Studies concerning the value of therapeutic drug monitoring programs for theophylline have yielded contradictory findings. This study employed a prospective, single-blind, randomized, crossover design to evaluate the effect of a therapeutic drug monitoring program for theophylline. There were 410 serum theophylline assays performed in 169 patients. The average number of assays drawn incorrectly and the average number of assays used inappropriately were significantly reduced (from 1.87 to 1.32 and from 2.07 to 1.45, respectively) with therapeutic drug monitoring. When the program was withdrawn, no statistically significant change resulted, and when the program was reinstituted, there was improvement in audit criteria performance but the change was not statistically significant. Although the savings from the program would equal the cost in the first year, cost savings would be anticipated in the future.