Pediatric Sepsis ID Improves With Electronic Alert

Veronica Hackethal, MD

June 20, 2017

An electronic alert decreased missed cases of sepsis in a pediatric emergency department (ED), but did not completely replace frequent assessments by physicians, a new study at Children's Hospital of Philadelphia has found. The researchers published their findings online June 2 in the Annals of Emergency Medicine.

The study is the first prospective study to evaluate whether an electronic sepsis alert can improve sepsis identification in the pediatric ED.

"The electronic sepsis alert required bedside evaluation of just 1% of the ED census and had 85% sensitivity and 99% specificity. Sensitivity was optimized to 99% when clinician identification was used to augment sepsis recognition of electronic sepsis alert–negative patients," Fran Balamuth, MD, PhD, from the University of Pennsylvania Perelman School of Medicine, Philadelphia, and colleagues write. "After implementation of the electronic sepsis alert, we decreased the proportion of missed ED cases from 17% preintervention to 4% postintervention."

The alert consisted of two stages: Tachycardia or hypotension at any time during the ED visit triggered the first stage, and the second stage was triggered by hypotension or tachycardia, fever or risk for infection, and one or more of the following: a high-risk condition, altered mental status, or abnormal capillary refill time.

Studies have suggested that bedside clinical evaluation misses about one quarter of pediatric sepsis cases, because most children initially compensate for shock with elevated pulse. However, children without sepsis can also have elevated pulse as a result of fever, anxiety, and pain.

Patients with positive first- and second-stage alerts were considered to have a positive electronic sepsis alert. That prompted a team assessment, or huddle, to decide whether to activate the sepsis protocol and order set.

During the intervention, 1.2% of ED visits (1112/182,509) triggered sepsis alerts; of those, 23.8% (n = 265) of patients received treatment with the sepsis protocol after the team huddle. Clinicians retained decision-making power to activate the sepsis protocol at any time in patients with negative sepsis alerts.

The electronic sepsis alert correctly identified true cases of sepsis (sensitivity) 86.2% of the time (95% confidence interval [CI], 82.0% - 89.5%) and correctly weeded out children without sepsis (specificity) 99.1% of the time (95% CI, 99.0% - 99.2%). The probability that patients with positive alerts had sepsis was 25.4% (95% CI, 22.8% - 28.0%). The probability that patients with a negative alert did not have sepsis was 100% (95% CI, 99.9% - 100%).

Including physician-identified sepsis cases in the analysis improved the sensitivity to 99.4% (95% CI, 97.8% - 99.8%), although the specificity remained at 99.1% (95% CI, 99.1% - 99.2%).

Use of the electronic sepsis alert increased sepsis identification in the ED from 83% to 96%.

"Although our overall sepsis recognition improved after alert implementation, clinician identification remains an important modality for recognition of patients with severe sepsis and identified 43 patients who were electronic sepsis alert negative," the researchers conclude. "This underscores the point that a vital sign–based sepsis screen is not sufficient to fully capture all patients with severe sepsis and that, currently, a component of clinician identification remains critically important."

In a linked editorial, Andrea Cruz, MD, MPH, from Baylor College of Medicine, Houston, Texas, points out several limitations of the alert, including a low positive predictive value indicating that about 25% of the children with triggered alerts had sepsis.

The authors acknowledge the relatively low positive predictive value, but explain that it is much higher than those found in past pediatric sepsis screening studies, which have generally run between 2.5% and 4%. Given the high financial and personal cost of missing one case of sepsis, a "hit rate" of 1:4 may be "reasonable" to prompt rapid clinical assessment, but may not be enough to trigger treatment, they say.

Dr Cruz also writes the alert missed 13.1% of children with severe sepsis, who were only identified after clinical assessment.

"Ultimately, the value of an electronic sepsis alert may be in increasing a clinical team's index of suspicion for sepsis and marshaling resources for treatment, rather than in identification of the child with severe sepsis per se," she concludes.

The study has several limitations. It took place at a single academic medical center, and other settings may show different results. The researchers selected vital sign cutoffs from norms within this one institution, and results may differ with other cutoffs. The study could not distinguish between patients who did not develop sepsis because they received appropriate care and those who would have improved without treatment and who may have been overtreated. However, sensitivity analyses accounting for this issue suggested the alert did not result in overtreatment.

The intervention also included ongoing educational efforts and feedback from ED providers.

Sepsis is a medical emergency that results from a systemic inflammatory response to infection. More than 75,000 children undergo treatment for severe sepsis in the United States each year, and up to 20% of them die from the condition, the authors write.

Three prior years of quality improvement efforts failed to improve sepsis recognition; therefore, the results suggest the electronic alert most likely was the cause of improved sepsis recognition, the researchers explain. However, the study could not rule out other interventions or actions that may have affected results.

The authors and Dr Cruz have disclosed no relevant financial relationships.

Ann Emerg Med. Published online June 2, 2017. Article extract, Editorial extract

For more news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.