International Journal of Drug Policy 22 (2011) 267–273 Contents lists available at ScienceDirect International Journal of Drug Policy jo ur n al homep age: www.elsevier.com/locate/drugpo Research paper Use of respondent-driven sampling to enhance understanding of injecting networks: A study of people who inject drugs in Sydney, Australia Dana M. Paquette a,∗ , Joanne Bryant a , John De Wit a,b a National Centre in HIV Social Research, University of New South Wales, Sydney, NSW 2052, Australia b Department of Social and Organizational Psychology, Utrecht University, Utrecht, The Netherlands a r t i c l e i n f o Article history: Received 12 January 2011 Received in revised form 18 March 2011 Accepted 22 March 2011 Keywords: Respondent-driven sampling Social networks Injecting drug use HIV Hepatitis C a b s t r a c t Background: Respondent-driven sampling (RDS) is a method for recruiting hidden populations, such as people who inject drugs (PWID). In RDS, participants recruit their peers into the study; who recruited who into the study is tracked, and thus information is gathered on the population’s social networks. The purpose of this study was to use information collected from an RDS study of PWID to determine the size and structure of injecting networks and whether network characteristics are associated with sharing injecting equipment. Methods: A study was launched in Sydney, Australia in 2009 with five seeds, who were asked to recruit three participants each into the survey. This process was repeated until the target sample size was reached. The median size of injecting networks and the homophily (a measure of in-group affiliation) of different subgroups were calculated. Participants’ information was linked with that of their recruiter to form dyads, and multivariate analysis was conducted to determine whether dyad and injecting network characteristics were associated with sharing injecting equipment within the dyads. Results: The injecting networks were large, the lowest median subgroup network size being 12. Homophily estimates indicated a lack of strong ties both within and across groups. In the multivari- ate analysis, factors significantly associated with sharing injecting equipment within dyads were feeling very close to their recruiter and having one or both members of the dyad identify as Aboriginal or Torres Strait Islander and one or both members having not been tested for hepatitis C in the previous year. Conclusion: RDS provided valuable information on injecting networks in Sydney. PWID were shown to be socially connected with a large number of other injectors, and affiliations were formed without regard to demographic or drug use characteristics. Linking information from the recruits with that of their recruiter was a useful way of organizing information to gain a more complete understanding of risk behaviour. © 2011 Elsevier B.V. All rights reserved. Introduction People who inject drugs (PWID) have a high risk for acquir- ing bloodborne viruses, such as hepatitis C and HIV (Abou-Saleh & Foley, 2008; Mathers et al., 2008). Outside of sub-Saharan Africa, injecting drug use is one of the main modes of HIV transmission (Kilmarx, 2009), and, in many countries, PWID are the popula- tion at highest risk for hepatitis C (Abou-Saleh & Foley, 2008). In Australia, in 2009, the prevalence of hepatitis C was 50% amongst PWID attending needle and syringe programmes (National Centre in HIV Epidemiology & Clinical Research, 2010). For PWID, the main risk behaviour for acquiring bloodborne viruses is sharing needles, ∗ Corresponding author at: National Centre in HIV Social Research, Webster Build- ing, UNSW, Sydney, NSW 2052, Australia. Tel.: +61 2 9385 6397; fax: +61 2 9385 6455. E-mail address: Dana.Paquette@unsw.edu.au (D.M. Paquette). syringes and ancillary equipment, such as cookers, filters and water (Santibanez et al., 2006). Given the high rates of bloodborne virus infection in this population, it is important to understand the fac- tors that influence sharing of injecting equipment. There is increasing evidence that aspects of the risk networks of PWID (i.e., people with whom bloodborne virus-related risk behaviour occurs) and their broader social networks have a strong influence on whether injecting equipment is shared (Rhodes, Singer, Bourgois, Friedman, & Strathdee, 2005). For example, the size of a person’s risk network, their position in the network (e.g., in the core or periphery), network member characteristics and the network’s injecting norms have all been associated with inject- ing risk behaviours (De, Cox, Boivin, Platt, & Jolly, 2007). However, despite these findings, and despite the fact that risk behaviours for bloodborne viruses usually happen at the dyad level (between sexual or injecting partners), behaviours are most often measured at the individual level. Surveys are rarely structured so that part- ner or network characteristics can be examined when describing 0955-3959/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.drugpo.2011.03.007