Physician Documentation of Sepsis Syndrome Is Associated with More Aggressive Treatment

INTRODUCTION Timely recognition and treatment of sepsis improves survival. The objective is to examine the association between recognition of sepsis and timeliness of treatments. METHODS We identified a retrospective cohort of emergency department (ED) patients with positive blood cultures from May 2007 to January 2009, and reviewed vital signs, imaging, laboratory data, and physician/nursing charts. Patients who met systemic inflammatory response syndrome (SIRS) criteria and had evidence of infection available to the treating clinician at the time of the encounter were classified as having sepsis. Patients were dichotomized as RECOGNIZED if sepsis was explicitly articulated in the patient record or if a sepsis order set was launched, or as UNRECOGNIZED if neither of these two criteria were met. We used median regression to compare time to antibiotic administration and total volume of fluid resuscitation between groups, controlling for age, sex, and sepsis severity. RESULTS SIRS criteria were present in 228/315 (72.4%) cases. Our record review identified sepsis syndromes in 214 (67.9%) cases of which 118 (55.1%) had sepsis, 64 (29.9%) had severe sepsis, and 32 (15.0%) had septic shock. The treating team contemplated sepsis (RECOGNIZED) in 123 (57.6%) patients. Compared to the UNRECOGNIZED group, the RECOGNIZED group had a higher use of antibiotics in the ED (91.9 vs.75.8%, p=0.002), more patients aged 60 years or older (56.9 vs. 33.0%, p=0.001), and more severe cases (septic shock: 18.7 vs. 9.9%, severe sepsis: 39.0 vs.17.6%, sepsis: 42.3 vs.72.5%; p<0.001). The median time to antibiotic (minutes) was lower in the RECOGNIZED (142) versus UNRECOGNIZED (229) group, with an adjusted median difference of -74 minutes (95% CI [-128 to -19]). The median total volume of fluid resuscitation (mL) was higher in the RECOGNIZED (1,600 mL) compared to the UNRECOGNIZED (1,000 mL) group. However, the adjusted median difference was not statistically significant: 262 mL (95% CI [ -171 to 694 mL]). CONCLUSION Patients whose emergency physicians articulated sepsis syndrome in their documentation or who launched the sepsis order set received antibiotics sooner and received more total volume of fluid. Age <60 and absence of fever are factors associated with lack of recognition of sepsis cases.