Injecting drug users (IDUs) should be considered a 'partially hidden population' at high risk for HIV infection. In almost all locations it should be possible to locate and conduct research with IDUs, but it will probably never be possible to enumerate or draw random samples from an IDU population. Surveillance research studies with IDUs should include risk behaviors, as surveillance of HIV infection only will not be sufficiently time sensitive, and be used to develop and refine HIV prevention programming for the population. Contacts with IDUs can be developed at multiple settings, including voluntary treatment programs, law enforcement settings, and through 'street outreach.' Each type of setting has different advantages, disadvantages and ethical concerns. HIV testing as part of surveillance also raises additional important ethical concerns. The primary risk behaviors that should be included in surveillance studies are 'sharing' of drug injection equipment, the potential for rapid partner change among risk partners, and sexual risk behavior. Additional important objectives for surveillance research include: (1) the size of the local IDU population, (2) patterns of drug use, (3) availability injection equipment, (4) participation in prevention activities, and (5) access to and use of anti-retroviral treatments. HIV incidence is an ultimate objective for surveillance research, but there are no currently available cost-efficient methods for studying HIV incidence, so estimation from indirect measurements is usually required.