ANRS-COM'TEST: description of a community-based HIV testing intervention in non-medical settings for men who have sex with men.

OBJECTIVE: To describe a community-based HIV testing programme. DESIGN AND
SETTING: An intervention of HIV voluntary testing conducted in non-medical settings in four French cities. PARTICIPANTS: Men who have sex with men (MSM). INTERVENTION: Counselling and rapid HIV testing staffed by trained personnel from an HIV/AIDS community-based organisation. PRIMARY AND SECONDARY OUTCOME MEASURES: The population that has taken hold of the intervention and the satisfaction of participants. Data were collected on demographics, HIV testing history, sexual practices and satisfaction with the testing programme.
RESULTS: 532 MSM were tested between February 2009 and June 2010, of whom 49 (9%) were tested two or more times. 468 MSM (88%) had casual male partners in the previous 6 months, and 152 (35%) reported having unprotected anal intercourse with risky casual partners (HIV infected or HIV serostatus unknown). 159 men (30%) had not been tested in the previous 2 years, and 50 (31%) of whom had unprotected anal intercourse with risky casual partners. Among the 15 patients who tested positive (2.8%), 12 (80%) received confirmation and were linked to care (median CD4 cell count =550/mm(3)). Satisfaction was high: 92% reported being 'very satisfied' with their experience. Steps of counselling and testing procedure were respected by testers and difficulties in handling tests were rare.
CONCLUSIONS: This community-based HIV testing programme reached high-risk MSM, of whom a substantial proportion had not been tested lately. This novel service supplements pre-existing HIV testing services and increases access to HIV testing in high-risk groups.

Introduction

Until very recently in France, only physicians could prescribe, perform and provide the results of HIV tests. Although current HIV testing rates in France rank second in Europe, 1 roughly 50 000 of an estimated 135 000–170 000 people infected with HIV remain unaware of their infection.2 Among people the most concerned by HIV, men who have sex with men (MSM) account for half of new HIV infections approximately.3 4 The HIV incidence in MSM is 60-fold higher than in the overall population.3 Moreover, a recent study demonstrated that 32% of MSM were diagnosed at advanced stages of disease.5

The most significant barrier to early HIV testing is the absence of perception of risk for HIV.6–8 In an effort to overcome this barrier, the French ministry of health recently recommended that physicians perform one-time routine voluntary HIV tests in the general population and annually in population groups at high risk of infection such as MSM.9 10 However, barriers to HIV testing remain at the individual level: fear of the disease, its disclosure and subsequent social stigma, as well as poor access to HIV testing.6–8 In addition, the gay community highlight inappropriate counselling and some moralistic attitudes face to their sexual practices and testing habits in conventional testing services as barriers to regular HIV testing.11

The recent availability and acceptability of rapid HIV tests12 13 offer an opportunity to implement new HIV testing strategies. A community-based HIV screening programme, for instance, may increase access in some populations by offering a more attractive and convenient location than doctor's offices. The presence of peers and non-clinical staff members who address sexuality more openly and avoid medical jargon during counselling sessions could also offset cultural barriers and reduce fears of HIV and associated stigma. Furthermore, some community-based organisations (CBOs) have been engaged in outreach prevention in which sexual practices, HIV exposures and testing are addressed. Continuing the prevention counselling with the awareness of the HIV serostatus could include testing in a comprehensive strategy of HIV exposure reduction.

In recent years, several European countries14–18 have begun implementing community-based HIV testing with rapid tests. They propose rapid testing in CBOs in large urban areas. The principle behind this strategy is the same than the one applied in developing countries where testing is conducted by lay counsellors from the community to facilitate access to testing to vulnerable populations.19 However, most of the reported programmes in developed countries involve medical staff, and although welcoming and support are conducted by community peers, HIV tests are performed by healthcare workers. To our knowledge, in Europe, the only ongoing community-based HIV testing programmes that do not involve medical staff is Checkpoint in Barcelona, Spain,20 and LASS in Leicester, England (http://www.lass.org.uk). However, data on these programmes (evaluation of an existing programme or set up into a study) have not yet been published.

The hypothesis was that a community-based HIV testing intervention may reach high-risk MSM, a high proportion of whom have not been tested lately; consequently, in addition to other existing HIV testing services, it may increase access to HIV testing in high-risk groups. The ANRS–COM'TEST Study describes a community-based HIV testing and counselling programme performed by peers in non-medical settings that targeted MSM. The purposes were to describe characteristics of the MSM who has taken hold of this programme and the satisfaction of participating MSM.

Methods

Ethics statement

ANRS–COM'TEST Study was approved by the French Comité de protection des personnes Nord-Ouest III and the Agence française de sécurité sanitaire des produits de santé. All participants were informed of the study and had to sign a consent form to be included. The survey was anonymous.

Intervention

The community-based HIV testing was managed by AIDES, a French CBO that focuses on outreach and prevention among exposed population and notably MSM. Although not being healthcare workers, trained AIDES CBO staff members performed the whole testing procedure including pre- and post-test counselling, rapid HIV tests as well as delivery of test result (figure 1). Prior to HIV testing, testers spoke with participants about their sexuality, risk perceptions and sexual safety. Once the results were available, participants learnt their HIV status, as well as strategies to reduce sexual risk taking.

Figure 1

ANRS–COM'TEST Study diagram.

AIDES CBO staff members were trained specifically in risk assessments, risk-reduction strategies and counselling using published motivational interview methods.21 They were also trained in performing the rapid HIV test, reading the test and delivering the result and referral for confirmatory test or other services if needed.

The VIKIA® HIV1/2 BioMérieux rapid HIV test kit (sensitivity: 99.8%; specificity: 99.9%22) was used to analyse a self-drawn whole blood sample from the participant's fingertip. Results were available within 30 min. Participants who tested positive were referred to HIV clinics for confirmatory blood tests and linkage to care.

The intervention offered HIV tests during 3 h sessions once or twice a week in the evening and/or on the weekend in the AIDES CBO locations in the French cities of Montpellier, Lille, Bordeaux and Paris. No appointment was required, and HIV tests were performed on a first-come first-served basis. We informed the MSM community about the intervention through communication campaigns (posters, flyers, web banners and ads) at commercial and non-commercial gay venues, as well as in gay websites, magazines and organisations. The study sites were the settings of the AIDES CBO. The possibility of performing an HIV test was, however, not advertised outside the setting to preserve confidentiality.

Study population

ANRS–COM'TEST exclusively targeted MSM. Eligibility for the study included age ≥18 years and pursuit of HIV testing at one of the four participating AIDES CBO locations. MSM who reported potential exposure ≤48 h prior to enrolment were not included. Rather, they were immediately referred to medical settings for HIV testing and post-exposure prophylaxis.

One of the most important goals of the intervention evaluated in this study was to target MSM who are not regularly tested for HIV (or never tested). In the French 2004 Gay Press survey—a survey investigating lifestyle and sexual behaviours in MSM who read the gay press—17% of MSM stated that they have never been tested for HIV in their life. Among those with at least a history of one HIV test, 27% stated that they were not tested in the previous 2 years. Based on these results, we therefore anticipated that 30% of MSM enrolled in our study would not have a history of HIV testing in the previous 2 years. We calculated the number of patients to be enrolled in this study to have a precision of 4% around this estimated point. The calculated sample size was 504 MSM; given the highest number of participants that could be tested by session in each centre, the enrolment time was estimated at approximately 1 year.

Study outcomes

The population that has taken hold of the intervention was described, in particular its demographic characteristics, HIV testing habits and sexual practices.

The satisfaction of the participants with the programme was collected just after the testing. The difficulties occurred in handling rapid tests, the respect of counselling and testing procedures and the proportion of participants tested positive who were linked to care were also assessed.

Data collection

Participants completed one questionnaire before and one after the test. The pre-test questionnaire assessed demographic characteristics, previous HIV testing history, sexually transmitted infections (STIs) diagnosis in the previous 6 months and sexual practices in the previous 6 months. Using a 4-point Likert scale, the post-test questionnaire assessed satisfaction with the testing experience, that is, satisfaction globally and with each step of counselling and testing procedure. It also assessed satisfaction during the interviews with counsellors, satisfaction with words used, items addressed and information learnt. Finally, it addressed stress and comfort of an HIV testing performed by peers in non-medical settings.

Staff members used a form that outlined every step of the intervention: welcome and description of the study, participant's signature of the consent form, major concerns to address during pre- and post-test interviews and difficulties faced at each step of HIV testing procedure. For every participant, staff members marked the completion of each step (done or addressed: yes/no/partially) and mentioned any comments. This form was in particular used to identify problems faced during testing and counselling.

AIDES CBO staff members conducted face-to-face interviews with patients who were tested positive 3 months later to evaluate their linkage to care.

Data analysis

The main analysis was descriptive: medians and IQR were used to describe quantitative data, and numbers and proportions were used to describe qualitative data.

Some comparisons were performed between participants tested negative and positive, and participants retested during the study period at study sites and those were not retested. For this purpose, the Fisher exact test was used to compare proportions.

Statistical analyses were performed with SAS V.9.2 software (SAS Institute Inc.).

Results

Overall, 598 men sought voluntary HIV testing at AIDES CBO involving locations. We excluded 66 participants. The three main reasons to not be included were (1) reported sex exclusively with women (n=25), (2) women (n=14) and (3) refused to participate due to the amount of time they should have spent for testing and research procedures (around 2 h, n=10). Among the 17 remaining men not enrolled in the study, 10 refused for different reasons (afraid of lack of confidentiality, no need to be tested for HIV, need time to think a possible participation) and the seven other were excluded because of age <18 years, a risk exposure <48 h or they did not understand French speaking. The man who reported a potential HIV exposure <48 h was referred to a post-exposure prophylaxis service.

We enrolled 532 MSM in the study and performed 592 tests; 49 participants (9%) were tested two or more times throughout the study period. More than half of the tests were performed in Paris (285/592, 54%). Socio-demographic characteristics of participating men are shown in table 1. Although 94% of men defined themselves as MSM, 128 (25%) stated that their sexual identity was unknown to their family and 64 (13%) that they had not revealed it to anyone.

Table 1

Participant characteristics

Participant characteristics	Median (IQR)
Age	31 (25–38)

The study enrolled 532 men who have sex with men (MSM). Percentages are calculated based on the number of respondents to each question. For each question, there were less than five missing data points (<1%); for questions regarding acceptance of homosexuality and bisexuality, there were between 22 (4%) and 35 (7%) missing data points according to the question.

Parents and/or siblings.

Divorced or separated from partner.

Of the 527 MSM who answered questions regarding their HIV testing habits, 368 (70%) reported having been tested in the previous 2 years. Among these men, the median number of tests in the previous 2 years was two (IQR, 1–4) and the last test was performed for a median of 8 months (IQR, 4–14) prior to enrolment (table 2). Among the 159 MSM (30%) who had not been tested in the previous 2 years, the last test was conducted for a median of 46 months (IQR, 34–62) before enrolment.

Table 2

Risk profile of the 532 ANRS–COM'TEST participants

HIV test in the previous 2 years?	No test	≥1 test
	n=159	n=368
History of HIV testing in the previous 2 years*
Months since previous test, median (IQR)	46 (34–62)	8 (4–14)
Number of tests in the previous 2 years, median (IQR)	0	2 (1–4)
Casual male partners and sexual behaviour 6 months prior to testing, n (%)	132 (83)	333 (90)
Number of casual partners, median (IQR)	11 (5–20)	14 (6–30)
Partners were HIV infected or serostatus was unknown, n (%)	123 (77)	307 (83)
Number of casual partners who were HIV infected or whose serostatus was unknown, median (IQR)	8 (3–13)	10 (4–20)
Unprotected anal intercourse with partners who were HIV infected or whose serostatus was unknown†, n (%)	50 (31)	100 (27)
History of STIs in the previous 2 years, n (%)‡	8 (5)	44 (12)

Except for history of STIs in the previous 2 years, the differences between men who have sex with men tested or not tested for HIV in the previous 2 years were not statistically significant (p>0.05).

Five missing data points (<1%).

Overall, 152 men reported having unprotected anal intercourse with partners with HIV serostatus unknown or positive, but for two of them, the information about the last test was missing.

One missing data point (<0.2%).

STI, sexually transmitted infection.

Nearly all participants (96%) reported having had at least one male partner in the previous 6 months; 92% said that these were casual partners. The median number of casual partners within the last 6 months was 12 (IQR, 6–25). Overall, 152 (35%) men reported having unprotected anal intercourse with casual male partners who were HIV infected or whose serostatus was unknown. Among MSM who were tested in the previous 2 years, 100 (27%) had unprotected anal intercourse with casual partners who were HIV infected or whose serostatus was unknown (table 2). This proportion was 31% among participants who had not been tested in the last 2 years. During the prior 6 months, 415 men (78%) stated using at least once recreational drugs before or during sex and 205 men (39%) reported regular use of them. The recreational drugs the most used here were alcohol (336; 65%), poppers (236; 46%) and cannabis (140; 27%).

The most frequent reasons for participating in ANRS–COM'TEST and seeking HIV testing were (1) reassurance (86%), (2) routine testing (42%) and (3) recent risky sexual exposures (41%). Forty-nine men (9%) returned to ANRS–COM'TEST Study sites to be retested. Of these, seven (14%) returned more than twice. The median time between two tests among these participants was 4 months (IQR, 2.5–6.7). MSM who returned did not differ demographically from those who came only once, but a larger proportion of returners had tested for HIV within the previous 2 years (94% vs 68%, p<0.0001).

Of the 532 participants, 15 (2.8%; 95% CI 1.4% to 4.2%) were tested positive (table 3). Among these, 12 (80%) received confirmatory test results and linked to care and three (20%) lost to follow-up. Their median CD4 count at diagnosis was 550/mm3 (IQR, 484–571). Among the 15 men with positive results, eight (57%) had not been tested for HIV in the previous 2 years (vs 30% among HIV-negative men; p=0.03).

Table 3

Clinical characteristics of men who have sex with men who tested positive for HIV

Positive rapid HIV test, n (%)	15 (2.8)
Loss to follow-up, n (%)	3 (20)
Confirmation of positive rapid HIV test, n (%)	12 (80)
Linkage to care, n (%)	12 (80)
CD4 count at diagnosis (cells/mm3), median (IQR)	550 (484–571)

We collected 514 post-test satisfaction questionnaires; 92% of participants who tested negative (464/504) and 70% of those who tested positive (7/10) reported being ‘very satisfied’ with the intervention. Three-quarter would recommend ‘certainly’ community-based HIV testing and counselling to a friend and 54% of those who tested negative stated that they would choose ‘certainly’ the same venue in the future. The main reasons for which some patients were not ‘very satisfied’ (43; 8%) were the amount of time spent at the testing facility (median, 2 h, including a 45 min explanation of the study and questionnaires completion) and the hours during which testing was available. Among the 440 MSM who had reported to have performed HIV tests in the past, 55% found community-based HIV testing less stressful than traditional HIV testing, while only 5% found it more stressful. Eight-eight per cent found that testing in non-medical settings offers a best welfare. More than 98% of participants attested they could address sexuality openly with peers and no one reported feeling judged.

From the testers' point of view, peers were satisfied from their new activities. The different steps of the 592 testing and counselling procedures performed were respected (more than 90% of completion for every step). Difficulties in handling tests were rarely reported (<2%) by testers, except concerning self-drawn blood samples (19%) and blood collection by testers (14%). A second test had to be performed for eight of the 592 tests (1.5%) because an insufficient amount of blood had been collected. The results of each of these second tests were negative. No other adverse or unexpected events were reported. During pre-test counselling, major concerns were not addressed rarely: risk perception in 4% of interviews, information about HIV transmission routes in 7% and the anticipation of test results in 8%. During post-test counselling, the test result was not explained in 6% of cases, strategies for a better prevention was not discussed in 8% and information about STIs testing was not given in 10% of cases.

Discussion

ANRS–COM'TEST was an HIV testing programme that targeted MSM and studied community-based HIV testing using rapid HIV tests. Counselling and testing were performed by community members who were not healthcare workers. The programme reached MSM who reported high rates of risky sexual behaviour known to be risk factors of seroconversion for HIV.23–25 Sexual behaviours in this population is similar to that of MSM who attend commercial gay venues in Paris in whom HIV prevalence was estimated recently at 18%.26

A substantial proportion of men enrolled in this study had not been tested for HIV recently, even though they were at high risk of HIV infection. These results may suggest that community-based HIV testing programmes may be attractive and convenient for population groups that have not been reachable with traditional HIV testing methods as it has been shown recently by a community-based HIV testing programme in the UK.17 The authors demonstrated that MSM at the community-based programme were less likely to have been tested previously compared with people who were seen at genitourinary medicine clinics.

The programme also reached MSM regularly tested. Moreover, although our study lasted only 12–17 months according to the city, almost 10% of participating men were tested twice or more. A community-based voluntary counselling and testing programme for MSM in Geneva has shown that the proportion of MSM who return for testing is likely to increase over time.15 The MSM who returned for testing in the COM'TEST programme were also tested significantly more often for HIV than men who came once. Increased availability and selection of HIV testing services may therefore encourage even those who already test regularly in traditional programmes to test more often, thereby moving HIV diagnoses to earlier in infection. In addition to community-based HIV testing, other HIV testing strategies such as home tests or tests available in pharmacies may also be interesting to supplements pre-existing HIV testing services and increases access to HIV testing in high-risk groups.27 28 However, additional data are needed on benefits and harms of these strategies.

Counselling and testing were performed by community members who are considered as peers. Although they were not healthcare workers, they were able to handle tests correctly, deliver test results, refer men according to their test result and respect counselling and testing procedures along the whole study period. They added testing and awareness of HIV serostatus to comprehensive prevention counselling in line with participants' lives that is one of the added values in comparison with HIV testing in the healthcare system. The high rate of satisfaction in MSM who participated in this study also shows that community-based HIV testing and counselling was largely acceptable. Overall, participants reported feeling more comfortable with testing and counselling with peers. The Barcelona Checkpoint in Spain that is completely staffed by peers as ANRS–COM'TEST drawn the same conclusions.20 Reasons for not being satisfied of the programme were linked to the study part that was too long and imposed tight opening sessions. The study part may curb some men to come for testing; attendance may be higher in the real life.

In France, the HIV prevalence was 1.6% among MSM who attended voluntary counselling and testing clinics29 and 2.2% in a community-based testing in Paris both run by medical staff.18 In our study, 2.8% of men tested positive for HIV, and all those who underwent confirmatory testing were found to be HIV infected. The HIV prevalence was high, and consistent with rates seen in other recent community-based HIV testing programmes in Europe. During the first year of implementation, these programmes found MSM prevalence rates of 2.4% in Geneva, Switzerland,15 3.2% in Brighton, UK,17 3.2% in Barcelona, Spain,20 and 5.2% in Amsterdam, the Netherlands.14 These results suggest again that community-based HIV testing programmes like ANRS–COM'TEST could reach MSM who are at high risk of HIV infection. MSM with an HIV-positive test have been tested less often in the previous 2 years than men with a negative test; this result suggests that also this programme could reach MSM at high risk who were not tested recently in other testing services.

A large proportion of those men who tested positive for HIV at AIDES CBO were referred and linked to care. They were diagnosed at early stages of the disease, with median CD4 counts of 550/mm3—higher than previously seen in traditional testing programmes. Data from the French HIV surveillance system in 2009 demonstrate that 20% of MSM were diagnosed with CD4 counts <200/mm3 and 62% had CD4 counts <500/mm3.2 4 These results are consistent with the British study, which found a median CD4 count of 431/mm3 among MSM who were diagnosed in community-based programmes compared with 311/mm3 among those who were diagnosed in genitourinary medicine clinics.17 When HIV testing is performed by peers at CBOs, MSM who show evidence of repeated risky sexual behaviour may feel less judged than in a traditional HIV testing programme run by healthcare workers. This difference in attitude may lead MSM to check their serostatus more frequently and sooner after exposure, resulting in earlier diagnoses.

This study has several limitations. First, ANRS–COM'TEST was conducted in four French cities but more than half of the participants enrolled in Paris. Community-based HIV testing programmes may be attractive and efficient in large urban areas, but perhaps less so in smaller cities, where an outreach approach may work better. Second, we certainly underestimated the proportion of MSM who would return for regular testing at CBOs since the study duration was short and many participants found the time spent completing questionnaires too lengthy. Third, the quality was assessed with a form completed by the community staff. The completion rates of different steps of the testing and counselling procedure may be overestimated. Finally, the number of HIV diagnoses was small. The prevalence and median CD4 count among the few HIV-infected participants should therefore be interpreted with caution.

The ANRS–COM'TEST Study was an HIV testing and counselling programme exclusively based on MSM community. It showed first that this type of programme could reach high-risk MSM, of whom a substantial proportion had not been tested recently; second, peers testers involved in such programmes may have the capacity to perform the test into a comprehensive prevention approach; third, participants tested positive are at early stages of disease and that a high proportion of them are linked to care. The programme demonstrated that community-based HIV testing delivered by non-medical staff could increase access to and choice of HIV testing facilities, supplementing existing HIV testing programmes. Based on this study results, French ministry of health recently authorised community-based HIV testing by non-healthcare CBO workers30 as a complementary HIV testing service in France. Community-based HIV testing programmes should also be assessed in other populations at high risk for HIV infection and late presentation to care, such as sub-Saharan African immigrants.31 32