Quick SOFA Scores Predict Mortality in Adult Emergency Department Patients With and Without Suspected Infection

Study objective: The Quick Sequential Organ Failure Assessment (qSOFA) score (composed of respiratory rate >= 22 breaths/min, systolic blood pressure <= 100 mm Hg, and altered mental status) may identify patients with infection who are at risk of complications. We determined the association between qSOFA scores and outcomes in adult emergency department (ED) patients with and without suspected infection. Methods: We performed a single-site, retrospective review of adult ED patients between January 2014 and March 2015. Patients triaged to fast-track, dentistry, psychiatry, and labor and delivery were excluded. qSOFA scores were calculated with simultaneous vital signs and Modified Early Warning System scores. Patients receiving intravenous antibiotics were presumed to have suspected infection. Univariate and multivariate analyses were performed to explore the association between qSOFA scores and inpatient mortality, admission, and length of stay. Receiver operating characteristics curve analysis and c statistics were also calculated for ICU admission and mortality. Results: We included 22,530 patients. Mean age was 54 years (SD 21 years), 53% were women, 45% were admitted, and mortality rate was 1.6%. qSOFA scores were associated with mortality (0 [0.6%], 1 [2.8%], 2 [12.8%], and 3 [25.0%]), ICU admission (0 [5.1%], 1 [10.5%], 2 [20.8%], and 3 [27.4%]), and hospital length of stay (0 [123 hours], 1 [163 hours], 2 [225 hours], and 3 [237 hours]). Adjusted rates were also associated with qSOFA. The c statistics for mortality in patients with and without suspected infection were similarly high (0.75 [95% confidence interval 0.71 to 0.78) and 0.70 (95% confidence interval 0.65 to 0.74), respectively. Conclusion: qSOFA scores were associated with inpatient mortality, admission, ICU admission, and hospital length of stay in adult ED patients likely to be admitted both with and without suspected infection and may be useful in predicting outcomes.