A very commonly-implemented intervention when attempting to “mistake-proof” a process involving humans is the double-check.

Double-checks are, by their nature, waste.  Often, double-checks are put in place as temporary stop-gap measures while more effective, value-adding countermeasures are devised and implemented.  While a perfectly-implemented double-check does indeed improve error rate, it does not come close to serving as effectively as a prevention-based mistake proof device (poka yoke) and should therefore be considered a weak intervention.

The prevailing theory behind a double-check is that, while a person will occasionally err when undertaking a task, with probability A, the process will be made more robust by inserting a double-check of the work of the original person.  The presumption is that the person doing the double-check will also occasionally err, with probability B, but that the combined probability of error emanating from the double-checked task will be geometrically lower (improved).  For example if the error rates of the original person performing the task is 10% (A=0.10) and the error rate of the double-checker is 15% (B = 0.15), the combined error rate of the double-checked process is now (0.10)(0.15)=0.015, nearly an order of magnitude improvement.

Actual results of such interventions in practice rarely match the theoretical improvement.  By understanding the factors leading to gaps in performance of double-checks, we can set out to implement double-checks in such a manner that efficacy is enhanced.

Factors leading to sub-optimal performance of a double-check

Often, a double-check step is inserted into a process “on the fly” without thoughtful planning to eliminate possible failure modes of the check itself.  Following are some common situations to avoid when considering a double-check for your process.

• Lack of clear definition (standard work) of how the double-check is to be performed
• Ineffective training of staff who are to perform the double-check
• Failure to design an independent double-check; one that only involves the person originating the work has a high likelihood of error due to confirmation bias, which can blind the person performing a check on his/her own work .
• When a second person is assigned to perform the double-check, confirmation bias can blind this person.  This can be caused by specific information being passed from the first to the second person prior to validation, the second person subconsciously deferring to the authority of the first person, or simply trusting that the first person could not have erred.  Any of this leads to the second person missing errors.
• Creeping complacency, due to a person knowing that someone else is in place to catch any mistakes he or she will make
• Interruptions during the double-check
• Rushing the double-check

Related, and equally common, is the issue of nobody being tasked with examining errors caught during the double-check and using that insight to drive focused root cause analysis and interventions upstream to eliminate as many of these failure modes as possible.

What can you do?

1. Assess and analyze any double-checks you already have in place.  Look for evidence of any of the above issues.  Ask to see the data showing what errors are being found.  Investigate what actions have been initiated as a result.
2. Before assuming that a double-check is the strongest mistake-proof device (poka yoke) that can be designed and implemented for a particular task within a given value stream, challenge yourself and your team to seek out stronger interventions.  Estimate the relative strength between multiple intervention ideas.
3. Where a double-check is the only course of action, either for the short term or for a longer time period, infuse these design principles:

• independence of the check, separated by time and space if possible to mitigate risk of “contaminating” the checker with information about the person originating the work, inputs, outputs, calculations, and judgement applied
• to simplify and make visual the double-check, create a checklist showing the specific steps to be undertaken
• create a physical environment for the double-check that will be free from distraction and which will make the task simplest
• carefully document the standard work, train those involved, and assess skill/capability of those who will be performing the check
• avoid having the same two people serving as checkee-checker on a sustained basis
• periodically assess efficacy of the check, in a blinded fashion and provide feedback
• instrument your check and specify roles to include those who will prepare, analyze, and act upon the errors caught by the check

4.   Seek to continually improve, through successive cycles of mistake proofing, until you can evolve beyond the use of double-checking

Principles and Techniques for Sustained Improvement

Thursday June 13th at 1PM Eastern.

Register Now

Mistake-Proofing (or Poka Yoke) prevents mistakes from being made or makes the mistake obvious at a glance, so that corrective action can be taken before a patient is harmed. It’s a common sense approach tocreate low-cost, effective design changes that reduce errors.

Brian Nass has worked in quality and continuous improvement for over 20 years.  He held several leadership positions at Mayo Clinic in Rochester including Executive Director of Continuous Improvement and COO of Lab Services. Most recently as VP, Performance Improvement at Mission Health System in North Carolina.

Learn techniques for eliminating defects, improving quality and streamlining work.

Brian will discuss mistake-proofing within the specific context of healthcare; however, this webinar is not specific to only the healthcare industry. Many examples and references will come from healthcare, however this webinar will benefit managers from all industries who want to better understand the widespread application of mistake-proofing. Join Us!

Register Today