Abstract
Keywords
The Working Group on CGM of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) has developed a comprehensive guideline for the clinical assessment of CGM systems. 1 The guideline defines requirements for study design and procedures, characteristics of comparator measurements, minimum requirements for accuracy, and performance characterization.
This letter addresses minimum requirements for accuracy and performance characterization.
Generic accuracy requirements
In the Milan conference, 2 a group of pathologists and clinical chemists proposed criteria for accuracy requirements. The advocated hierarchy is:
clinical outcomes
biological variability
state of the art
The IFCC paper uses state of the art requirements in spite of the fact that clinical outcomes are available with an error grid.
Minimum Acceptance Requirements for Accuracy
The IFCC guideline requires the percentage of comparison results to be greater than a stated percentage (ranging from 67% to 94%) for either 20% or 40% differences, depending on the reference value. I have published 3 a critique of these minimum acceptance requirements for CGM accuracy.
Note that, a 40% difference is in the B zone of an error grid. This region is associated with adverse events (including injuries). Moreover, allowing for a percentage of results to be greater than 40% suggests that these results are also acceptable. With more than 100 000 CGM annual results per person, even a small number of results above 40% can be a large number. For results less than 70 mg/dL, the requirement is less than 6% above 40% which is 6000 results! Using 65 mg/dL as an example, 40% low is 26 mg/dL (severely hypoglycemic) and 40% high is 91 mg/dL (normal). A better wording for this requirement would be—accuracy requirements that must be met to avoid being rejected.
Additional Performance Reports
There are no requirements for these metrics. Some concerns.
MARD is recommended as a performance report in spite of concerns with this metric. 4
That an error grid is mentioned as another performance metric (same as MARD) is hard to understand. An error grid has been shown to be the best assessment of clinical accuracy. 5 The error grid should replace the minimum acceptance requirements for accuracy. If an error grid were used as a requirement, not all results would be in the A zone (no risk). Using the currently proposed minimum acceptance requirements for accuracy will have one trying to see where results fall in an error grid rather than using the error grid as an accuracy requirement.
Footnotes
Acknowledgements
None
Abbreviations
IFCC, International Federation of Clinical Chemistry and Laboratory Medicine; CGM, continuous glucose monitoring.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
