Severity scoring systems: are they internally valid, reliable and predictive of oxygen use in children with acute bronchiolitis?

BACKGROUND: Severity scores are commonly used in research and clinically to assess the severity of bronchiolitis. However, there are limitations as few have been validated. The aim of our study was to: (i) determine the validity and reliability of a bronchiolitis scoring system, and (ii) examine if the score predicted the need for oxygen at 12 and 24 hrs. Children aged <24 months presenting to Royal Darwin Hospital with a clinical diagnosis of bronchiolitis were eligible to participate. STUDY
DESIGN: We reviewed published papers that used a bronchiolitis score and summarized the data in a table. We chose the Tal score that was easy to use and encompassed clinically important parameters. Three research nurses, trained to assess children, used two scoring systems (Tal and Modified-Tal; respiratory rate, accessory muscle use, wheezing, cyanosis, and oxygen saturation), blindly evaluated children within 15 min of each other.
RESULTS: The children's (n = 115) median age was 5.4 months (IQR 2.9, 10.4); 65% were male and 64% were Indigenous. Internal consistency was excellent (Tal: Cronbach α = 0.66; Modified-Tal: α = 0.70). There was substantial inter-rater agreement; weighted kappa of 0.72 (95% CI: 0.63, 0.83) for Tal and 0.70 (95% CI: 0.63, 0.76) for Modified-Tal. For predicting requirement for oxygen at 12 and 24 hrs; area under receiver operating curve (aROC) was 0.69 (95% CI: 0.13, 1.0) and 0.75 (95% CI: 0.34, 1.0), respectively.
CONCLUSION: The Tal and Modified-Tal scoring systems for bronchiolitis is repeatable and can reliably be used in research and clinical practice. Its utility for prediction of O2 requirement is limited.