BACKGROUND:
In developing countries, most people infected with HIV do not know their infection status. We aimed to assess whether HIV testing could be increased by combination of community mobilisation, mobile community-based voluntary counselling and testing (VCT), and support after testing.

METHODS:
Project Accept is underway in ten communities in Tanzania, eight in Zimbabwe, and 14 in Thailand. Communities at each site were paired according to similar demographic and environmental characteristics, and one community from each pair was randomly assigned to receive standard clinic-based VCT (SVCT), and the other community was assigned to receive community-based VCT (CBVCT) plus access to SVCT. Randomisation and assignment of communities to intervention groups was done by the statistics centre by computer; no one was masked to treatment assignment because the interventions were community based. Intervention was provided for about 3 years (2006-09). The primary endpoint of HIV incidence is pending completion of assessments after the intervention. In this interim analysis, we examined the secondary endpoint of uptake in HIV testing, differences in characteristics of clients receiving their first HIV test, and repeat testing. Analyses were limited to clients aged 16-32 years. This study is registered with ClinicalTrials.gov, number NCT00203749.

FINDINGS:
The proportion of clients receiving their first HIV test during the study was higher in CBVCT communities than in SVCT communities in Tanzania (2341 [37%] of 6250 vs 579 [9%] of 6733), Zimbabwe (5437 [51%] of 10,700 vs 602 [5%] of 12,150), and Thailand (7802 [69%] of 11,290 vs 2319 [23%] 10,033). The mean difference in the proportion of clients receiving HIV testing between CBVCT and SVCT communities was 40·2% (95% CI 15·8-64·7; p=0·019) across three community pairs (one per country). HIV prevalence was higher in SVCT communities than in CBVCT communities, but CBVCT detected almost four times more HIV cases than did SVCT across the three study sites (952 vs 264; p=0·003). Repeat HIV testing in CBVCT communities increased in all sites to reach 28% of all those testing for HIV by the end of the intervention period.

INTERPRETATION:
CBVCT should be considered as a viable intervention to increase detection of HIV infection, especially in regions with restricted access to clinic-based VCT and support services after testing.

FUNDING:
US National Institute of Mental Health, HIV Prevention Trials Network (via US National Institute of Allergy and Infectious Diseases), and US National Institutes of Health.