Does selected ecological evidence give a true picture of HIV transmission in Africa?

Current reconsideration of the contributions of various modes of transmission to HIV epidemics in sub-Saharan Africa has important implications for HIV prevention. In recent reviews, we argue that accumulated evidence supports the hypothesis that unsafe health care fuels Africa's HIV epidemics. In response, critics have presented selected ecological evidence - age and sex distribution of HIV infections and geographic distribution of hepatitis C virus infections - to support the conventional hypothesis that sexual contact accounts for most HIV infections. In this communication, we examine critics' evidence and arguments. Critics ignore or reject important evidence including, for example, large numbers of unexplained HIV infections in children, strong associations between incident HIV and injections, and genetically-linked infections in persons with no known sexual contact. We urge that research projects in Africa disclose unpublished relevant evidence on risks for incident HIV. In any case, because each iatrogenic infection causes subsequent linked infections, the impact of HIV transmission through health care has been underestimated. We commend the emerging consensus to improve the safety of health care delivery. In countries with generalized HIV epidemics, we urge public education about the risks for HIV acquisition from unsterile health care.