Summary: 1. A clinical trial is a powerful tool for showing whether an intervention works, but the heterogeneity of trial participants means it is a mistake to assume that the overall (or group) benefit of an intervention found in such a trial is the same for every participant. 2. The absolute benefit of an intervention is greater for trial participants (and for patients generally) whose baseline risk for a bad outcome is high than it is for those whose baseline risk is low. 3. A quality improvement program in any one organization is like an individual patient, in the sense that it is highly complex, is unstable (i.e., changes over time), and its local circumstances are unique, all of which make it hard (although not impossible) to judge whether a quality improvement program in any particular setting actually works, and to know whether it would work elsewhere.
