Most HIV surveillance has been performed through serologic surveys in relatively stable, accessible populations. Similar surveillance, with or without counseling and testing, in populations that are hard-to-reach, presents logistical challenges, including the selection of laboratory testing strategy and algorithm. The advent of rapid serologic assays for HIV now allows for on-site testing, including confirmatory testing, and rapid provision of test results and counseling. The possibility of only a single contact makes repeat sampling, which current diagnostic testing recommendations include, difficult. To address the logistical complexities in surveillance in hard-to-reach populations and the increased availability of rapid tests, we propose adapting the testing strategies for HIV of the World Health Organization/the joint United Nations Programme on HIV/AIDS in order to facilitate this surveillance, including, where carried out, the provision of test results back to individuals. The choice of enzyme-linked immunosorbent assay (ELISA) versus rapid testing for these settings is discussed, as is the choice of specimen - blood, oral fluid, or urine. Three appendices summarize: (1) test algorithms for the various testing strategies; (2) advantages and disadvantages of ELISA and of rapid test formats, and (3) the characteristics and status of currently available rapid HIV tests. We also discuss the potential application of the recently developed 'detuned' methodology for estimating HIV incidence in hard-to-reach populations.