Santa checks his list twice. Shouldn't we?
From the December 17, 2009 issue
Risk management procedures typically include recommendations for double-checks of critical work. Yet attitudes among those responsible for conducting them haven’t always been positive. The process takes extra time that some may feel is not justified by what they believe to be a relatively small number of problems that are missed initially. Others may believe that double-checks could lead to more mistakes as staff learn to rely upon others to catch problems. Or the checker may, based upon their colleague’s past performance, assume that the work is sound and be less observant during the double-check process. Some feel that double-checks just don’t work. Now, with staffing shortages and other time resource issues plaguing healthcare, the value of double-checks again has been called into question, especially since most facilities have experienced errors that reached patients despite double-checks. Re-search in community pharmacies and hospitals, however, suggests that double-checks are a good idea if conducted properly.
Double-checks identify a higher rate of errors than people realize. Among community pharmacists who randomly checked completed prescriptions awaiting pick-up, 4.2% of the 5,700 prescriptions were found to have errors.(1) Of these, 2.1% were considered to be potentially clinically significant. In another study, when pharmacists randomly compared 15,600 completed prescription forms against the attached computer labels, discrepancies were identified in 2.6% of them, and 1.4% were potentially clinically significant.(1)
The average “checking error rate” (errors missed during a double-check) is about 5%. In studies where “artificial errors” were introduced into medication carts and sample pharmacy orders, 93% to 97% of such mistakes were identified during an independent double-check.(1-2) Thus, even with relatively small error rates, a significant number of mistakes can be identified over time.
It’s hard to find your own mistakes. In a pharmacy simulation, for example, participants were significantly better at finding other’s mistakes rather than their own.(1) One reason is that confirmation bias fogs our ability to always see mistakes that are present. For example, the person preparing a medication may not be able to recognize that he’s selected the wrong drug because he can only see what his mind thinks he should be seeing—the intended look-alike product. But putting a fresh pair of eyes on something improves checking accuracy.
Double-checks work best when they are conducted independently. The person checking has to form an independent judgment without cues from the person doing the initial work (which is why ISMP encourages “independent double-checks”). Thus, telling someone to “Check this prescription for ZyrTEC 10 mg” would not result in an independent judgment. Such instructions create a mental image for the checker that can lead to verifying the originator’s conclusion (that it’s ZyrTEC 10 mg), even if erroneous. We’re more likely to use the information provided by others rather than form our own independent judgment. It would be more effective, for example, to have two people calculate a dose separately and then compare their answers, than to have them perform the calculation together or to have one share his answer with the other before the double-check occurs. People may be swayed by the opinions of others, even more so if the other person holds a position of authority or has more experience.
Limit checks to high-risk situations. While it’s true that independent double-checks are not as effective as system changes that make it hard for people to make mistakes, if performed properly, they can reduce the risk of an error reaching the patient. However, they should be limited to situations that involve certain high-alert medications, very complex processes, and high-risk patient populations. For example, independent double-checks might be limited to preparation and infusion rate settings for IV chemotherapy, IV and epidural opioids, IV insulin, IV heparin, neonatal parenteral medications, TPN compounding, and manual compounding of electrolyte solutions. Also maximize the inherent independent double-checks that naturally occur during medication use whenever possible. Examples include the independent double-check that occurs when a pharmacist dispenses medication and a nurse checks the accuracy, or when certain designated drugs are returned unadministered and pharmacy staff investigate to learn why the drug wasn’t given.
Too many checkpoints dilute the effectiveness of the process and are a sure sign of an ailing medication system with lots of Band-Aids downstream instead of prevention measures upstream. That’s why relying on double-checks alone without system changes has lead naysayers to the conclusion that double-checks don’t work.
Learn from captured errors. Finally, we can’t overemphasize the importance of documenting mistakes discovered during the independent checking process. Recognized errors must be immediately addressed, but there also must be assurance that these close calls are further analyzed to understand the reasons behind them and to identify precautions that need to be taken to reduce risk in the future. Unfortunately, some fatal errors have occurred in hospitals where it was later learned that the exact situation had previously resulted in a close call.
Simulations may be valuable in training staff to carry out independent double-checks effectively. Completing worksheets and recording initials also can formalize the process and enhance documentation. In the end, double-checks have their place in healthcare. Let’s not risk getting coal in our stockings for our failure to employ them properly!
The above article was modified from an original contribution by the late Anthony Grasha, PhD, who at the time of his death was a professor of psychology at the University of Cincinnati.
The Lilly Hospital Group offers a program called Double Check Safety Zone, which contains CDs and other materials to aid hospitals in promoting and implementing independent double-check systems. Contact your Lilly hospital representative for a copy.
References: 1) Grasha AF, et. al. Delayed verification errors in community pharmacy. Tech Report Number 112101. Cognitive Systems Performance Lab. 2) Campbell GM, Facchinetti N. Using process control charts to monitor dispensing and checking errors. Am J Health-Syst Pharm 2000;55:946-952.