Saturday, June 4, 2011

Stopped clocks and Uncertainty of Measurement

Uncertainty of Measurement in the medical laboratory is not all bad. Think of UM as being like the stopped clock which tells the right time twice a day.

To “calculate” the uncertainty of measurement one develops an uncertainty budget to determine the impact of all the steps that might have an impact on the variation in measurement values. While this may make some theoretical sense, the reality is that errors that occur especially in the pre-examination phase tend to occur randomly and intermittently and on a periodicity best described as irregularly irregular. Trying to make calculations on random events is almost impossible, unless one studies each step many many times, to get, as the statisticians say, a sufficient n value. And that is an unrealistic and bizarre expectation to be inserted into a standard that is treated as a requirement for accreditation.

So at what point can this be a good thing, even fleetingly? Well, I will tell you.
In the implementation of a Quality system, there are two processes that will make significant improvement to an organization. Some refer to this as CAPA as if the two are closely linked (which is incorrect).
The first is corrective action; once a problem is detected, it needs to be examined and analyzed for likely cause, and then be addressed to reduce the risk of repeated and ongoing weakness. This makes infinite sense, and is generally a pretty easy requirement for folks to understand, and is relatively easy to implement. Corrective actions are reactive responses to detected problems.

Preventive actions are different because they are more assertive or pro-active. One takes a process and looks at all its steps for potential weak points which could fail and result in error or failure. One is looking for potential opportunities for future error and preventing them from occurring. Quality system standards expect organizations to regularly take the time to not only be pro-active in preventing accidents, injuries, errors.
In order to plan a medical laboratory preventive action program, first you have to know what happens at each step along the way to the samples. You need to think about processes looking for potential error. In the world of risk management this process is call looking for potential failure modes and their effects analysis (also known as FMEA). And this step is essentially the first step that leads to developing an uncertainty budget. So while trying to calculate a value for uncertainty is generally a waste, we end up with information that we can use for other purpose.

Looking for weak points and addressing them is a useful exercise and links risk management, and quality management preventive action programs. It is described and recommended in a guideline (called a technical specification) published by the International Organization for Standardization as ISO/TS 22367:2008 titled “Medical laboratories -- Reduction of error through risk management and continual improvement”.

To come back to my point that CAPA is a poor acronym, most institutions find doing corrective actions relatively easy.  Doing preventive actions are tough.  Many organizations find them the toughest part about quality management.  Maybe approaching preventive actions by working through a budget approach will make them easier to develop. 

So in my world the following is clear: Requiring uncertainty of measurement in the medical laboratory is almost a total waste of time, with the singular exception that it does motivate us to take a closer look at individual processes with the view for finding potential weak points that we can amend through an active Preventive Action process.

Which brings me back to the stopped clock.

No comments:

Post a Comment

Comments, thoughts...