If total error – calculated by the usual models – is set to equal 95% of the results, then how can 95% instead of 100% represent total error? Shouldn’t “total” imply all results? Putting things another way, what do you call the additional 5%?
I was working on a paper and decided to comply with the nomenclature expected by the journal and used the word “measurand.” The word was underlined as unknown by the dictionary used by “Word.” I went to an online version of the Merriam Webster dictionary and no match for found for measurand. So much for ISO nomenclature.
Clinical Chemistry has published one of its Q&A pieces (subscription required) with the title “Triangulating Dynamic of Clinical Laboratory Testing.” With a title like that it’s not surprising that 8 pages contained no useful information. Given that some of these people work for laboratory medicine consulting companies, it would have been nice to provide information such as test volume by analyte, company revenues and profits, revenue by analyte, etc. But don’t expect to see any Letters about this piece because Clinical Chemistry won’t accept Letters about anything other than articles.
I was cleaning out some old files and came across a strategy document. The strategy happened when our company had been bought and a consulting company was there to assess the various groups. The summary for our group was that we had lost focus and would be better served if we stuck to our core competence of evaluations and reliability. Evaluations meant product performance evaluations and statistical analyses related to getting a product approved by the FDA and reliability meant improving the reliability of instrument systems under development through the use of data analysis methods.
This was pretty standard stuff back then. When a company was bought, you brought in a consulting company started by Harvard Business School professors (this one was called Integral) to suggest a bunch of changes.
The problem was that our original core competence was just evaluations. The reliability component had been started by our group as a means of expanding our services where we thought we could provide value. In fact, our reliability services had become our most valuable service. Had this consulting company been brought in when our reliability program was in its infancy, undoubtedly, we would have been told to drop it and stick to evaluations.
The point is that most organizations look to grow and the attempts to grow involve risk. Some projects succeed – most fail. Our company back then tried to expand in other areas – pulmonary function, pulse oximetry, coagulation: all failed. One other project – the automation of an immunoassay analyzer – the ACS 180 – was a huge success.
So some efforts should be devoted to high risk high reward projects.
The Westgard web has some comments about IQCP.
Here are mine.
- There is no distinction between potential errors and errors that have occurred. This is non-standard. In traditional risk management different methods are used for potential errors vs. errors that have occurred. For example on page 12 of the IQCP book which focuses on specimen risks, “Kim” reviewed log books and noted errors. Yet on the same page, Kim is instructed to ask “What could go wrong.” The problem is that there are clearly errors that have occurred yet there could be potential new errors that have never occurred.
- The mitigation steps to reduce errors look phony. For example, an error source is: “Kim noted some specimens remained unprocessed for more than 60 minutes without being properly stored.” The suggested mitigation is: Train testing personnel to verify and document: Collection time and time of receipt in laboratory and proper storage and processing of specimen. The reason the mitigation sounds phony is that most labs would already have this training in place. The whole point of risk management is to put in place mitigations that don’t already exist.
- There is no measurement of error rates. Because there is no distinction between potential errors vs. errors that have occurred, there is a missed opportunity to measure error rates. In the real world, when errors occur and mitigations are put in place, the error rate is measured to determine the effectiveness of the mitigations.
- The word “Pareto” cannot be found in IQCP. Here is why this is a problem. In IQCP, for each section, a few errors are mentioned. In the real world, for either potential errors or those that have occurred, the number of errors is much larger. So much larger that there are not enough resources to deal with all errors. That is why the errors are classified and ranked (the ranking is often displayed as a Pareto chart). The errors at the top of the chart are dealt with. In the naïve IQCP, there is no need to classify or rank errors because all are dealt with. The same problem occurs in CLSI EP23 and ISO 14197.
Conclusion: One might infer that no one who participated in the writing of IQCP has ever performed actual risk management using standard methods or perhaps any methods.
I had occasion recently to look at IQCP – here are some comments
It seems that if a lab performs IQCP, they can change their QC frequency (read reduce) from the CLIA minimum of twice per day. The lab can’t reduce it more than the manufacturer’s recommended frequency.
This is interesting as it seems that manufacturers will now have the ability to set QC frequency!
Say, the QC frequency turns out to be once a month (which was the case allowed for EQC).
To me, this makes no sense – IQCP or not – since this could mean that a month’s worth of patient samples are incorrect. How can this be good. Moreover, I remember talking to lab directors and asking them, does QC ever go out and the answer was yes, it does.
IQCP requires performing risk management. This should be performed anyway! The notion that performing risk management – which is not that easy to be done well – will somehow obviate the need to check performance by running QC is delusional and not in the best interests of clinicians and patients.
This article is now online (subscription required).