Objectives: To simultaneously examine adherence to long-term controller and quick-relief medications and to contrast patterns of medication use in children with asthma.Study design: Cross-sectional, 1-month follow-up study conducted with 75 children ages 8 to 16 years diagnosedwith persistent asthma and prescribed quick-relief and long-term controller medications by metered dose inhaler.Participants were a subsample of a larger adherence study.The primary outcome measure was adherence to both medications as measured by electronic monitoring devices.A classification framework for contrasting adherence patterns between medication classes was developed to identify cases for individual analysis.Results: High levels of nonadherence to long-term controller medications (median = 46%of prescribed doses taken) and variable patterns of quick-relief medication use (range = 0 to 251 doses over the month) were documented,whereas consistent relationships between patterns of medication use across both classes were not found.Individual cases identified by the classification scheme illustrated the complexity and clinical utility of contrasting adherence patterns.Conclusions: Monitoring long-term controller medication adherence may be more predictive of morbidity than quick-relief medication use, except in outlier cases in which monitoring both medication types may be valuable for clinical and empirical purposes. Objectives: To simultaneously examine adherence to long-term controller and quick-relief medications and to contrast patterns of medication use in children with asthma. Study design: Cross-sectional, 1-month follow-up study conducted with 75 children ages 8 to 16 years diagnosed with persistent asthma and prescribed quick-relief and long-term controller medications by metered dose inhaler. Participants were a subsample of a larger adherence study. primary outcome measure was adherence to both medications as measured by electronic monitoring devices. A classification framework for contrasting adhere adherence patterns between individuals who have been diagnosed for individual analysis. Results: High levels of nonadherence to long-term controller medications (median = 46% of prescribed doses taken) and variable patterns of quick- to 251 doses over the month) were documented, where consistent relationships between patterns of patterns h classes were not found .Individual cases identified by the classification scheme illustrated the complexity and clinical utility of contrasting adherence patterns. Conclusions: Monitoring long-term controller medication adherence more likely predictive of morbidity than quick-relief medication use, except in outlier cases in which monitoring both medication types may be valuable for clinical and empirical purposes.