Therefore, based on the presented data, 126 of 817 SIRS alerts (15 percent) provided accurate, potentially valuable information. Unfortunately, another 80 patients in the hospitalized cohort received discharge diagnoses of sepsis despite never triggering the tool. Finally, these data only describe patients requiring hospitalization and not those discharged from the emergency department. We can only speculate regarding the number of alerts triggered on the diverse ED population not requiring hospitalization, as prior work has estimated infection constitutes no more than a quarter of patients with SIRS in the emergency department.³ If such estimates hold true, the potential utility of such a SIRS-based tool drops below 5 percent.

The lead author proudly concludes their tool is “an effective approach toward early recognition of sepsis in a hospital setting.” Of course, the author, employed by Cerner, also declares he has no potential conflicts of interest regarding the publication in question.

Now, for a serious condition such as severe sepsis, it is reasonable to accept some limitations in specificity in the interests of maximal sensitivity. But the Cerner tool is simply an early warning tool for sepsis, the less critically important cousin of severe sepsis. Fewer than half of the alerts produced by the Cerner tool were for instances detecting severe sepsis, the condition targeted for immediate interventions pioneered by early goal-directed therapy and its cousins. These are the cases for which we suffer alert fatigue, and yet, as a recent New England Journal of Medicine (NEJM) article further describes, these criteria will still miss one in eight patients with significant mortality risk.⁴

These authors of the NEJM article retrospectively reviewed 13 years of patient data from 172 intensive care units in New Zealand and Australia. After identifying those patients with organ failure (eg, severe sepsis), they categorized them by the number of manifested SIRS criteria. They found that even in ICU patients with an infection and organ dysfunction, nearly 12 percent met zero or one SIRS criteria. That is to say, even in a critically ill population of which 42 percent exhibited septic shock, a SIRS-based tool, such as the Cerner instrument, would not be triggered one out of eight times.

Even worse for the SIRS criteria, these authors found the traditional threshold of two or more conferred minimal prognostic value. Patients with zero or one SIRS criteria had a combined 16.1 percent risk of in-hospital death. Those with SIRS-positive sepsis had 24.5 percent in-hospital mortality, but the adjusted odds of death were statistically no different between those with zero or one SIRS criteria and those with two or three SIRS criteria.