The goal of this 2-part study was to quantify the frequency of allergy warnings that were overridden, specify the reasons why a clinician chose to override an alert, and evaluate allergy warnings that were accepted by prescribers.
This was a retrospective analysis of computerized prescriber order entry decision support allergy warnings and the corresponding prescriber decisions during four 1-week intervals and another 16-week interval from August 2003 to February 2005 at a 993-bed tertiary care teaching hospital. The following information was collected: patient allergy status, allergy category, allergy severity, warning provided, clinician category, and the reason warning was overridden.
A total of 29,483 evaluated orders triggered 777 allergy warnings, of which, 44% of the warnings were accepted, and 56% of the warnings were overridden. Of the 435 warnings that were overridden by the ordering clinician, 6% of the warnings were drug specific, and 94% were drug class warnings. The primary reason selected for overriding an allergy warning was that the patient tolerated the medication in the past (70%). Of the 342 warnings that were accepted by ordering clinicians, 24% of the warnings were drug specific, and 76% were drug class warnings.
Clinicians were much less likely to override a warning if it was an exact drug match compared with a drug class warning. Limiting the number of alerts for less significant drug class warnings would help to improve the positive predictive value of computerized prescriber order entry decision support for allergy warnings.
From the *Division of Pharmacy Practice and Administration, College of Pharmacy, The Ohio State University, and †Department of Pharmacy, The Ohio State University Medical Center, Columbus, Ohio.
Correspondence: Craig A. Pedersen, RPh, PhD, Division of Pharmacy Practice and Administration, College of Pharmacy, 500 West 12th Ave, Columbus, OH 43210-1291 (e-mail: firstname.lastname@example.org).