The degree of procedural compliance and the clinical outcome associated with a therapeutic interchange of extended-spectrum penicillins were assessed. Pharmacists conducted a concurrent chart review on all patients receiving mezlocillin as a therapeutic alternative for piperacillin or ticarcillin over a five-month period. The pharmacists assessed whether (1) the pharmacy appropriately dispensed mezlocillin when orders were written for piperacillin or ticarcillin, (2) physicians and nurses were properly notified of the therapeutic interchange, and (3) the bacterial isolates were susceptible to mezlocillin. Pharmacists and physicians evaluated clinical outcomes of all patients receiving mezlocillin through a retrospective chart review and classified the responses as "success," "failure," or "indeterminate." Fifty-one orders for piperacillin were written during the study period, and mezlocillin was selected as a therapeutic alternative in all cases. No orders for ticarcillin were written. Proper notification was made to nurses and physicians in 98% and 65% of cases, respectively. No mezlocillin-resistant gram-negative bacteria were found. Eighty-three courses of therapy were reviewed for clinical outcome; 68 were classified as successes, 0 were failures, and 15 were indeterminate. The cost savings after nine months of therapeutic interchange of mezlocillin for piperacillin was $6600. This study indicates that the pharmacy made a proper therapeutic interchange of mezlocillin for piperacillin and followed the correct procedure for notification of the nursing staff. However, more effort is needed to ensure communication with the prescribing physicians.