Because asthmatics have the highest utilization rate (11%) kin our emergency department (ED) observation unit (OU), we conducted a study correlating predictors of the need for OU therapy to initial disposition (ID) and final disposition (FD) using chart audit of treated asthmatics. Twenty-four clinical variables. (historical, physiological, laboratory, therapy response) were examined utilizing chi-square and Student's t tests. Forty-six asthmatics were treated during a four-month period in 1980. The ID breakdown was as follows: 1) home, 17; 2) OU, 23; and 3) admit, 6. Twenty-seven (59%) of the patients received treatment in the OU at some point in their attack (initial or rebound); 18 (39%) were definitively treated in the ED, and nine (20%) were admitted. The mean OU stay was 19 hours at a cost that was 34% of that incurred for a hospital admission. The FD differed from the ID in 14 of 46 (30%): 1) home, 12; 2) holding, observation, and short-term therapy, 18; and 3) admit, 16. Clinical variables correlating significantly with definitive therapy based on ID and FD were historical; symptoms greater than 24 hours, prior OU admissions, and prior hospitalizations. We conclude that the OU is appropriate, safe, and less expensive than admission; is not used for procrastination in decision making and decreases the hospitalization rate. Historical data correlated significantly with both ID and FD, while clinical variables were of little predictive value.