BACKGROUND: We explored the relevance of simple markers (clinical or laboratory markers not requiring sophisticated laboratories) in the decision of initiation of therapy in resource-poor settings. METHODS: Among HIV-infected Ethiopian cohort participants, simple markers predicting short-term death were examined using time-dependent Cox proportional hazards models. Timing of hypothetical treatment was compared between guidelines using the simple markers (based on presence of at least one marker), guidelines recommended by the United States Department of Health and Human Services (based on CD4 cell count and viral load), and guidelines for resource-limited settings recommended by the World Health Organization (WHO). RESULTS: From February 1997 to August 2001, 35 deaths were recorded among 155 HIV-positive participants. Simple independent predictors of death were low body mass index, HIV-related conditions, anaemia, and lymphocyte count < 1500 x 106/l. In such time as was covered by our study, 135 (87%) of 155 cohort participants would have had the same management under both the simple markers and the DHHS guidelines, i.e., would have been treated (n = 114, 74%) or not treated (n = 21, 14%). Of the 114 participants hypothetically treated under either set of guidelines, 91 (80%) would have started treatment at the same time. Application of the WHO guidelines for resource-limited settings (without CD4 cell counts) would have resulted in 11 participants dying without ever meeting a treatment indication during regular follow-up visits. CONCLUSION: Simple markers for the initiation of highly active antiretroviral therapy were identified among HIV-infected Ethiopian patients. The validity of these markers for monitoring patients' improvement following therapy remains to be evaluated.