Comparative effectiveness of physician diagnosis and guideline definitions in identifying sepsis patients in the emergency department

PURPOSE: The purpose of our study was to compare the agreement of emergency physician diagnoses relative to the 1991 American College of Chest Physicians (ACCP)/Society of Critical Care Medicine (SCCM) and 2001 ACCP/SCCM/European Society of Intensive Care Medicine/American Thoracic Society/Surgical Infection Society internationally accepted definitions of sepsis, severe sepsis, and septic shock.

MATERIALS AND METHODS: This study was an observational cohort study of adult patients presenting to the emergency department (ED) with a chief complaint suggestive of infection over a 6-week period, during a daily enrollment schedule from 7:00 am to 10:00 pm. Patients were categorized as having "no sepsis," "sepsis," "severe sepsis," or "septic shock" based on ED physician diagnosis, the 1991 definitions, or 2001 definitions. Agreement statistics were performed.

RESULTS: A total of 1275 patients were enrolled with age 50.1 ± 21.7 years and 59.1% were female. Among the enrolled patients, 228 were identified as having a source of infection. Temperature, heart rate, and white blood cell count were significantly higher in patients with infection, compared with those without (P < .001). The odds ratio for disagreement between a physician-designated no sepsis diagnosis and the 1991 definitions was 4.47 (95% confidence interval, 3.01-7.53) and 5.96 (3.78-9.46) between the same physician-designated diagnosis and the 2001 definitions. The odds ratios for disagreement of a severe sepsis physician diagnosis in relation to the 1991 and 2001 definitions were 0.06 (0.01-0.19) and 0.06 (0.01-0.20), respectively. The 1991 and 2001 consensus definitions had strong agreement, with κ = 0.86 and 91.2% agreement. No agreement was found between the physician diagnosis and 1991 consensus sepsis definitions (κ = 0.11 and 52.2% agreement) or between the physician diagnosis and the 2001 consensus sepsis definitions (κ = 0.13 and 50.0% agreement).

CONCLUSIONS: Our study showed that ED physician diagnosis of sepsis may disagree with the international definitions such that severe sepsis is underrecognized by clinical judgment alone. Although these results are limited to a single center, we raise concern that early treatments for these high-risk patients may be delayed due to inaccurate clinical diagnosis. Efforts are needed to increase the application of sepsis guideline definitions to better identify ED patients with this potentially deadly condition.