As the summer winds down, respiratory season will soon be on the doorsteps of pediatric offices, urgent care clinics, and emergency departments (ED). Leading the way will be bronchiolitis, a respiratory illness leading to nearly 120,000 emergency department visits annually.1 In this month’s Pediatrics, Freire et. al. (10.1542/peds.2017-4253) study whether or not we can reliably predict which patients with bronchiolitis will require escalation of or intensive care at the time of ED presentation.
This multi-national collaboration performed a planned, retrospective secondary analysis of a cohort from the Pediatric Emergency Research Networks (PERN). Using 12 predictors previously associated with bronchiolitis severity, study authors aimed to determine which of these predictors, if present at the time of evaluation in the ED, would be associated with a greater odds of requiring escalated care, including high flow nasal cannula (HFNC), continuous or biphasic airway pressure, and admission to a pediatric intensive care unit. A clinical risk tool from these results was created by assigning weighted values to each predictor based on the magnitude of the individual odds ratio. The diagnostic ability of the score to predict the need for escalation of care was determined by calculating the area under the curve on a receiver operating characteristic curve.
The study included over 2,700 infants less than 12 months of age presenting to one of 38 EDs in the final analysis. Nearly 10% (n=261) of the infants received escalated care, the majority (63%) of which were treated with HFNC. Of the 12 predictors studied, only three were not associated with an escalation in care: sex, prematurity, and duration of distress. The remaining nine had varying odds ratios with oxygen saturation <90% (OR 8.98; CI 5.08-15.66), history of or observed apnea (OR 3.01; CI 1.89-4.78), nasal flaring/grunting (OR 3.76; CI 2.64-5.35), and retractions (OR 3.02; CI 1.59-5.73) being the most significant. Incorporating the predictors into a weighted total risk score resulted in an AUC of 84.7% (CI 81.7% to 86.8%) for predicting a need for escalation of care. Study authors validated the risk tool using a bootstrapping statistical technique, where study participants were randomly re-sampled from the original pool and subjected to the predictive risk tool. Analyses were subsequently performed on each re-sampling to test for significant association between the total risk score and escalation of care, resulting in a corrected average AUC of 84.2%. While no prospective studies evaluating the clinical utility of this tool have been performed, the results are nonetheless encouraging. Interested? Check out Pediatrics this month and learn more about how we might be able to better predict which infants with bronchiolitis will require more than reassurance and observation!
Reference
1. Hasegawa K, Tsugawa Y, Brown DFM, Mansbach JM, Camargo CA. Trends in Bronchiolitis Hospitalizations in the United States, 2000-2009. Pediatrics. 2013;132(1):28-36. doi:10.1542/peds.2012-3877
