Background:  Accurate HIV surveillance of hard-to-reach populations is critical to the global “know your epidemic” mandate. Although half of new HIV cases occur in youth globally, reliable data on HIV seroprevalence among street and out-of-school youth, a subpopulation of youth who are more vulnerable to acquiring HIV, are largely unavailable. We adapted time-location sampling to assess HIV seroprevalence among street and out-of-school youth in 4 Eastern European cities.

Methods:  Cross-sectional assessments were conducted in St. Petersburg, Russia (April to May 2006) and Kiev, Odessa and Donetsk, Ukraine (May 2008 to December 2008). Comprehensive mapping of all street youth locations in each city was conducted to develop city-specific sampling frames. Since the population of street youth occupying a given site did not vary by day of the week or time of day, random selection of sites did not require allowances for day or time. Sites selected into the sample were visited during hours of maximum street youth activity. The assessment included rapid HIV testing for all consenting 15 to 24 year olds (15 to 19 year olds in St. Petersburg) street youth at selected sites and an interviewer-administered survey. Adjusted odds ratios (AOR) were calculated using logistic regression.

Results:  A total of 132 sites were identified, of which 96 were randomly selected to enroll approximately 300 youth per city. Common locations included metro or train stations, street markets, feeding centers or fast-food sites, parks and computer clubs. Of 1,499 youth approached, 1,300 (86.7%) were eligible for participation and 1,242 (95.5%) of those were enrolled. HIV seroprevalence in street youth was 37.4% in Russia and 18.4% in Ukraine. Characteristics independently associated with HIV infection among street youth in Russia and Ukraine, respectively, included: injection drug use (AOR 23.0 and 9.0), sharing needles (AOR 13.3 and 6.8), being diagnosed with a sexually transmitted infection (AOR 2.1 and 2.8), being an orphan/lack of parental support (AOR 3.3 and 1.7).

Conclusions:  Time-location sampling is a probability sampling methodology that can be successfully utilized in hard-to-reach populations. It is an efficient, inexpensive, reproducible and accurate method of HIV surveillance in street and out-of-school youth. Using this methodology, we found that street youth in Russia and Ukraine have extremely high HIV seroprevalences associated with social, sexual and drug-related factors.