Article Text

Original research
Online self-sampling kits to screen multipartner MSM for HIV and other STIs: participant characteristics and factors associated with kit use in the first 3 months of the MemoDepistages programme, France, 2018
  1. Delphine Rahib1,2,
  2. Heloïse Delagreverie3,
  3. Audrey Gabassi3,
  4. Thanh-Thuy Le Thi4,
  5. Eleonore Vassel4,
  6. Pierre Vodosin4,
  7. Benjamin Leveau4,
  8. Amandine Pisoni5,
  9. Edouard Tuaillon5,
  10. Julien Digne6,
  11. Vinca Icard4,
  12. Constance Delaugerre3,
  13. Nathalie Lydié1
  14. pour le groupe MémoDépistages
  1. 1 Sexual Health Unit, Santé publique France, Saint Maurice, France
  2. 2 iPLESP UMRS1136, INSERM, Paris, France
  3. 3 Microbiology Department, Hopital Saint-Louis, Paris, France
  4. 4 Centre de ressources biologiques Nord, Hospices Civils de Lyon, Lyon, France
  5. 5 Département de bactériologie-virologie, Inserm UMR 1058, CHRU de Montpellier, Montpellier, France
  6. 6 Alphabio Européen 13003, Marseille, France
  1. Correspondence to Delphine Rahib, Sexual Health Unit, Santé publique France, Saint Maurice 94415, France; delphine.rahib{at}


Objectives In 2017, to reduce the proportion of men who have sex with men (MSM) in the undiagnosed HIV population in France (38%), HIV screening is advised each 3 months and STI screening is advised each year in multipartner MSM. Despite the range of testing solutions, over 40% of MSM were not tested for HIV and over 50% for STIs in the past year. Based on international experiments that offer screening solutions via online advertising, the French National Health Agency launched a programme (MemoDepistages) to provide a free self-sampling kit (SSK) for HIV and STIs. This article analyses the sociodemographic and behavioural characteristics of MSM in terms of kit acceptance and sample return.

Methods Participants were registered for the programme online after ordering an SSK. The study included men aged over 18 years, living in one of the four selected French regions, and willing to disclose their postal and email address; they had health insurance, acknowledged more than one male partner in the past year, indicated a seronegative or unknown HIV status and were not taking medically prescribed pre-exposure prophylaxis drugs. Samples were collected by users and posted directly to the laboratory. Characteristics associated with kit acceptance and sample return were analysed using logistic regression.

Results Overall, 7158 eligible MSM were offered to participate in the programme, with 3428 ordering the kit (47.9%) and 1948 returning their sample, leading to a return rate of 56.8% and an overall participation rate of 27.2%. Acceptance and return rates were strongly associated with sociodemographic characteristics, mainly education level but not with behavioural characteristics. Non-college graduates had lower acceptance (44.2%) and return rates (47.7%).

Conclusion The programme rapidly recruited a large number of MSM. It removed geographical inequalities related to screening access.

  • screening
  • service delivery
  • men

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


In France, the population of men who have sex with men (MSM) is the most affected by HIV and STIs. In 2017, they had the highest undiagnosed HIV prevalence in the country,1 while the number of STI cases in this population has increased for several years. To reverse this trend, the French Haute autorité de santé recommended quarterly HIV screening in high-risk MSM.2 This strategy is based on a wide range of testing solutions: in addition to reimbursed laboratory tests and free testing in sexual health clinics, rapid HIV tests can be performed by family practitioners and trained non-governmental organisations. In September 2015, a blood-based HIV self-test was approved for over-the-counter sales, and among MSM tested in 2017, 5% used it for their last screening.3 Despite this wide range of solutions, in 2017, only 53% of MSM were tested for HIV in the past 12 months, while very few (15%) had used three or more tests.4

At the time of this study, no national guidelines were in place for a global approach to STI testing in MSM. However, the French Society of Dermatology recommended for sexually active MSM to be tested annually for syphilis, hepatitis C (HCV), as well as Chlamydia trachomatis (CT) and Neisseria gonorrhoea (NG) in pharyngeal, urine and anal samples.5 Hepatitis B (HBV) should also be tested, with vaccination proposed depending on the result. STI testing options are limited in comparison to those for HIV. Tests are available at sexual health clinics or laboratories on prescription. Rapid testing performed by non-governmental organisations has existed for HCV since 2016. Currently, however, there is no self-managed solution for STI screening comparable to the HIV self-test in France. The screening rates for STIs are poorer than those for HIV. For example, in 2017, 46.4% of sexually active MSM were screened for HCV, 43.6% for syphilis, 36.3% for Ct and 32.5% for Ng during the past 12 months (personal communication, Eras Study6).

To meet the new French guidelines and increase the frequency of HIV and STI testing among MSM, there is a need to develop novel approaches and engage MSM where testing was not already offered. Novel outreach design needs to be conceived for modern dating approaches such as dating apps. Using online communities, services can now offer self-sampling kits (SSKs) for HIV screening to key populations through distance facilities.7 Since 2016, the UK has provided a national home SSK service free of charge.8 Several studies have analysed the factors associated with HIV SSKs use, showing that users have a higher education level, better income and are younger7 9 than those who do not use it. Compared with heterosexual men, MSM also tend to be more willing to use such services.7 9–15 Most SSKs propose blood sampling for HIV screening or local sampling for Ct/Ng screening. Only a few propose both types of sampling in the same kit to detect both viral and bacterial infections.7

Building on the results from these projects, the French public health agency developed the MemoDepistages programme. With an 18-month follow-up, this programme aims at increasing quarterly testing among multipartner MSM recruited trough dating apps and targeted websites. It offers an SSK for HIV, HBV, HCV, syphilis, Ct and Ng screening. The SSK included a microtainer tube and two finger prick tests for blood collection, a uriswab device to collect urine and two swabs and PCR-compatible medium for anal and pharyngeal samples. To our knowledge, this was the most comprehensive STI SSK described in international studies.

To date, few studies have explored the step-by-step process of screening using these services: participants first need to subscribe to the service, perform the self-sampling and then send it to the laboratory. This article analyses the different factors that impact the overall participation in the first screening round of the programme.


Study population

Men were recruited between 11 April 2018 and 10 June 2018 through targeted online advertisement, MSM sex apps and community websites. Geolocated advertisements redirected viewers to a 5 min questionnaire investigating the sociodemographic and behavioural characteristics of MSM and verifying their eligibility. Eligibility criteria were as follows: males over 18 years, living in one of the four French regions with the highest HIV prevalence1 (Auvergne Rhone Alpes, Ile de France, Occitanie Est and Provence Alpes Côte d’Azur), and willing to disclose their postal and email address; they acknowledged more than one male partner during the past year, indicated a seronegative or unknown HIV status. Men using pre-exposure prophylaxis (PrEP) with prescription were excluded from the study because an appointment with their practitioner every 3 months was mandatory for PrEP prescription. They benefited from an STI screening during this consultation. However, men who used PrEP without any medical follow-up could be recruited. According to French regulation on biomedical research, only participants who declare to be affiliated to the national health insurance have been recruited.

Definitions and measures

Eligible men were offered to participate in the MemoDépistages programme, in which case an STI SSK would be sent to them within 24 hours (online supplemental file 1). Those who accepted to participate in the programme and ordered the kit are referred to as subscribers. The acceptance rate was defined as the proportion of subscribers out of eligible men.

Supplemental material

They collected their samples and sent them to the laboratory in a prepaid envelope. The men who sent at least one sample to the laboratory before 1 September 2018 are referred to as participants. The return rate was thus defined as the proportion of participants out of subscribers.

The overall participation rate was defined as the proportion of participants among eligible men (figure 1).

Figure 1

Flow chart of the first screening of the MemoDepistages programme. PrEP, pre-exposure prophylaxis.

Statistical analyses

Statistical analyses were performed using Stata V.14.16 As data were collected online using an automatic completeness check, there were no missing data in the dataset. Sociodemographic and behavioural characteristics associated with men’s acceptance and return of the kit were tested using a χ² test. Since Stata V.14 treats collinearity in models, and the number of events was >10 times the sum of all the terms of interest, variables with a p coefficient ≤0.2 were considered for integration and interpreted in the logistic regression. Possible interactions between independent characteristics were tested by including cross-product terms in the regression, with p<0.05 deemed to be indicative of statistical significance. Significant interaction terms included in the final model are presented in the tables.


Over the 2 months of the recruitment period, 12 758 questionnaires were completed. The screening programme and kit were offered to 7158 eligible men (56.1%). The acceptance rate was 47.9% with 3428 subscribers. Among them, 1948 participants sent samples to the laboratories, leading to a return rate of 56.8% and an overall participation rate of 27.2%.

Description of eligible men

The main reason for ineligibility was not living in one of the four regions of the study (60.6%), followed by having only one male partner during the past year (16.9%).

The median age of eligible men was 31 years (table 1). A majority (68.7%) had a college-level education, lived in a large city of >100 000 inhabitants (52.1%) and considered their financial situation to be good or average (82.1%). In terms of sexual behaviour, 68.1% of participants had more than five partners in the past year, and most had their last intercourse with a non-steady partner (78.5%). Nearly one out of four (21.2%) eligible men had never visited a gay meeting place (ie, bar, club, sauna with or without sex, outdoor gay sex venue). In terms of HIV prevention, 44.1% did not use protection, that is, condoms, PrEP or treatment as prevention during their last sexual intercourse, including 38.6% whose last intercourse was with a non-steady partner. Half the eligible men (49.3%) were tested for HIV in the last 12 months, half of them being tested several times. Screening for Ct and Ng was less frequent, with 29.1% of eligible participants reporting a urinal or urethral test, 13.0% a pharyngeal test and 11.5% an anal test over the past year.

Table 1

Sociodemographic characteristics of men offered to participate in the MemoDepistages programme, 2018 (n=7158)

Acceptance rate and associated factors

The acceptance rate ranged from 41.3% to 51.7% depending on the sociodemographic and behavioural characteristics (table 2).

Table 2

Association between sociodemographic and behavioural factors and overall participation in MemoDepistages first screening, 2018 (n=7158)

In a multivariate analysis, the acceptance rate was strongly associated with sociodemographic data. It was lower in men aged over 30 years (adjusted OR (ORa)=0.79, 95% CI (0.72 to 0.88) vs <30 years) and in those who had not attended college (ORa=0.80; 95% CI (0.72 to 0.89) vs college graduates). However, the acceptance rate was better in men declaring an average (ORa=1.13; 95% CI (1.02 to 1.40)) or bad financial situation (ORa=1.22; 95% CI (1.06 to 1.40)) in comparison with those with a good situation.

Considering the socialisation patterns in the community, the men who visited gay meeting places signed up for the kit and the programme less frequently than the others. The highest difference in the acceptance rate was observed between men who frequently visit gay meeting places (51.7%) and those who never did (42.6%, p<10−3) (ORa=0.65; 95% CI (0.50 to 0.84)).

Acceptance was better in men who had sought HIV screening in the past. Men who had never been tested had a lower acceptance rate than those who were tested once in the past 12 months (ORa=0.74; 95% CI (0.62 to 0.88)). Nevertheless, those who were tested several times in the past year also accepted the kit less often (ORa=0.87; 95% CI (0.76 to 0.99)).

Return rate and associated factors

The return rate varied from 47.7% to 61.4% among the subscribers.

In a multivariate analysis, the return rate was strongly associated with sociodemographic data. It was lower for men who had not attended college (ORa=0.65; 95% CI (0.56 to 0.76) vs college graduates) and for men who reported a difficult financial situation (ORa=0.7; 95% CI (0.58 to 0.86)) compared with those with a good situation.

Men with a high number of partners in the past year (between 11 and 20) returned their kit more often than those with fewer than five partners (ORa=1.29; 95% CI (1.05 to 1.57)). There was no association with preventive behaviours.

As a result of these associations, the overall participation rate was found to be strongly associated with being aged under 30 years (ORa=0.66; 95% CI (0.58 to 0.75) for men aged 30 years and older) and having a college education (ORa=0.83; 95% CI (0.74 to 0.92) for men without a college education). No association was made with the perceived financial situation. A lower overall participation was observed in the Provence Alpes Cote d‘Azur region.

Never visiting gay meeting places and not being tested for HIV in over a year were associated with a decreased overall participation rate (respectively ORa=0.72; 95% CI (0.53 to 0.76) and ORa=0.78; 95% CI (0.64 to 0.95)). However, this association is weaker than that observed with sociodemographic data.


Our results show that 47.9% of eligible men accepted to receive the SSK, with 56.8% of them returning at least one of their samples to the laboratory, thus leading to an overall participation rate of 27.2%. More than one out of four eligible MSM completed the self-sampling for HIV and STIs.

Few studies report the overall participation rate as defined in our study. Usually, evaluations focus mainly on the return rate without considering the acceptance rate. However, studies that investigate different populations or use different recruitment processes may still be used to discuss our results. In the SH:24 experiment conducted over a period of 9 months and targeting people aged 16–30 years without consideration of their sexual orientation, the overall participation rate was 37.6% (388/1,031) after 6 weeks.17 This experiment used a variety of recruitment methods: online as well as advertising displays in university bars and nightclubs and outreach methods. Our internet-only recruitment had a narrower scope but the MemoDepistages programme was able to recruit a large number of high-risk MSM (n=7158) over a short period of time. The main reason for ineligibility was not living in an area of interest. This shows that the targeting worked efficiently and that the programme was properly designed for this population.

Among men who were offered to participate in MemoDepistages, nearly half accepted. While the acceptance rate was lower than that found in a phone-administered health study (82.4% for MSM15), it is consistent with that observed for a Ct screening kit offered to French youths in the Chlamyweb study (around 50%).17

In our study, the recruitment period was 2 months. The return rate of 56.8% is consistent with the rates observed for different long-term SSK services. For example, in the Dean Street @Home Service, 55.2% of participants returned their kit during the 2 years of the evaluation.18 In the evaluation of the Umbrella Health Service, 63.2% of MSM returned the SSK.7 Calculated at least 1 year after implementation, this rate may have further improved over time. In the first year of the national UK HIV self-sampling programme, the return rate was 52.7%, all populations combined, with the rate increasing to 60.4% after 3 years.8 Consequently, the return rate calculated in our study may be seen as a baseline return rate for such a programme in France.

In France, men living in a city with >50 000 inhabitants are usually tested more often than men from rural areas.19 With the SSK, no differences were found according to the size of the place of residence. The online outreach and proposition of the programme cancelled out the territorial inequality usually observed in screening. However, a lower overall participation rate has been observed in the Provence Alpes Cote d‘Azur region. Our investigations revealed that this region has some specificities in terms of delivery services (different work organisation, more of non- standardised mailbox and so on). This result highlighted how the service could be impacted with local characteristics.

The acceptance process is crucial in defining the profile of participants, as it was strongly associated with individual sociodemographic characteristics. As previously found, a young age and high school diploma were associated with a better overall participation rate.7 12 Men aged under 30 years had a better acceptance rate in our study, which impacts their overall participation, even if they returned the kit at the same rate as older participants. Non-college graduates participated less in the programme and also returned their sample less frequently, leading to the highest difference between groups in terms of the overall participation rate (9 points).

Similar associations were also found in a self-sampling study in the general population.20 These characteristics were also associated with screening in the standard health services.4 21 Such individuals were more likely to take part in this new screening offer with the SSK. Nevertheless, despite the disappointing low overall participation rate observed in the population who had never been tested for HIV or had not been tested for >1 year, the benefit in increasing the testing rate for HIV and STIs using SSK could still be attractive when compared with the traditional system, as previously shown in younger populations.22

When looking at the behavioural characteristics, participants with 11–20 partners during the past year used the SSK more often after receiving it. This behaviour corresponds to a global risk perception instead of a reaction to a recent risk, since no difference was observed in the acceptance or return of the SSK for people with and without protection against HIV during their last intercourse. This conclusion is reinforced when considering men’s history of HIV testing. MSM who most frequently used the SSK were already familiar with HIV screening and had already been screened at least once during the last year, although the screening needs to be repeated to meet the national health guidelines. If considering the relation of men with gay meeting places, those who frequently visited those venues took advantage from the non-government organisation’s outreach and screening offer in those venues. Consequently, they already have a better access to screening and education about screening benefits. They order less, but when they did, they are more willing to return the samples.

Our study has several limitations. First, the campaign was advertised online and focused on providing SSKs for HIV and STI screening. We therefore did not recruit men who did not use the internet, and it is probable that those who clicked on the advertisement had a greater general interest in STIs. As a result, our sample is a convenience sample, and thus the results cannot be applied to the entire MSM population. As the recruitment only took place over 6 weeks, men who were resistant to change, whether in the form of new products or screening propositions, would not have shown their interest in the programme. This short-term recruitment, proposing a new way to get screened, may have led to a selection of early adopters profiles. However, it mimics the design of several long-term interventions developed internationally, whose results could be used to inform policymakers on the potential of such initiatives in France.

Second, no information was available to us regarding the difficulties faced during the sampling. The reasons for non-return were not investigated, and we therefore cannot draw on any original data to explain the difference in the return rates between groups. However, qualitative study will investigate keys of the programme success at the end of the programme.

Finally, return rates were calculated using data available on 30 August 2018. After this date, some SSKs were still being returned. This choice led to a small underestimation of the return rate in our sample and may have slightly impacted the population characteristics.


The MemoDepistages study reached its main target of MSM frequently exposed to STIs: a high number of partners during the past 12 months and a frequent unprotected last intercourse. With half of them not being tested during the last 12 months, this population needed to increase its screening frequency. Results of MemoDepistages regarding SSK used suggest that this approach may succeed in removing the traditional geographical inequalities associated with screening access.19 Studies have shown that populations with lower screening rates are those who benefit the most from such interventions in terms of rate increase.17 22 Several factors linked to screening in conventional settings19 were also found in MemoDepistages, and further analysis on the effect of the intervention are needed to conclude on its global impact. Following this initial screening, participants will be able to choose whether they prefer to use the SSK or another screening option for routine quarterly testing over 18 months. An observation of the various screening iterations will allow us to better understand how the SSK completes the current French screening strategies.

Key messages

  • STI self-sampling kit (SSK) is more used when people already have a recent HIV screening experience.

  • Several factors linked to screening in conventional settings were also found to be associated with SSK.

  • It contributes to address the unequal access to testing across the country.

Abstract translation

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Handling editor Jason J Ong

  • Collaborators MemoDepistages group is composed of : Nathalie Lydié, Delphine Rahib, Constance Delaugerre, Héloïse Delagreverie, Béatrice Berçot, Iris Bichard, Hannane Mouhim, Hélène Salord, Vinca Icard, Thanh Thuy Le Thi, Christine Fernandez, Fatima Oria, Hervé Richaud, Sarah Lablotière, Grégoire Eiberlé, Julien Digne, Hacène Khiri, Edouard Tuaillon, Amandine Pisoni, Vincent Tribout, Marie-Noëlle Didelot.

  • Contributors DR and NL conceptualised and managed the study. HD, CD, T-TLT, VI, JD and ET contributed to the study design. DR, HD, AG, EV, PV, BL and AP managed the data collection. DR cured, analysed and interpreted the data to wrote the first full draft of the manuscript. NL and HD provided intellectual input for interpretation and reviewed first draft of the manuscript.

  • Funding This work was supported by the Agence Nationale de Recherches sur le Sida et les Hépatites Virales (ANRS) (grant number ECTZ47249).

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval The protocol was approved by local authorities under the number ID RCB 2017-A00838-45 and ethics committee CPP-Ouest II-ANGERS.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data could be requested to the corresponding author.