Numerous studies have documented a positive association between weekend hospitalization and mortality. Some researchers have argued that these associations are causal and arise from weekend reductions in hospital staffing. Others have suggested that the observed correlations reflect differences in the unobservable characteristics of weekend versus weekday patients. The existing literature has provided only limited evidence of the specific pathways through which weekend hospitalization might affect health outcomes. This article outlines the limitations of the existing research and highlights the need for in-depth condition-specific studies that acknowledge the potential threat posed by selection bias. © 2008 Future Drugs Ltd.