Randomized controlled trials (RCTs) remain the accepted "gold standard" for determining the efficacy of new drugs or medical procedures. Randomized trials alone, however, cannot provide all the relevant information decision makers need to determine the relative risks and benefits when choosing the best treatment of individual patients or weighing the implications of particular policies affecting medical therapies. To demonstrate the limitations of RCTs in providing the information needed by medical decision makers, and to show how information from observational studies can supplement evidence from RCTs. Qualitative description of the limitations of RCTs in providing the information needed by medical decision makers, and demonstration of how evidence from additional sources can aid in decision making, using the examples of deciding whether a 60-year-old woman with mildly elevated blood pressure should take daily low-dose aspirin, and whether a hospital network should implement carotid artery surgery for asymptomatic patients. Even the most rigorously designed RCTs leave many questions central to medical decision making unanswered. Research using cohort and case-control designs, disease and intervention registries, and outcomes studies based on administrative data can all shed light on who is most likely to benefit from the treatment, and what the important tradeoffs are. This suggests the need to revise the traditional evidence hierarchy, whereby evidence progresses linearly from basic research to rigorous RCTs. This revised hierarchy recognizes that other research designs can provide important evidence to strengthen our understanding of how to apply research findings in practice.