Assessment of Medication Errors That Involved Drug Allergies at a University Hospital

T. Aaron Jones, Pharm.D., Jackson A. Como, Pharm.D.


Pharmacotherapy. 2003;23(7) 

In This Article


A total of 340 patients were reviewed for potential analysis. Of these, 133 patients (39%) reported allergies to one or more drugs. Seventy-three of these patients met the inclusion criteria, and an attempt was made for an interview. Owing to patient unavailability and time constraints, 23 patients were not interviewed. Demographics for the remaining 50 patients are reported in Table 3 . From the 50 patients interviewed, 70 allergies were assessed, as summarized in Table 4 . More than half of the assessed allergies (64.3%) listed in the computer system did not indicate a specific agent, but rather a class of drugs (i.e., 27 allergies listed as penicillin or PCN, 17 allergies listed as sulfa, and 1 allergy listed as cephalosporins); only 8 of these were further clarified on patient interview. Specific agents identified in the computer system were codeine (10 patients), sulfamethoxazole-trimethoprim preparations (5), hydrocodone (3), cephalexin (2), morphine (2), furosemide (1), hydromorphone (1), and oxycodone (1). Also, allergies to -lactams were reported in 43 (12.6%) of the 340 patients in the sample population, which is close to the 10% often reported in the literature. Allergies to sulfonamides and opioid narcotics were reported in 31 (9.1%) and 49 (14.4%) patients in the sample population, respectively. Other reported drug allergies included nonsteroidal anti-inflammatory drugs, skeletal muscle relaxants, anticonvulsants, contrast dyes, angiotensin-converting enzyme inhibitors, vancomycin, and benzodiazepines.

Of the reported allergies to -lactams that were assessed ( Table 5 ), most reactions occurred more than 10 years previously. Of these, 43% were determined to be true allergic reactions. Another 43% could not be ruled out as allergic reactions, but these reactions occurred so long ago they were labeled as indeterminate for purposes of clinical reactivity. Most of these patients described receiving a penicillin as a child; they had some type of unknown reaction and were told not to take the drug again or had received a penicillin around the time it was first introduced (1940s). In addition, two of the reported allergies to -lactams were determined to be anaphylactic, and these had occurred within the past 10 years.

Reported allergies to sulfonamides are summarized in Table 6 . Most reported allergic reactions were determined to be due to sulfonamide antibiotics, and when questioned about possible rechallenges, many patients reported taking other sulfonamide-containing drugs (i.e., diuretics and/or sulfonylureas) without similar reactions noted. Contrary to what is often found in the literature, most reactions were reported to have begun within 1-2 days of starting therapy.

As in other published reports, most documented allergies to opioid narcotics were determined to be due to intolerance of the drug, primarily nausea and vomiting ( Table 7 ). Pruritis was typically described as itching with a nonurticarial rash. When questioned as to whether or not the patient had received other opioid narcotics in the past, or on review of their drug profile, three of four patients had received another opioid without incident.

Overall, 70 medication errors were identified in which a drug was prescribed for a patient with a documented allergy to that drug or drug class. The results from contacting prescribers on 54 occasions are presented in Table 8 . Most agents involved were -lactam antibiotics, with an overwhelming number of these errors due to piperacillin-tazobactam (51.4%, 36 errors). Other drugs involved were ampicillin (10%, 7 errors), other -lactams (24.3%, 17 errors), opioid narcotics (10%, 7 errors), and sulfonamides (4.3%, 3 errors).

From direct communication with prescribing physicians, the primary contributing factors to these errors were identified ( Table 8 ). Most contributing factors were classified as "MD [prescribing physician] not aware of allergy." This was attributed to a variety of reasons, including workload, cross-coverage, and forgetting to review the drug profile. On a number of occasions, the physician was unaware of patient allergies until notification by either a nurse or pharmacist. In addition, when asked about the specific nature of the patient's reaction (i.e., date and severity), most physicians were unable to provide detailed information.

In all but a couple of cases, the medication error was prevented by a pharmacist who contacted the physician before order verification. When the physician was contacted, the prescribed drug was changed 87% of the time. In three cases, a patient's drug allergy profile was overlooked and at least one dose was given to that patient. The drug was continued in only one patient, and when contacted, the physician stated that the allergy was questionable. In addition, most piperacillin-tazobactam orders were changed to either a third-generation cephalosporin (ceftazidime) or a fluoroquinolone (ciprofloxacin or levofloxacin) with or without another antimicrobial agent.