Diagnostic accuracy study assessing the ability of paediatric asthma scores to predict admission following initial emergency department bronchodilator therapy: a Clinical Asthma Scoring systems in Paediatric Emergency (CASPER) study
Charmaine Gray, Leah Armit, Franz E Babl, Stuart Dalziel, Meredith Louise Borland, Sharon O’Brien, Doris Tham, Mohammad Asghari-Jafarabadi, Simon CraigBackground
The ability of paediatric asthma scores to predict admission following initial bronchodilator therapy has not been comprehensively evaluated.
Methods
A multicentre observational study including children aged 2–18 years who received at least a single dose of bronchodilator following presentation to four Australian emergency departments (EDs). Observations were obtained by trained research nurses before and after initial treatment to calculate ten common asthma scores (Pediatric Respiratory Assessment Measure (PRAM), Siriraj Clinical Asthma Score (SiCAS), Clinical Asthma Score (CAS), Acute Asthma Intensity Research Score (AAIRS), Modified Pulmonary Index Score (MPIS), Pulmonary Index Score (PIS), Pediatric Asthma Score (PAS), Pediatric Asthma Severity Score (PASS), Woods and Downs and Pulmonary Score (WDCAS)). Primary outcome was determining the global ability of scores to discriminate between those admitted and discharged, assessed by an area under receiver operating curve (AUC ROC) ≥0.80. Secondary outcomes included adjusted ORs and the ability to provide an adequate balance between sensitivity and specificity (defined as Youden Index ≥50%).
Results
Of the 1238 patients enrolled, 792 (64.0%) were male, median age was 3 years (IQR 2–6), and 513 (41.4%) met the admission criteria. None of the asthma scores demonstrated an AUC ROC ≥0.80, with the four best performing (MPIS, SCAS, PIS and PAS) scoring between 0. 70 and 0.73. Results for adjusted OR analysis indicated that each one-point increase was associated with increased odds of admission for all 10 asthma scores (p=<0.001). No asthma score achieved a Youden Index ≥50%.
Conclusion
None of the asthma scores assessed in this study met the predetermined primary outcome of an AUC ROC ≥0.80.