Early predictors of admission or prolonged emergency department treatment for children with acute asthma

Objective. To determine whether demographic, historical and clinical information available at the time of presentation to the pediatric emergency department (ED) can be used to predict which children with acute asthma are likely to require extended treatment (>5 hours in ED or hospital admission). Design. Concurrent cohort study. Setting and sample. Inner-city, university-based pediatric ED. Subjects were 1-18 years old (n = 181) receiving standardized asthma therapy with frequent beta-agonists and corticosteroids. Measurement. Upon ED presentation, demographic information, asthma history and seven clinical variables were assessed. Bivariate analysis and multivariable logistic regression were used to identify significant predictors of extended treatment. Positive predictive values (PPVs) for individual and combined variables were calculated. Results. Overall, 30% (54 of 181) subjects required extended treatment; 8% (15 of 181) required treatment in the ED > 5 hours and 22% (39 of 181) were admitted. All but one of the asthma severity score items were significantly associated with prolonged treatment (suprasternal in drawing, P=0.07; all others, P < 0.05). When these items were combined into the asthma scores from which they were originally derived, PPVs for extended treatment were only 45 and 50%, respectively. These PPVs for extended treatment were no better than those for individual items, which ranged between 36 and 50%. Demographic information and prior asthma history were not associated with extended treatment. Conclusions/implications for practice. Although individual asthma severity score items and asthma severity scores assessed at ED presentation were associated with extended treatment, no variable, alone or in combination, had a clinically useful PPV. Decisions regarding observation unit admission for pediatric asthmatics should not solely be based on initial clinical assessment.