This study was undertaken to determine if a computerized pharmacokinetic program for adjusting theophylline infusion rates could attain a goal serum theophylline level more accurately than physician-derived adjustments and what clinical impact this would have. Thirty-five patients with diagnoses of asthma or chronic obstructive pulmonary disease were randomized to a control group (empiric) or experimental group (kinetic) after initial theophylline levels were drawn from each group. After second levels were drawn, patients in the kinetic group had their infusion rates adjusted by the computerized pharmacokinetic program to achieve a level of 15 mg/L, whereas patients in the empiric group had their infusions adjusted empirically by the primary care physicians to achieve a serum theophylline level of 15 mg/L. A final theophylline level was obtained just before the infusion was discontinued. The kinetic group was closer to the goal level of 15 mg/L than the empiric group, but this was not statistically significant (14.8 +/- 4.4 versus 12.6 +/- 4.1; p > 0.05). The total number of days that patients were receiving intravenous theophylline was slightly longer for the kinetic group (4.1 +/- 3.3 versus 3.2 +/- 1.5; p > 0.05) as was the total number of hospital days, but neither of these were statistically significant (11.4 +/- 21.6 versus 8.8 +/- 15.4 days; p > 0.05). There were no differences between the two groups in the number of subtherapeutic or toxic levels, and there were no significant differences in arterial blood gas measurements.(ABSTRACT TRUNCATED AT 250 WORDS)