Of 1,015 patients seen, 63 percent were homeless, and 37 percent were not. Using the HEADSS interview instrument, homeless and runaway youth were compared with nonhomeless youth in several areas including risk for HIV infection. Results showed that homeless teenagers tended to be younger, female, and white in comparison to their nonhomeless counterparts. They were more likely to have dropped out of school and were far more likely to be depressed and actively suicidal. They demonstrated all forms of drug abuse, engaged in first sexual intercourse at an earlier age, experienced a higher incidence of sexual abuse and prostitution, and were six times more likely to be at risk for HIV infection. A primary advantage of the HEADSS instrument is that it prompts health care professionals to ask questions they might otherwise forget or ignore. It also helps to coordinate patient care and to plan and raise funds for more effective intervention programs. Future research is suggested that includes more extensive standardization of the HEADSS instrument, comparison of data collected by physicians and nonphysicians, and analysis of data collected through interactive computer-assisted programs. 25 references and 2 tables