In severe sepsis, several studies, a recent meta--analysis and studies evaluating multifaceted strategies for quality improvement, have shown that bundled care can improve survival. Here, the effectiveness of treatments for severe sepsis included in the bundles was analyzed. Despite the observational design of the studies, possible biases are minimized using propensity scores and other adjustments. The results of are very consistent: early sepsis recognition with the administration of broad-spectrum antibiotics in all patients and specific treatments for patients in shock, such as activated protein C, or those on mechanical ventilation play a role in improving sepsis outcome. Hospitals should recognize this new evidence and design strategies to guarantee bundled care for severe sepsis.