From the March 6, 2003 issue

Has your double check system ever failed, leading to a medication error that escaped your detection and ultimately reached a patient? If you answered “yes” to this question, you’re not alone. Here’s one recent example. A pharmacist correctly calculated the dose and volume of interferon for an infant, but entered 0.68 mL into the computer instead of the correct volume of 0.068 mL (a common mistake documented in the literature). A second pharmacist double checked the calculation. He arrived at the correct volume of 0.068, but misread the computer entry of 0.68 by the first pharmacist as 0.068 due to confirmation bias – seeing only what one expects to see and overlooking any disconfirming evidence.

As this example shows, there’s no question that double checks carried out by people fail at times. But have these failures led you to doubt the overall value of double check systems? Given how busy healthcare professionals are, do you wonder if this error reduction strategy is even worth your time to carry out? We asked Dr. Anthony Grasha, Professor of Psychology at the University of Cincinnati, to offer comment on this issue.

Research shows that people find about 95% of all mistakes when checking the work of others.1,2 Mathematically, the benefit of double checks can be demonstrated by multiplying this 5% error rate during the checking process and the rate in which errors occur with the task itself (the checking error rate x the task error rate). For example, if a pharmacy dispensing error rate is 5% (based on research findings), and a double check occurs before medications are dispensed, then the actual chance of a dispensing error reaching the patient is 5% of 5%, or only 0.25%.

Human factors suggest that double checks are more effective if they are performed independently. For example, an error in the concentration of a drug will be detected more often if the person checking the product performs all calculations independently, without knowledge of any prior calculations. In fact, sharing prior calculations or performing a double check together with the person who originally completed the task is fraught with problems. In these instances, if a mistake is present, the person checking the work is more easily drawn into the same mistake, especially if it appears to be correct at first glance (e.g., numbers correct but decimal point placement wrong, correct drug but wrong concentration selected during PCA pump set-up).

Dr. Grasha also points out that the effectiveness of double check systems depends on training staff to carry them out properly – as an independent cognitive task, not a superficial routine task. And with workload issues looming heavily over practitioners, double checks should only be applied strategically to situations that most warrant their use – prescribing, dispensing, and administering select high alert medications. These have the greatest chance of harming patients if misused. Fewer well-placed double checks will be much more successful than an overabundance of double checks.

References: 1) Grasha AF, et. al. Delayed verification errors in community pharmacy. Tech Report Number 112101. Cognitive Systems performance Lab. Contact: Tony.Grasha@UC.Edu. 2) Campbell GM. and Facchinetti N. Using process control charts to monitor dispensing and checking errors. Am J Health-Syst Pharm 2000; 55: 946-952.