Errors conference: executive summary.

T ARTICLES published here outline a daunting task, one that might take a generation, even given adequate funding and support. So that we are not paralyzed by the enormity of the job, priorities are selected and outlined here. Some additional topics and controversies are also highlighted here. Selected goals are proposed and general summary points are offered. One difficulty that came up repeatedly during the conference, and shows in the work product, is the difficulty in distinguishing between errors and adverse events. Sometimes actions become ‘‘errors’’ only in retrospect. An action might appear entirely reasonable, but when an adverse outcome ensues, it is later identified as an error. For example, is all medical care that is proximally related to a death subject to criticism? After all, everyone will eventually die, and most deaths will occur in some proximity to medical care. To get beyond nonproductive debates would be to concentrate on improving patient safety by reducing adverse outcomes, and then worry later about to what extent they are due to error. This is elaborated upon in this issue of AEM. The work group on defining and measuring error has outlined an ambitious scope. In the short term, it would be reasonable to set the following goals: 1. Accept the Institute of Medicine report definition for case finding and identification. This promotes consistency, but does not preclude the parallel use of another definition if circumstances warrant. 2. Each institution should develop an investigative function, empowered to make system-wide im-