Representativeness of injecting drug users who participate in HIV surveillance: results from Australia's Needle and Syringe Program Survey.

Objective: Australia's annual Needle and Syringe Program (NSP) Survey forms the basis of HIV surveillance among injecting drug users (IDUs) by providing serial point prevalence estimates of patterns of infection and risk behaviors. This study examined the representativeness of NSP Survey samples.

Design: National cross-sectional survey of 3920 NSP clients.

Methods: Demographic and drug use characteristics of respondents and nonrespondents to the 2006 Survey were compared.

Results: Relative to NSP clients who last injected heroin, methamphetamine injectors were significantly more likely to complete the Survey, as were people who had last injected an opioid maintenance pharmacotherapy or morphine. Other variables independently associated with Survey completion were female gender and being aged >or=35 years. Although the median age of NSP Survey samples has increased by 1 year per annum since 2002, the increase has occurred among both repeat and first-time respondents, allaying concerns that an ageing cohort of NSP clients repeatedly completes the Survey and unduly influences its results.

Conclusions: Inferences derived from the Survey results can reasonably be applied to the population of NSP clients, although because older female pharmaceutic injectors may be overrepresented among NSP Survey participants, recruitment strategies to target specific subpopulations (younger male participants) and stratification of main outcomes by age and gender in future analyses may usefully be considered. Although the extent to which Survey results can be generalized to Australia's broader IDU population cannot be ascertained, their consistency with other sources of surveillance data suggests that NSP Survey samples reflect the changing characteristics of Australia's illicit drug markets. Consequently, these are likely to be as representative samples of injectors as it is practical to obtain, and the Australian NSP Survey provides a useful model for blood-borne virus surveillance among IDUs.