The prevalence of injecting drug use in a Russian city: implications for harm reduction and coverage.

AIM

This study sought to estimate the prevalence of injecting drug users (IDU) in Togliatti city and to examine the implications of these estimates for HIV prevalence and harm reduction.

DESIGN

Routine data sources of IDUs were identified. Covariate capture-recapture techniques were used on the individuals identified on the three data sources and used to estimate the number of IDU 'not observed' by the data sources, and thereby estimate the prevalence of IDU.

SETTING

Togliatti City, Samara Oblast, Russian Federation.

PARTICIPANTS

IDUs recorded on three data sources (narcology records, HIV positive test results and police arrest data) during 2001.

MEASUREMENTS

Poisson regression models were fitted to the observed data, with interactions between data sources fitted to replicate 'dependencies' between those data sources. To select the best model the goodness of fit was approximated by chi2 distribution and the best-fitting model was selected on the basis of standard information criteria and log likelihood ratio tests.

FINDINGS

The total estimated population of IDUs is 20 226 [95% confidence interval (CI): 16 971-24 749] giving a population prevalence of 5.4% (95% CI: 4.5-6.6%) of the registered population and 2.7% (95% CI: 2.4-3.5%) of the population (including migrants) aged 15-44 years. For every one IDU in contact with a service there were three out of contact.

CONCLUSIONS

There is a high prevalence of IDU which, in the context of a fast-emerging IDU-associated HIV epidemic, will have serious public health implications.