We evaluated hospital readmission as an indicator of the quality of management of asthma patients, between July 1989 and June 1990. Using hospital separation data, we constructed a matched data set to identify early (within two weeks of discharge) readmissions. Of over 14,000 admissions for asthma in the 1-to-44-year age group, 2.8 per cent were classified as early readmissions. Admissions and readmissions were more common in rural than metropolitan areas. Admissions were most common during autumn, but early readmissions occurred most often during spring. Patients staying more than one day were 0.5 times (95 per cent confidence interval (CI) 0.37 to 0.68) as likely to have an early readmission than patients staying less than one day. Using the same data set, we identified patients who had the potential for readmission within a six-month period. Of the 5052 patients, 17.8 per cent were readmitted at least once during the period; 3.7 per cent had at least one early readmission, and 15.8 per cent had at least one late readmission (more than two weeks following discharge). A length of stay of more than one day was associated with 0.41 times (CI 0.24 to 0.70) the risk of early readmission in this cohort. A length of stay of more than one day was associated with a higher risk of late readmission (1.52, CI 1.09 to 2.12), which was less likely to occur in rural than metropolitan areas (0.45, CI 0.37 to 0.55).(ABSTRACT TRUNCATED AT 250 WORDS)