by the study’s scientific committee for their experience and expertise in the area of HIV and HIV screening for each population. Each group developed recommendations for a specific population, including six high HIV prevalence populations: men who have sex with men (MSM), migrants from sub-Saharan Africa, substance users (including both injection and non-injection substance use), transgender people, French West Indies, French Guiana; and two low prevalence populations: the general population, people aged under 25. The choice of including a specific group on young people, in spite of their current low prevalence rates, was linked to the results of a 2010 survey [18] which revealed that today’s younger generation in France seems to be less aware of HIV risks than their elders. The procedure for choosing experts was as follows. Firstly, an extensive list of potential participants was compiled for each population group. All members of the scientific committee then scored the experts according to their perceived expertise with regard to HIV testing for the population group in question. For each population group, the representatives with the highest mean scores were invited to participate, but at the same time including at least one expert from each of the four key expert categories: researchers publishing in the area of HIV and screening; clinicians working in HIV medical services or in screening centers; policy-makers in regional or national health authorities; board members or employees of community organizations addressing the sexual health concerns of the population group in question. No expert participated in more than one group. The aim was to include a total of ten to twelve experts for each group, thus constituting a pooling of judgments, as much as possible from each of the four expert categories, as well as taking into account the information processing capability of the research team, as recommended by Hsu and Sandford [24]. The entire Delphi process was conducted by email, thus allowing a broad geographical representation: apart from the two expert groups specifically addressing the two overseas French departments with high HIV prevalence, each group included experts from the Greater Paris Area (Ile-de-France) as well as from the other provinces.

In the first round, experts were asked to propose ten or so statements that, in their opinion, constituted “good practices for responding to the information and support needs of HIV self-test users”. For each factor, experts were asked to explain briefly why they considered this to be important. The final list of factors for each expert group was then analyzed by two researchers using the following method: (a) factors that were identical or that used different words to describe the same phenomenon were grouped into one factor; (b) factors that covered more than one issue were divided into distinct entities. The researchers took pains to respect each expert’s nuances in describing his or her different recommendations using, as much as possible, the terms and expressions chosen by the experts themselves to formulate final versions of each recommendation. After the within-group analysis for each of the eight groups of experts, the same process was used to identify factors that were common to more than one expert group and to harmonize the terminology used across different groups.

In the second round, the complete list of factors identified by all the experts within each group were sent back to all the members of that group. Experts were then asked to score each factor on a scale from 1 (not at all relevant) to 9 (highly relevant) with regard to the degree of importance they attributed to that factor for informing and supporting self-test users when the self-test would be officially approved for sale and would come onto the market in France. The mean score for each factor was then calculated as the group score for that factor for the group in question.

Finally, in the third round, experts were invited to reconsider their scores, if they wished, in the light of the mean group score for each factor.

For the purposes of the present study, recommendations with scores ≥7 out of a possible maximum of 9 were considered to be relevant. Recommendations are presented by decreasing mean score. Standard deviations are used to describe the level of agreement between experts of the same group.

The Paris Descartes University Health Research Ethics Board (CERES) discussed this project and conveyed that no ethical committee review was required for this study, the methodology used presenting no problems of any ethical nature. Participants were deemed to have consented by returning their recommendations.

Results

Ten to twelve experts were identified and contacted for each target population. Although the participation of the experts from mainland France proceeded as planned, the response and acceptance rates of experts from the French West Indies and Guiana proved to be considerably lower, despite the repeated entreaties of the research team. Overall, from February to May 2014, a total of 72 experts completed the Delphi process, including only 6 for the French West Indies and 5 for French Guiana. The general characteristics of the experts for each expert group are presented in the Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the experts in each population group.

https://doi.org/10.1371/journal.pone.0152567.t001

Using the qualitative methods as described above, a final total of 263 recommendations were identified and aggregated into eight themes (Table 2). The two major themes concerning communication and information about the self-test and how to use it accounted for 47% of the total number of recommendations.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Number of recommendations for each major theme.

https://doi.org/10.1371/journal.pone.0152567.t002

Table 3 shows the total number of recommendations and the number of recommendations with a mean score ≥7 for each group of experts. The MSM expert group stood out from other groups with regard to their high total number (90) of recommendations, compared to 60 and 57 respectively for the same number of experts for Substance Users and the General Population, or to 48 for the nine experts in the Transgender group. This was a first indication of the considerable interest of the self-test for MSM experts compared to the experts for other population groups. With regard to high-scoring recommendations (mean score ≥7), the General Population expert group stood out with only 35% of recommendations with mean scores ≥7, compared to the migrants from sub-Saharan Africa or Transgender groups with 79% and 88%, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Number of recommendations for each expert group.

https://doi.org/10.1371/journal.pone.0152567.t003

Discussion

The present study used the Delphi process with 72 experts on HIV screening and populations vulnerable to HIV to develop recommendations for good practice with regard to HIV self-testing in France. The main strength of the study was its scope, allowing a comparison of points of view of expert groups for eight different populations, including populations with both high and low HIV prevalence rates in France, and addressing a broad range of issues associated with providing access to self-testing.

Although a large number of recommendations were common to several groups of experts, the study highlighted a series of recommendations specific to each different population group, particularly with regard to information content and access both to information and to the self-tests themselves. The present results have thus already proved useful to health authorities at a national level in making recommendations and setting up standards for marketing self-tests in France. These issues ranged from accessing self-tests and accessing information and support on how to use self-tests in highly diverse community contexts, through to the key question of access to care for individuals facing positive self-test results]. In the present study for example, the recommendation that the instructions must indicate what to do if you get a positive test result was made by seven of the eight expert groups and received the highest overall mean score of recommendations common to three or more groups.

The sheer number of different recommendations made by the MSM experts compared to other expert groups in the present study is to be noted, potentially underlining the importance of HIV self-testing for this population group. For other groups, the same number of experts produced significantly lower numbers of different recommendations. This was particularly the case for the transgender experts who stressed the fact that HIV testing was often not a priority issue for this population group, being HIV positive being felt to be a barrier to accessing hormonal therapy or surgery. These experts insisted on the fact that, for transgender people to want to take care of themselves and look after their health, all barriers to the transitioning process needed to be removed.

The cost of self-tests is a significant issue, as has already been highlighted in several studies internationally. In 2012, in Singapore, with a screening center population amongst whom 40% earned less than US$1,500 per month, only 28% of subjects declared that they would accept to pay more than US$15 for a self-test [25]. This finding is in contrast with results published the same year in a Spanish study, where 40% of participants declared that they would be willing to pay €20 (US$23) or more [26]. In the present study, the recommendation that the self-test should be at a moderate and accessible price for all was brought up and validated by seven of the eight expert groups. However, less consensual was the proposition, debated in six groups, that self-tests should be available free-of-charge in screening centers, family planning centers, community organizations, and services for people with substance abuse problems. This recommendation scored highly for three groups who had also recommended self-tests being free-of-charge for key populations with higher HIV risk (sub-Saharan migrants, SU and FWI). However it was only given a mean score of 5.0 by the MSM expert group and 6.9 by the General Population and Young People experts. Clearly, the question of the price of self-tests is an issue with varying significance from one population group to the next. With regard to costs involved in accessing counselling and support, opinions were more consensual: four expert groups (sub-Saharan migrants, Trans, FWI and Young People) underlined the importance of there being zero cost for the user to access telephone support services, including from smartphones.

Another major issue was that, although self-testing is clearly a significant step towards user empowerment and sense of self-efficacy with regard to screening [27], experts in the present study insisted upon the need to provide access to information and support at each and every step of the self-testing process, underlining the importance of support and training for effective empowerment. In France, in the context of the recent government approval of HIV self-testing, the principle government-funded hotline providing information and support on HIV/AIDS free-of-charge and 24/7 will also ensure information and support for self-test users. However, linkage to counseling and care should the test result be positive is important, in concordance with recent findings from a New York study on MSM’s anticipated and actual reactions to obtaining positive self-test results [28]: subjects underlined not only the need to be able to access information, counselling and care, but also the question of handling their immediate emotional reaction when discovering a positive result and then dealing with potential stigma associated with being HIV positive.

The present study has a certain number of methodological limitations. First, the fact that only a limited number of experts accepted to participate in the French West Indies and Guiana expert groups clearly reduced the validity of results concerning these two populations. Seemingly, HIV was not amongst current priorities, no doubt in part due to the arrival of the chikungunya epidemic in the Caribbean area at the very moment the Delphi study was being launched. However, it is to be noted that (1) the four areas of expertise (research, screening and care, policy-making, community groups) were represented in both these expert groups and (2) a high proportion of recommendations were in common with the other expert groups. Second, the fact that all expert groups systematically included experts in HIV screening may have introduced a certain bias with regard to maintaining or defending existing screening services. Studies involving the different population groups in question with access to self-tests in real-life situations need to be undertaken to validate the findings derived from the experts participating in the present study. Third, the large number of recommendations and the fact that, for example, only one in three were common to three or more of the eight groups of experts, might well be a methodological artifact: indeed, researchers in the present study made a considerable effort to respect the different nuances expressed by each expert group. For example, the recommendation Create a telephone hotline providing support on how to use the test, accessible 24/7, common to all eight experts groups, was identified as a separate recommendation from that made by experts for young people and for substance users concerning providing a hotline specifically addressing support for users who discover positive self-test results: Create a telephone hotline providing support in case of positive result, accessible 24/7. Merging these separate but similar recommendations into one single recommendation might well have reduced the total number of recommendations and increased between-group concordance rates. In other cases, it is clear that on certain issues, in-depth interviews with experts might have provided more accurate and detailed recommendations. However, doing so many interviews would have been impossible given the importance of producing these recommendations before the HIV self-tests came onto the market in France. Furthermore, it must be underlined that, in the first round of the Delphi process, the experts were asked to give a short explanation with each of their initial recommendations. The researchers who analyzed the round one recommendations took all these comments into consideration when formulating the recommendations for round two of the Delphi process. Fourthly, in spite of the fact that the use of HIV self-tests in populations with low HIV prevalence like young people may lead to an increasing number of false positive test results, only two expert groups made recommendations concerning the reliability and performance of self-tests. This may well be due to the simple fact that this issue may have been seen by many experts as not being related to the principle research question concerning “good practices for responding to the information and support needs of HIV self-test users”. This should clearly not be interpreted as meaning that the other expert groups considered issues of specificity and sensitivity as not being important.

Finally, extrapolations from this French study to other national contexts are clearly problematic. The fact that the French health and social care system includes free access to HIV screening, treatment and care with no out-of-pocket payment will have inevitably influenced experts’ priorities in the present study.

Conclusions

Results from the current study have already made a significant contribution to policy decisions concerning catering for the specific access, information and support needs of different potential HIV self-test user groups when these tests became publicly available in France in September 2015. Providing adapted access, information and support will contribute to facilitating screening both for people from high risk groups and for the general population, as well as potentially making an inroad into the hidden epidemic in France by bringing in vulnerable populations that had been reluctant to use standard testing options so far.

Supporting Information

Acknowledgments

The authors would like to express their sincerest thanks to the 72 experts who participated in this Delphi study, to the Maison Blanche Hospital Research team, in particular Dominique Granger and Emmanuelle Jouet, for contacting the experts in question and organizing the three rounds of the Delphi process, and to Christopher Daniell for his precious help with translation issues.

Initial results of the study were presented at the IAS conference in July 2014 in Melbourne, Australia and the HepHIV (HIV in Europe) conference in October 2014 in Barcelona, Spain.