OBJECTIVES To investigate how the cost-effectiveness of preventing or treating HIV/AIDS varies by a series of potentially relevant external factors. METHODS We reviewed the cost-effectiveness evidence for HIV/AIDS prevention and treatment in the Cost-Effectiveness (CEA) Analysis Registry at Tufts Medical Center (www.cearegistry.org) from 2002-06, and then constructed efficient frontier curves (EF curve) in terms of incremental cost per QALY ratios based on extracted evidence using methods introduced by Institute for Quality and Efficiency in Health Care (IQWiG), Germany. All articles that report cost-effectiveness as USD/QALYs were selected, including information on the following factors: payer perspective; prevention stage; intervention type, and country of study. We excluded articles with time horizon < 30 years and those without discounting. RESULTS Of 237 eligible articles in the CEA Registry, we extracted 11 HIV/AIDS-related articles, which included 84 individual cost-effectiveness (CE) ratios. Plotted EF curves were visually distinct, depending on the prevention type, with primary prevention interventions being most efficient. With respect to country, the EF curve for the U.S. and South Africa were separate, with the curve for South Africa lying near the vertical axis, reflecting low cost per QALY ratios for interventions studied. Subgrouping by payer perspective (societal or health care payer) and intervention type (pharmaceutical, education, and screening) did not affect the separation of the lines. CONCLUSIONS The CE ratio clustering in HIV/AIDS was found on the C-E plane, which suggested separate EF curves should be considered by stratifying by prevention stage and country. These results may differ for other diseases, but this analysis shows that stratified EF analysis could help to develop a deeper appreciation of cost-effectiveness beyond crude cost per QALY ratios without stratification.