This paper reports on data from a double-blind, randomized controlled study of out-patient use of corticosteroids following an acute asthma attack. Issues related to compliance are examined, including: (1) the extent of non-compliance; (2) impact of non-compliance on interpreting the drug trial results; and (3) correlates of non-compliance. Of the 102 cases enrolled in the study, 25.5% were excluded from analysis because they were lost to follow-up (10.8%) or non-compliers (14.7%). Based on data for compliers, the drugs were found to reduce relapse rates and asthma symptomatology; when non-compliers were included in the analysis, the steroid drug appeared ineffective for reducing relapses and less effective for improving overall illness status. Examination of 24 potential correlates of compliance yielded a few significant associations, and only the "usual habit of compliance" correlation suggests an avenue for future action. The implications of the study findings for design and interpretation of clinical trials, as well as for improved management of chronic diseases, are discussed.