A great comment from Dr. Cory Franklin:
As an ICU physician and one of the people who drafted the initial “Sepsis guidelines” in 1991-1992, I want to point out they were not drafted with the intent of clinical use, rather as study criteria for the evaluation of sepsis drugs, in that case Xigris (unsuccessful).
i was one of the people who said that they should not be used as clinical criteria, since the whole concept of sepsis is ill-defined and difficult to validate. This fell on deaf ears and within a short period of time I saw ER diagnoses of R/o SIRS. IT soon became a routine clinical diagnosis, even though it was not originally meant to be.
Once started there is nothing to be done. unfortunately, the concept of defining sepsis has taken hold. I wrote about it in Thorax in the mid 1990’s I believe) and I despaired of a definition, preferring to opt for teaching people what to look for and how to evaluate possible infection.
(In fact, the term “sepsis” is rather new. Throughout the first two thirds of the 20th Century the terms used were “septicemia” – presumably a positive blood culture with clinical signs of infection, or “antisepsis” – ways to sterilize an environment). Sepsis came from the Greek for “putrid” – the way many infections were diagnosed in the late 19th century. It was not a particularly sensitive or specific method but it was pretty much all they had for things like puerperal infections).
It is predictable that this approach will result in much of what the IDSA fears. But they are not without some culpability – since the ID doctors have not devoted a lot of attention to teaching non-Id doctors about the diagnosis of infection in different arenas.
What goes around comes around. AndI have no idea what to do about it.
This comment frames the issue in an important way. Many hospitals have become “sepsis obsessed”. So we now are willing to overdiagnosis a poorly defined condition. The story reminds me of the community acquired pneumonia story. We used to diagnose pneumonia; now we overdiagnosis CAP.
This story is a diagnostic nightmare. We have code sepsis that leads to automatic blood draws (for lactate) and emergency decision making. Sometimes the patient does have a serious infection, but too often “code sepsis” leads to a diagnosis of sepsis, which (as Dr. Franklin makes clear) is not really a diagnosis. We need to carefully consider whether patients have infections and possibly septicemia.
We try to automate diagnosis and seemingly believe that clinical judgment is worthless. Performance measures and guidelines should not induce diagnostic errors, yet at least in the cases of sepsis and CAP they seem to lead to many false positives.
