Social mixing and correlates of injection frequency among opioid use partnerships

January 2018



As resources are deployed to address the opioid overdose epidemic in the USA, it is essential that we understand the correlates of more frequent opioid injections-which has been associated not only with HIV and HCV transmission, but also with overdose risk-to inform the development and targeting of effective intervention strategies like overdose prevention and naloxone distribution programs. However, no studies have explored how characteristics of opioid use partnerships may be associated within injection frequency with opioid partnerships.


Using baseline data from a trial of a behavioural intervention to reduce overdose among opioid users in San Francisco, CA, we calculated assortativity among opioid use partnerships by race, gender, participant-reported HIV- and HCV-status, and opioids used using Newman’s assortativity coefficient (NC). Multivariable generalized estimating equations linear regression was used to examine associations between individual- and partnership-level characteristics and injection frequency within opioid use partnerships.


Opioid use partnerships (n=134) reported by study participants (n=55) were assortative by race (NC=0.42, 95%CI=0.33-0.50) and participant-reported HCV-status (NC=0.42, 95%CI=0.31-0.52). In multivariable analyses, there were more monthly injections among sexual/romantic partnerships (β=114.4, 95%CI=60.2-168.7, p<0.001), racially concordant partnerships reported by white study participants (β=71.4, 95%CI=0.3-142.5, p=0.049), racially discordant partnerships reported by African American study participants (β=105.7, 95%CI=1.0-210.5, p=0.048), and partnerships in which either member had witnessed the other experience an overdose (β=81.8, 95%CI=38.9-124.6, p<0.001).


Social segregation by race and HCV-status should potentially be considered in efforts to reach networks of opioid users. Due to higher injection frequency and greater likelihood of witnessing their partners experience an overdose, individuals in sexual/romantic opioid use partnerships, white individuals in racially homogenous partnerships, and African American individuals in heterogeneous partnerships may warrant focused attention as part of peer- and network-based overdose prevention efforts, as well as broader HIV/HCV prevention strategies. Developing and targeting overdose prevention education programs that provide information on risk factors and ways to identify overdose, as well as effective responses, including naloxone use and rescue breathing, for more frequently injecting networks may help reduce opioid morbidity and mortality in these most at risk groups.