Interrater Reliability of the Pediatric Asthma Score.

Abstract

The aim of this study was to determine the interrater reliability (IRR) of the Pediatric Asthma Score (PAS) and to evaluate the discriminative performance of this score to predict the need for hospital admission among children with acute asthma. A secondary analysis of prospective data was performed to compare triage nurse and study personnel PAS scores among children aged 6 to 18 years presenting to the emergency department with acute asthma. The IRR was determined by calculation of weighted Cohen κ with differences evaluated by Wilcoxon ranked pairs. Receiver operating characteristic curves were created to evaluate the predictive ability of PAS to determine the need for hospital admission. One hundred one subjects were evaluated by both study personnel and a triage nurse with PAS score recorded. The IRR of the total PAS score was determined to be moderate (κ = 0.57) and acceptable, although lower than previously reported. Individual components of the PAS score demonstrated fair to substantial agreement. Receiver operating characteristic analysis demonstrated total PAS at emergency department triage to have poor test characteristics in predicting the need for hospital admission, whether PAS was determined by study personnel, triage nurse, or an average score (area under the curve, 0.62-0.65). In this study, total PAS score demonstrated a moderate and acceptable level of IRR with a poor discriminative ability to determine the need for hospital admission at the time of ED triage.