1) A physician ordered a heparin infusion with directions to follow a weight-based nomogram for laboratory monitoring and dose adjustments. Later that evening, the nomogram indicated that a bolus dose of heparin 1,700 units IV should be administered based on the patient's aPTT level. The patient's nurse removed a 10 mL vial of heparin (1,000 units/mL) from an automated dispensing cabinet to prepare the dose. However, she miscalculated the volume that was needed as 17 mL, not 1.7 mL. The nurse, concerned that she would be using a second vial of heparin to prepare the bolus, quickly asked another nurse to "look at my math" to make sure she had not made an error. But the other nurse did not actually recalculate the volume needed, so she made the same error when "looking over" her colleague's work. The patient received 17,000 units of heparin. A physician's assistant discovered the error after the patient developed severe epistaxis.

2) An epidural infusion of fentanyl 2 mcg/mL) with bupivacaine (0.125%) was started on a 62-year-old man who had just undergone a lobectomy for cancer of the lung. The drug was supplied as a premixed product manufactured by Baxter Compass. Several nights later, a supervisor went to an automated dispensing cabinet to retrieve a replacement bag. But she accidentally picked up a premixed Compass bag of morphine (1 mg/mL) intended for intravenous use, which was located in the same drawer as the fentanyl/bupivacaine bags. Both the IV morphine and epidural fentanyl/bupivacaine bags were supplied by Baxter Healthcare Corporation, which recently sold the Compass products to PharMedium Healthcare Corporation. Both bags were packaged in identical brown plastic overwraps to shield the compounded solutions from light. The labels, located on one side of the brown overwraps, were also similar in appearance, and both products were packaged in the same size bags (100 mL in 150 mL container). The supervisor brought the bag to the nursing unit. A second nurse doublechecked the product, but also failed to notice the mistake since the bag was packaged in the brown overwrap, as she'd come to expect. The morphine was hung, and several hours later the patient's respiratory status began to deteriorate, so the epidural infusion was temporarily turned off. Even then, staff did not notice the error. Another nurse, who was documenting the waste after the patient's epidural catheter was removed, finally discovered the error.

While multiple system failures clearly contributed to these errors, in both cases, failed double-checks allowed the errors to reach the patients. Why did the double-checks fail? In part, the answer lies with how the double-checks were performed and the differences between endogenous and exogenous errors (1).

Case 1 - an endogenous error. An endogenous error arises solely from within an individual, from a random and unpredictable cognitive event like miscalculating a dose or prescribing a drug at a dose appropriate for the next medication being contemplated. In Case 1, the nurse made an endogenous error when calculating the volume of heparin to administer. Because endogenous errors arise within a single person, another person performing the same function does not often make the same error. Thus, endogenous errors are likely to be detected if a double-check is performed independently by another person, as a separate redundant action. This way, the checker is not misled into the same faulty thinking as the person who originally made the error. In Case 1, had the double-check been performed independently as a redundant function without prior knowledge of the first nurse's work, it's far more likely that the error would have been detected.

Case 2 - an exogenous error. An exogenous error arises from conditions in the external environment, like poor design of packages and labels, complex task characteristics, or unclear presentation of information. In Case 2, the nurse ade an exogenous error related to the look-alike packaging of Compass bags. A subsequent check by another nurse did not uncover the error. Doublechecks are often less successful in detecting exogenous errors than endogenous errors, even when the check is performed independently. Some of the same external factors that initially led to the error are often still present, and people with similar training could easily make the same mistake during the doublecheck.

Avoid sole reliance on double-checks. While double-check systems will sometimes fail - more so for exogenous errors - they still play a vital role in error detection strategies when strategically placed at the most vulnerable points of medication use, and when performed independently. But the hoped-for improvement in system reliability will be illusory if you rely on these manual double-check systems alone to catch all errors. System changes must also be made to reduce the frequency of errors. "Sameness" distracts Auxiliary labels help differentiate Using the exogenous error in Case 2 as an example, better labeling of the Compass products is necessary to prevent errors. (Compass pain management products also include epidural morphine, IV and epidural hydromorphone, IV meperidine, IV midazolam, epidural bupivacaine, and others also packaged in brown overwraps. We spoke with PharMedium to suggest label improvements.) The hospital that reported this error now applies large yellow "FENTANYL/BUPIVICAINE For Epidural Use Only" labels (to match the yellow stripe in the epidural tubing) or blue "CONTAINS MORPHINE Not for Epidural Use" labels on the bags (see photo) and the overwraps. The labels are applied to both sides of the bags and overwraps so they can be seen regardless of the bag's orientation in the pump or storage area. These labels are also applied to the cartons stocked in the pharmacy. The products are stored separately in both the pharmacy and in automated dispensing cabinets. Pharmacy avoids having these look-alike products delivered on the same day to prevent mix-ups during the order fulfillment process or upon receipt in the hospital.

Likewise, for Case 1, system-based error reduction strategies can be employed to further prevent calculation errors: dosing charts that eliminate the need for calculations and pharmacy preparation of nonemergency drugs are just two examples.

Reference: Senders J. Essays on human error in medicine. ISMP-Canada, October 2000. www.ismp-canada.org/smp0010.htm