Unmeasured confounders and selection bias can significantly influence the results of retrospective observational analyses of asthma therapy. To evaluate the efficacy of oral montelukast and inhaled fluticasone propionate in a randomized, prospective 12-month "real-world" observational analysis of children with mild persistent asthma. Children (n = 104) between 6 and 15 years of age with mild asthma as determined by forced expiratory volume in 1 second, symptoms, and evaluation by an experienced pediatric allergist or pulmonologist, who were not currently receiving controller therapy, were randomly assigned to fluticasone or montelukast on an alternating basis. Subjects were asked to complete a questionnaire at 6 and 12 months; otherwise, medical care was identical to that of similar managed care patients. Outcome parameters were evaluated after 12 months by claims database analysis. An acute asthma attack requiring emergent care was the primary outcome parameter. Measures of adherence, symptoms, and asthma control, as measured by the pediatric Asthma Therapy Assessment Questionnaire, were secondary outcome parameters. Demographics, spirometry, symptoms at enrollment, emergent care visits, asthma hospitalizations, routine office visits, and symptoms at study completion were not significantly different between study groups. Adherence, as evaluated by the number of controller fills, was significantly (P = 0.0003) better for montelukast (7.65 +/- 3.01) than fluticasone (5.46 +/- 3.01). Similar numbers of subjects in each study group required beta-agonists and oral prednisone. These results suggest that oral montelukast and inhaled fluticasone have similar real-world efficacies in the treatment of children with mild asthma, possibly as a result of the significantly better adherence with oral montelukast therapy compared with inhaled fluticasone.