High levels of HIV testing in men who have sex with men (MSM) remain key to reducing incidence, particularly in men who have condomless anal intercourse (CAI) with multiple partners. There is little evidence about the effectiveness of free HIV self-testing (HIVST) to increase HIV diagnosis rates in MSM. We aimed to assess if the offer of a single free HIVST kit led to increased diagnosis of HIV infections that linked to care.
SELPHI is an internet based, open-label, randomized controlled that used online advertising to recruit men potentially interested in HIVST. Enrolment criteria were male (including trans), aged ≥16 years, ever had anal intercourse (AI) with a man, not known to be HIV positive and consent to link to national HIV surveillance databases (to ascertain new HIV diagnoses and linkage to care). Participants were randomly allocated 3:2 at enrolment to a free HIVST (Baseline Test [BT]) versus no free HIVST (no Baseline Test [nBT]). Online surveys collected data at baseline, 2 weeks (2w) (BT only) and 3 months (3m) post-enrolment. Men in BT were asked about HIVST use and linkage to care if reactive. Primary outcome was a confirmed new HIV diagnosis within 3m of enrolment.
10,111 men were randomized (6049 BT; 4062 nBT); median age 33 years (IQR 26-44); 89% white; 20% born outside UK; 0.8% trans men; 47% degree educated; 15% never HIV tested; 8% ever and 4% currently on PrEP. At enrolment 89% reported AI and 72% CAI with ≥1 male partner in previous 3m. 4194/4695 (89%) in BT reported using the HIVST kit. No significant difference at 3m in confirmed new HIV diagnoses (primary outcome)(p=0.64, 19 [0.3%] in BT vs 15 [0.4%] in nBT). Men randomized to BT were more likely to HIV test in 3m after enrolment (96% vs 42%; risk ratio 2.27 95%CI 2.13, 2.40), but a higher proportion in nBT tested for HIV in the 3m after enrolment (42%) compared to 3m before (21%). STI testing rates between arms were similar (22% BT vs 25% nBT).
Reflecting national declines in MSM, new HIV diagnoses were low in both arms by 3 months after enrolment, with no significant difference between men randomized to receive an HIVST kit (BT) and those who were not (nBT). Men randomized to nBT may have been motivated to HIV test through other routes in the 3 months after enrolment. However, HIV testing rates were overall higher in the 3 months after enrolment in those offered HIVST, with similar rates of STI screening.