Antiretroviral therapy is finally reaching millions of persons with HIV infection in Africa. While countless lives have been saved, it is becoming increasingly clear that deaths in the first 6 months on ART are dramatically elevated as compared to industrialized settings, - perhaps claiming as many as five-fold more than in industrialized settings. Even these alarming estimates, however, are likely underestimates because high losses to follow up in resource limited settings mean that many deaths are not ascertained. The causes of these early deaths, furthermore, are not well explained. Therefore the proposed Aims seek to (1) evaluate a sampling-based technique that we have developed to understand the true magnitude of early mortality; (2) determine whether undiagnosed opportunistic infections are the cause of early mortality on ART by collecting biological specimens and investigating them intensively in San Francisco and (3) develop a clinical prediction rule to risk stratify individual patients for early mortality at the time or ART initiation. Overall, this study seeks to understand the magnitude and determinants of early mortality on ART in HIV infected patients in Africa in order to develop a targeted intervention in the future to reduce it. These Aims will also provide a platform for completing further career development toward independence as a scientist. RELEVANCE: Early mortality on ART may claim 1 in 10 patients of the millions of patients starting ART in Africa. These deaths are likely due to treatable infections. Therefore, a better understanding of early mortality on ART in Africa has the potential to improve ART delivery and save many lives.