A Systems Approach to Reducing Medical Error

Abstract

To the Editor I was pleased to read the recent Viewpoint 1 highlighting the serious problem of diagnostic error-a crucial issue not just in emergency departments. I applaud the recommendations to move the focus away from individual physician failings and toward discussion of underlying causes.However, although some systems-based solutions can help reduce diagnostic error, they may also cause unintended consequences. For example, computer-based algorithms used for the diagnosis of sepsis are a widespread systems-based effort designed to improve the timely diagnosis of a particular condition that has arguably facilitated widespread misdiagnosis of others. 2,3 Certainly, diagnostic algorithms and computer-generated alert systems have raised clinician awareness of sepsis and in many cases improved care. However, I also believe they can inadvertently promote cognitive biases such as diagnostic anchoring and premature closure, thereby preempting consideration of alternate diagnoses. Proof of the negative impact of these diagnostic aids may be hard to come by, but many have expressed this concern. 4,5 Furthermore, these diagnostic aids may pose a particular concern about their effect on the development of medical trainees' clinical reasoning skills.Systems-based approaches to reducing diagnostic errorsuch as computer-based decision supports-have a lot of promise. However, if the experience with sepsis is any lesson, diagnostic reasoning remains at its core a human endeavor. And physicians and administrators need to remain vigilant to the potential that systems-based solutions to diagnostic error may produce unpredicted outcomes.