Potential risk factors for the transition to injecting among non-injectingheroin users: a comparison of former injectors and never injectors

To compare potential risk factors for the transition to injecting among non -injecting heroin users (NIUs) with different injecting histories. Design. Cross-sectional data from baseline structured interviews with NIUs in a stu dy on transitions to injecting. Sample recruited by outreach or chain-refer ral in New York City (NYC), 1996-1998. Setting. Recruitment of sample and i nterviews conducted in a NYC neighborhood where many drug users reside and/ or use drugs. Participants. Of 575 NIUs, 67% had never injected; 16% had in jected one to nine times (infrequent former injectors (IFI)); and 18% 10 or more times (frequent former injectors (FFI)). Measurements. Controlling fo r age and race/ethnicity, adjusted odds ratios were estimated in multivaria te logistic regression, and differences in means tested by ANCOVA. Findings . FFI (compared to never injectors and IFI) were more likely: to be homeles s; to be unemployed; to be long-time users; to be younger at first heroin u se; to not have initiated heroin use through non-injected routes; to not be afraid of injecting themselves with needles; to sniff heroin with former I DUs; and, for both men and women separately, to have sex partners who were former IDUs. Both FFI and IFI were twice as likely as never injectors to pe rceive that their friends thought that it was "OK" to inject drugs. Conclus ions. FFI have multiple individual and network characteristics that may inc rease their risk of injecting drugs. Interventions among NIUs to prevent tr ansitions to injecting need to ascertain NIUs' injecting history and addres s the many potential risks that FFI have for resuming injecting drug use.