‘ Test n Treat ’ (TnT): a cluster randomized feasibility trial of on-site rapid Chlamydia trachomatis tests and treatment in ethnically diverse, sexually active teenagers attending technical colleges Clinical Microbiology and Infection

Objectives: We conducted a cluster-randomized feasibility trial of 90-minute Chlamydia trachomatis tests and same day on-site treatment ( ‘ Test n Treat/TnT ’ ) in six technical colleges in London, England, to assess TnT uptake rates; follow-up rates; prevalence of C. trachomatis at baseline and 7 months; time to treatment; acceptability of TnT. Methods: Participants completed questionnaires and provided genitourinary samples at baseline and 7 months. Participants were informed that baseline samples would not be tested for 7 months and were advised to get screened independently. Colleges were randomly allocated 1:1 to intervention (TnT) or control (no TnT). One month and 4 months post recruitment, participants at intervention colleges were texted invitations for on-site free C. trachomatis tests. A purposive sample of students who did/did not attend for screening were interviewed (n ¼ 26). Results: Five hundred and nine sexually active students were recruited: median age 17.9 years, 47% male, 50% black ethnicity, 55% reporting two or more sexual partners in the previous year. TnT uptake was 13% (33/259; 95% CI 8.9 e 17.4%) at 1 month and 10% (26/259; 6.7 e 14.4%) at 4 months with overall C. trachomatis positivity 5.1% (3/59; 1.1 e 14.2%). Follow-up at 7 months was 62% (317/509) for ques- tionnaires and 52% (264/509) for samples. C. trachomatis prevalence was 6.2% (31/503) at baseline and 6.1% (16/264) at 7 months. Median time from test to treatment was 15 h. Interviews suggested low test uptake was associated with not feeling at risk, perceptions of stigma, and little knowledge of sexually transmitted infections (STIs). Conclusions: Despite high C. trachomatis rates at baseline and follow-up, uptake of testing was low. Like many countries, England urgently needs better sex education, including making STI testing routine/ normal.


Introduction
Chlamydia trachomatis is a common, often asymptomatic, bacterial sexually transmitted infection (STI) which can lead to pelvic inflammatory disease, ectopic pregnancy, and infertility [1,2] and may be associated with adverse pregnancy outcomes [3]. However, uptake of C. trachomatis testing by 16e24 year olds in many countries is too low to reduce infection rates [1,4e8], and there are often delays in treatment. Bringing novel 90-minute C. trachomatis tests [9,10] and same day on-site treatment ('TnT ¼ Test n Treat') to the community might get more young people treated faster [6,10]. This could reduce rates of infection, onward transmission and adverse reproductive health effects, and save healthcare costs [7,11].
In order to address a number of unknown parameters required for the design of a future definitive study, we conducted a cluster randomized feasibility trial (or pilot study) of frequent, rapid TnT in six technical ('Further Education'/FE) colleges in London, England, over the academic year 2016e17. (FE colleges offer both academic and practical courses such as plumbing and hairdressing, and take many students from socio-economically deprived backgrounds. C. trachomatis positivity may be 6e8% [12e14]. ) We assessed the following feasibility outcomes: recruitment rates TnT uptake rates follow-up rates prevalence of C. trachomatis at baseline and 7 months time to treatment acceptability of TnT.
We selected a cluster design for practical reasons for delivering screening, which would reflect the design of a definitive trial. This was a feasibility study and was not powered to assess the effectiveness of TnT. Although we used a combined C. trachomatis/ Neisseria gonorrhoeae rapid test (Cepheid CT/NG GeneXpert® system [9]), on-site treatment (TnT) was for individuals with C. trachomatis only [15] as participants with N. gonorrhoeae (or C. trachomatis/N. gonorrhoeae dual infection) were referred to a sexual health clinic. Detailed qualitative and economic analyses will be presented elsewhere.

Recruitment and baseline samples
All technical colleges/clusters were eligible and all six approached agreed to participate. As previously described [15], researchers recruited students from public areas at the six colleges. Students were eligible if they were aged 16e24 years and had ever had sexual intercourse. The participant information leaflet and consent form provided information about STIs and the study design (please see Supplementary Material). Participants provided written informed consent. They were asked to complete questionnaires (see Table 1 and Supplementary material), and to provide samples (for research purposes only) in the nearest washroom (urines for males, self-collected vaginal swabs for females) [15]. These samples were stored at e80 C and tested blind at St George's hospital after seven months using the Cobas 4800 CT/NG system (Roche diagnostics) [7]. All participants were warned of the risks of untreated C. trachomatis/N. gonorrhoeae and that their baseline samples would not be tested for seven months, and advised to get checked for STIs independently of the study.

Randomization
After recruitment of all participants, the six colleges were randomly allocated 1:1 into the intervention group (TnT) or control group (no TnT; Fig. 1) by the trial statistician using a computergenerated allocation sequence [15].

Intervention colleges: TnT at 1 and 4 months
One month and 4 months after recruitment (to fit with college Autumn and Spring terms), each of the three intervention colleges were visited on two consecutive days by the research team. We advertised the visit on college websites and notice boards, and texted/emailed participating students the day before the visit and on both days inviting them to come for TnT. Attendees came to a private room to collect a test kit. When they returned with a sample, it was tested for C. trachomatis/N. gonorrhoeae immediately on-site in a pop-up laboratory in a classroom using a 90-min test [15] (one test/participant). Negative results were texted to participants. The research team's nurse health adviser telephoned participants with positive results and met them in another private room in college (same day whenever possible) for confidential treatment for C. trachomatis, partner notification and/or referral.

Control colleges: no TnT
Participants from the three control colleges received texts 1 month and 4 months after recruitment thanking them for being in the study.

Outcome assessment at 7 months
All six colleges were visited again on two consecutive days in the summer term using the same methods as in TnT above, and participants from both groups were invited to provide repeat questionnaires and samples for immediate testing. Same-day results and treatment were provided for all attenders (but these were not part of the TnT intervention). Non-attenders were followed up by text/email and telephone questionnaire and asked to give an address (e.g. home/work/college) if they were willing to provide a postal sample for testing [15].

Honoraria
Participants received £5 in cash when they returned samples at recruitment and £10 after providing samples at the 7-month follow-up. Participants in intervention colleges did not receive honoraria for attending for TnT at 1 month and 4 months, as in the UK people are not usually paid for having an STI test.

Masking
Recruitment of colleges and participants was conducted blind to group allocation. After the first TnT intervention, participants and researchers were no longer blinded.

Main outcome measures
The key values to inform feasibility, sample size, and timescales of a definitive trial were recruitment rates TnT uptake in intervention participants at one and four-months follow-up rates at 7 months prevalence of C. trachomatis at baseline and 7 months time to receiving results and treatment (fidelity of TnT) acceptability of TnT in intervention colleges from thematicallyanalysed semi-structured interviews [16] with purposively sampled students (n ¼ 26 to ensure a range of ages, genders and ethnicities) who did/did not attend for TnT (to be published elsewhere).

Sample size and statistical analysis
Sixty to 100 subjects are sufficient to estimate an event rate with acceptable precision (i.e. sufficiently narrow confidence intervals) in a feasibility study [15,17]. As previously described [18], assuming a 30% recruitment rate [13], we aimed to approach 1600 students to recruit 480 overall (80 per college across six colleges).
Progression criteria to a definitive trial were TnT uptake 60% [13] at 1 and 4 months and TnT being acceptable to participants [16] (intervention colleges only), and follow-up rate 70% [12] at 7 months (all colleges).
Since this was a feasibility study, no significance testing was performed [19]. Descriptive statistics are presented, with corresponding exact 95% confidence intervals. Analyses [18] were performed in Stata version 14. As our analysis was of feasibility outcomes, the sample size and analysis were not adjusted for clustering.

Ethics approval and consent to participate
Bromley REC reviewed the study (reference 15/LO/1929). Parental consent for 16e18 year olds was not required.

Recruitment
Over 3 weeks in September/October 2016, we recruited 509 participants from six colleges (range 78e90 per college). We were unable to obtain information on all non-participants, but completed recruitment forms for 180 non-participants suggested that 67% (121/180) were ineligible due to never having had sexual intercourse, 14% (25/180) were ineligible for other reasons (e.g. not aged 16e24), and 19% (34/180) were eligible but declined.
Participants' median age was 17.9 years and 90% (458) were teenagers (aged 16e19 years). Participants described their ethnicity as black (50%), white (26%), or other ethnic groups (24%). Approximately half (47%, 240) were male, including 117 (23%) black male teenagers. Over half (55%) reported two or more sexual partners in the previous year, and a third (36%) said they had been tested for STIs in the past 6 months. Eligible non-participants (n ¼ 34) were similar to participants in age and ethnicity (median age 17, IQR 17e19; 53% black ethnicity), but a slightly higher proportion (67%) were male. Table 1 shows baseline characteristics of participants from intervention and control colleges.

TnT uptake at 1 month and 4 months in intervention colleges
Thirteen percent (33/259; 95% CI 8.9e17.4%) of intervention participants attended for on-site rapid tests and provided samples at 1 month and 10% (26/259; 95% CI 6.7e14.4%) at 4 months, despite implementing changes suggested by students and staff to increase uptake. These included brief information for tutors to give to their tutorial groups, educational posters (please see Supplementary Material), user-friendly texts, and free condoms. Five students provided samples at both 1 and 4 months. Of 59 tests, three (5.1%, 1.1e14.2) were positive for C. trachomatis. Two students with C. trachomatis only were treated on site (one same day, one next day), and one with dual C. trachomatis/ N. gonorrhoeae infection was referred for treatment as per protocol. Table 2 shows baseline characteristics of participants who did/did not provide samples for TnT were broadly similar, although more TnT attenders than non-attenders had a history of C. trachomatis (13% versus 6%), and more were men who had sex with men (MSM, 15% versus 3%).

Follow-up
Overall follow-up at 7 months was 62% (317/509; 95% CI 58e67%) for questionnaires and 52% (264/509; 95% CI 47e56%) for samples. (A further four participants provided invalid samples: three with no human DNA, one delayed postal sample.) Almost half the participants (46%, 232/509) completed follow-up questionnaires at college, a further 9% (46/509) subsequently completed an online questionnaire and 8% (39/509) a brief telephone questionnaire. These showed 29% of intervention participants and 25% of control participants reported STI testing outside the trial. (Other study-related behaviours reported at follow-up are shown in Table 3). Valid samples for testing were provided at college by 229 (45%) participants and later by post *Five par cipants provided samples at both one and four months. ** Two addi onal samples from each arm did not give a valid result *** One par cipant only returned a sample but did not complete a ques onnaire **** Limited ques onnaire data were collected while informing par cipants by telephone of a posi ve baseline test for five and two individuals in the interven on and control groups respec vely.  Table S1 gives baseline characteristics of those who did/did not provide samples at 7 months follow-up.

Time to results and treatment
For samples provided at college at 1, 4, and 7 months, most results (90%, 259/288) were received by participants the same day. Median time to being informed of a negative result (n ¼ 267) was 2.1 h (IQR 1.8e2.7 h, range 1.5 h to 23 days due to an administrative error). For the 15 cases of C. trachomatis only which were diagnosed in college (2 þ 13 at months 1/4, and 7), ten were treated on-site (six same day, four next day), three were confirmed treated later

Acceptability
Semi-structured interviews in JanuaryeMarch 2017 with 13 students who attended for TnT and 13 who did not suggested that low uptake of TnT was associated with not feeling at risk, perceptions of stigma, and lack of knowledge about STIs. However, all were positive about TnT: 'I think the service you provide is actually very good because like most kids I think they would be too shy to like go out and get checked ….' (male, 16, black, TnT non-attender). Comments from attenders included: 'amazing', 'educational', 'friendly', 'helpful'.

Principal findings
Rapid recruitment of sexually active teenagers was possible with £5 honoraria. However, despite high rates of C. trachomatis at both baseline and follow-up, the proportion of participants attending for non-incentivized college-based TnT was low: 13% at 1 month and 10% at 4 months. Although predetermined progression criteria for a definitive trial were not met, findings provide important insights for designing future studies and for public health policy.

Strengths and weaknesses
This was a unique study in a group of often socioeconomically deprived, ethnically diverse, inner-city teenagers. It included >100 black, sexually experienced teenage males, a group not often included in European STI research studies [4,7]. Participants had high rates of undiagnosed STIs including six participants with heterosexual N. gonorrhoeae, all from black and minority ethnic groups. It is also the first randomized study of rapid tests with on-site C. trachomatis treatment in FE colleges. It was a pragmatic study in a relevant setting to reach sexually active young people. Data on teenage lifestyles may inform future studies.
There are limitations. Opportunistic recruitment meant it was difficult to calculate a recruitment rate. We could not use the college population aged 16e24 (range approximately 500e3000 per college) as the denominator because assessment of eligibility required information on sexual history. As in other studies [4,20] we used self-reported data, which is subject to inaccurate recall. However, reported history of C. trachomatis was similar to rates in 16e24-year-old Londoners taking part in the populationbased National Surveys of Sexual Attitudes and Lifestyles (8.2%, 41/502 in our study versus 7.0%, 19/273 in Natsal-3 UK data archive). Only two-thirds (10/15) of C. trachomatis only positives diagnosed in college were treated on-site. A faster 30-min test might have encouraged more students to wait for results [21], but no such suitable test was available. Although all participants diagnosed with infections were informed that their partners needed treatment, we did not have partners' consent to confirm notification. The study design meant TnT was only available to those already recruited. This would not happen if TnT were rolled out in routine practice. Follow-up rates were lower than the 81% in the recent 'Safetxt' pilot trial [22], but most of their participants were white, and/or aged 20e24. Our findings may not apply to such groups.

Comparison with other studies
Rates of testing were lower than (54e60%) expected from our FE college-based pilot work [13,16,23], but similar to that in 16e29 year olds in a large Dutch register-based C. trachomatis screening trial: 16% in the first round decreasing to 11% in the second [8] with no substantial decrease in STI positivity rates. Another study from a Scottish FE college found 17% C. trachomatis testing uptake in teenagers [24], suggesting this is a challenging group to engage. By contrast, in the French Chlamyweb study [7] uptake by 18e24 year olds of an online offer of home-based C. trachomatis testing was 24% in males and 34% in females with positivity rates of 4.4% and 8.3% respectively. Similarly, in 'SH24', internet accessed postal testing almost doubled uptake of STI testing [20]. However, most participants were white and/or aged 20e30 years. As in other studies [7,16] many of our teenage participants did not want a test kit posted to their home. The high C. trachomatis positivity rates in Chlamyweb and our study were similar to those observed in STI clinics [7] and roughly double the rates in population-based studies in sexually experienced males and females aged 16e24 in England [5] (2.3% and 3.1%) and the USA [25] (1.7% and 3.2% respectively). Finally, there were more MSM among TnT attenders than non-attenders. MSM may be more aware of STI prevention [20].
The median time from diagnosis to treatment (within 1 day) was similar to a recent feasibility study of online C. trachomatis management via an eSexual health clinic [26]. Overall rates of confirmed treatment for C. trachomatis (87%, 13/15) were similar to ChlamywebII [7] (87%, 58/67) and 2014 English National Chlamydia Screening Programme results (91% within 6 weeks of test date [27]). Participants' lack of knowledge about STIs was in line with community-based studies from the USA, Europe, and Australia [16,24,28e30]. Sex education is optional in English state secondary schools.

Conclusions and perspectives
The low uptake of TnT despite high rates of STIs suggests that a definitive trial of TnT using this design is not feasible in FE colleges. It highlights both the difficulties of designing studies to reach sexually active young people, and the crucial need for better sex and relationships education [2]. This should include 'normalization' of STI testing [20] making it routine/acceptable to get checked. However, accessing testing is often problematic [1]. In the UK, funding cuts have closed many sexual health clinics, and relying on internet postal testing may disadvantage vulnerable teenagers [20]. Future trials might evaluate college-wide, multicomponent, combined Education/TnT interventions. This could include lessons offering user-friendly information on STIs, free condoms, and postal test kits perhaps followed by pop-up clinics offering confidential, on-site TnT.

Transparency declaration
This paper presents independent research funded by the Na-