Printer Friendly

Wrong thinking about glucose standards.

To the Editor:

In recent comments about the ISO 15197 glucose standard, experts have stated that the limits for 95% of the data should be tighter (1, 2). The unmentioned but important problem with this ISO standard is that it specifies limits for only 95% of the data. This approach leaves up to 5% of the data unspecified, meaning that up to 5% of results could have an error sufficiently large to lead to patient harm but still be within acceptable assay limits according to the ISO standard. Imagine specifying that up to a 5% rate of wrong-site surgery is acceptable! I have pointed out that it is necessary to treat assays that have continuous error as discrete event variables (such as wrong-site surgery) (3). In addition to the specification of limits for 95% of the data, limits should be specified for where no (or very little) data should occur. Accounting for both sets of limits creates an error grid (4).

The ISO standard's use of one set of limits implies that all values within those limits indicate low potential for patient harm and that all values outside of the limits indicate high potential for patient harm. Such an inference is illogical because values just inside the limits and values just outside the limits have almost the same amount of error and should indicate the same potential to harm patients. As is shown in Fig. 1, it is more realistic to expect an increasing potential for harming patients as the magnitude of the error increases, as is obtained with the Taguchi loss function. An error grid is an approach to approximate this loss function because it has multiple points.

It would be incorrect to think that large errors are unlikely because of the low probability of high multiples of an SD based on replication experiments. Large errors are often due to different processes, such as interferences. Rare large errors are unlikely to be observed in a short method comparison. To have 95% confidence that < 1 large error will occur in 1 x [10.sup.6] samples requires a sample size of 371 000 000. To evaluate the likelihood of rare errors requires risk-management techniques (5). Failure mode and effects analysis (FMEA) and fault trees are useful for such purposes and require a team to map the process; ask what can go wrong, how likely such an event is, and how severe such an event would be; and then decide what to do about it.


The goal in evaluating an assay is to estimate any error (large or small) that clinicians would observe in its routine use (e.g., total error). ISO 15197 has a separate section called "User Performance Evaluation." In such an analysis, a separate evaluation compares the results produced by a user and by a trained healthcare professional; however, the only analysis requirements stated are that "Results shall be documented in a report." Yet, individuals who use self-monitoring of blood glucose de vices experience pre- and post-analytical error in addition to analytical error alone. The implications are that in ISO 15197 user errors are not part of the quantitative part of the specification and that the ISO standard fails to inform clinicians of the true performance of self-monitoring of blood glucose.

Ideally, a specification should be quantitative, cover 100% of the data, and include a protocol and an analysis method. "Protocol" is used in a generic sense here because FMEA and a fault tree can be considered a protocol.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.

Authors' Disclosures of Potential Conflicts of Interest: No authors declared any potential conflicts of interest.

Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.


(1.) Sacks DB. Tight glucose control in critically ill patients: Should glucose meters be used? Clin Chem 2009;55:1580-3.

(2.) AACC. September 2009 clinical laboratory news: higher standards on the way for glucose meters? Pages/inside0909.aspx (Accessed November 2009).

(3.) Krouwer JS, Cembrowski GS. A review of standards and statistics used to describe blood glucose monitor performance. J Diabetes Sci Technol 2010;4:75-83.

(4.) CLSI. How to construct and interpret an error grid for diagnostic assays; proposed guideline. Wayne (PA): CLSI; 2009. CLSI document EP27-P.

(5.) CLSI. Risk management techniques to identify and control laboratory error sources; approved guideline--second edition. Wayne (PA): CLSI; 2009. CLSI document EP18-A2.

Jan S. Krouwer *

Krouwer Consulting Sherborn, MA

* Address correspondence to the author at:

Krouwer Consulting 26 Parks Dr. Sherborn, MA 01770 Fax 508-653-2379 E-mail

Previously published online at DOI: 10.1373/clinchem.2009.140277
COPYRIGHT 2010 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Letters to the Editor
Author:Krouwer, Jan S.
Publication:Clinical Chemistry
Article Type:Letter to the editor
Date:May 1, 2010
Previous Article:Bilirubin standardization in the Netherlands: alignment within and between manufacturers.
Next Article:Expanding and improving research attribution with the open researcher and contributor identification initiative.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters