In children with typical respiratory symptoms, the assessment of limitation in expiratory airflow is considered an essential component to support the clinical diagnosis of asthma. In this respect, a greater than or equal to 12% increase in postbronchodilator FEV1 has been proposed in international guidelines as evidence of airway reversibility. However, the validity of the 12% cutoff has been questioned in children. The aim of this study was to assess whether a thorough statistical model taking into account bronchodilator response (BDR) cutoff values and some child characteristics is associated with better performance than the traditional use of BDR cutoff values in supporting the clinical diagnosis through discrimination of outpatient children with steroid-naive asthma from those without asthma. More specifically, we propose to use a regression model with child covariates (eg, age, weight, height, and atopy) as usual linear terms and a piecewise linear relationship on BDR in which cutoff values are properly estimated from the data.