Can paramedics accurately diagnose sepsis and severe sepsis in the field [abstract]? Abstract uri icon


  • Background: Morbidity and mortality from sepsis and severe sepsis is a significant problem, with early recognition and treatment as areas of focus. Little work has been done to examine the ability of prehospital providers to identify sepsis and severe sepsis.
    Objectives: Determine the accuracy of paramedic diagnosis of sepsis and severe sepsis using a diagnostic screening tool and point of care (POC) lactate measurement.
    Methods: Following IRB approval, consented paramedics from a single, advanced life support (ALS) dual-role emergency medical services (EMS) agency underwent a one-hour training session on the universal criteria for sepsis diagnosis and the use of a POC lactate meter. Paramedics then screened every patient transported from three local nursing homes and assigned them a diagnosis of no sepsis, sepsis, or severe sepsis. The screening tool defined sepsis as an identified or suspected infection plus two of the following: temperature < 96.8 or >100.4F, heart rate > 90/min, respiratory rate > 20/min, or altered mental status. Severe sepsis was defined as meeting the sepsis criteria, as well as one of the following: mottled skin, capillary refill ‡ 3 seconds, lactate > 2mmol/L, or abrupt changes in mental status. Patient charts were abstracted and a blinded physician assigned a final diagnosis to each patient. Sensitivity, specificity, negative and positive predictive values, and Cohen’s kappa were used to evaluate the agreement between the paramedic and physician diagnosis.
    Results: One hundred and fifty-one patients were screened between October 2009 and February 2010, and 96 cases (64%) had both a paramedic and independent physician diagnosis assigned. Sensitivity, specificity, negative predictive value, and positive predictive value for sepsis diagnosis were 0.33 (95% CI 0.18, 0.53), 0.89 (95% CI 0.80, 0.94), 0.8 (95% CI 0.70, 0.87), and 0.50 (95% CI 0.28, 0.72). Sensitivity, specificity, negative predictive value, and positive predictive value for severe sepsis diagnosis were 0.20 (95% CI 0.05, 0.51), 0.94 (95% CI 0.87, 0.97), 0.91 (95% CI 0.83, 0.95), and 0.29 (95% CI 0.08, 0.64). The level of agreement between paramedic and physician diagnosis of sepsis and severe sepsis was low (kappa = 0.25 and 0.16, respectively).
    Conclusion: Paramedics were able to accurately identify patients not experiencing sepsis or severe sepsis, but were not able to identify those with sepsis or severe sepsis with high sensitivity.

publication date

  • 2011