Standardization, Harmonization, and Realization

Abstract

Point-of-care testing is useful when caring for patients in hospital settings and in emergency and disaster situations. However, point-of-care professional practice lacks components, such as standardization, harmonization, and consistency, which would substantially improve patient care if implemented. Therefore, we propose adoption of whole-blood standards, harmonization among testing methods, and tighter quality control constraints. Granting these 3 wishes will improve quality at the point of care and ultimately will improve diagnoses, treatment decisions, and patient outcomes.

Standardization: The World View

Standardization is a practical process aimed at achieving consistency among measurement procedures through application of high scientific standards. Standardization has become an important concept in clinical diagnostics, both nationally and globally. Key components of standardization include higher-order materials and measurement procedures, such as the Standard Reference Materials (SRMs) for diagnostic tests developed by the National Institute of Standards and Technologies (NIST) (Table 1).

TABLE 1.

Standard Reference Materials From the NIST

Matrix Category	SRM (NIST)	Clinical Marker
Serum	909b	Cholesterol, Cl, Na
	956a	Na
	965a	Glucose
	965b	Electrolytes
	967	Creatinine
	968c	Vitamins, carotenoids, cholesterol
	971	Hormones (ie, testosterone and estradiol)
	1951a,b	Lipids
	1955	Homocysteine, folate
	2921	Cardiac troponin I
	NA*	C-reactive protein
Urine	1507b	Multiple drugs
	2669*	Arsenic species
	2670a*	Na, Pb, toxic metals
Plasma	1950*	Metabolites (ie, cholesterol and glucose)
Whole blood	Not available	Not available

Information was found at www.nist.gov. Accessed April 4, 2008.

^*Indicates that NIST is currently undergoing research for validation.

Results of routine measurements are standardized through calibration to a reference method and/or material but, when the reference system is lacking, are only referred to a manufacturer's selected procedure and corresponding calibrator. Ideally, for example, manufacturers of metabolite assays (eg, plasma glucose) could use SRM 1950, once fully developed (Table 1), to assess the validity of their calibrators. This practice will result in standardized measurements with closer agreement of results in clinical practice. In support of the standardization practice, societies such as the International Federation of Clinical Chemistry and Laboratory Medicine have set forth to promote extensive global standardization in laboratory medicine,¹ and in the case of glucose, the current de facto sample matrix for standardization, generally, is accepted to be plasma.

The POCT notion of bringing the test immediately to the patient is accomplished via convenient handheld, portable, or transportable devices, many of which use whole-blood samples for results. The rapid technological advances in POCT permit measurement of multiple analytes in whole-blood samples. Additionally, whole-blood analysis in POCT decreases therapeutic turnaround time and facilitates shorter time to treatment.² However, the lack of whole-blood SRMs reveals a standardization deficiency in the POCT field that can adversely impact diagnoses, treatment decisions, and patient outcomes. Therefore, our first aspiration is for global standardization in POCT, which will ultimately result in an improvement of clinical interpretation of laboratory results to benefit patient care.

Harmonization: The Bedside View

Harmonization in medicine means bringing diagnostic standards and treatments into consonance or accord. However, at this point in time, no single reference or calibration method correlating plasma and/or whole-blood glucose and other analytes has been adopted universally among manufacturers. Additionally, if we adopt a bedside perspective, then most reference instruments should not show statistically significant bias relative to the point of care. Currently, however, there is a proven lack of harmonization of reference instruments used in hospitals across the United States (Fig. 1) when viewed from the bedside perspective. This deficiency in harmonization potentially introduces complications in clinical management, which can adversely affect treatment and care of patients on tight glycemic control (TGC).³ What counts is that all instruments, whether POC or reference, should be harmonized and better standardized.

FIGURE 1.

Quantitative comparison of clinical laboratory reference analyzers used at 35 United States Medical Centers—the bedside perspective. Bar graph illustrating glucose measurement differences (glucose meter − reference) calculated from comparing glucose values from 1 type of glucose meter versus various laboratory reference analyzers currently used in hospitals across the United States. The x axis represents the 7 clinical analyzer types. The y axis represents glucose difference. The SD bars are shown. Student t test for paired differences was used to determine the significance of the difference for each analyzer type. P values less than 0.001 are denoted by 3 asterisks (***), and P values less than 0.01 but greater than 0.001 are denoted by 2 asterisks (**).

Now, adding “global” into the equation would mean unifying international standards to create worldwide harmony in diagnostic testing. In the absence of global harmonization, the likelihood that a result may be misinterpreted and the patient may be misdiagnosed is increased because of different assays exhibiting nonequivalent analytical responses. A spectrum of different values may result from testing of the same specimen by different methods. Data reporting comparisons of global SRMs show an international bias resulting in noncomparable patient outcomes due to different field methods.^4,5 Bias in measurements has also shown to impact medical decision making.⁴

On the other hand, the improvement of precision in cholesterol standards since 1968 has been estimated to save $100 million per year in treatment costs.⁴ Now, imagine the potential savings in treatment costs and improvements in patient outcomes that could be made possible by implementing standards for POC. Without the immediate prospect for whole-blood standards (Table 1) or, in the case of glucose, for even plasma standards produced by the NIST, our second inclination is for national, followed by global harmonization.

Additionally, too long, the perspective has been clouded. Preanalytical delays and other confounding factors deteriorate samples before they are assayed in clinical laboratories. In many cases, bedside results may reflect actual in vivo conditions more accurately. Therefore, we recommend that harmonization adopt the bedside perspective in settings, such as critical care, where nurses and physicians need physiologically relevant and timely evidence for decision making and also will be obtaining that evidence using POCT.

Consistency: The Need for Quality

Common sense dictates the following: (a) quality control (QC) results should agree better, that is, be more consistent; and (b) tolerance bands generally are too wide for use in settings, such as TGC, where glucose meter results trigger rapid periodic adjustments in insulin doses. We observed that, although high- and low-level QC values may lie within manufacturer-expected tolerance limits, analysis of variance for test groups of QC results shows statistically significant (P < 0.001) differences between glucose means, demonstrating a lack of consistency (Fig. 2). Additionally, we have shown using locally smoothed median absolute difference curves that an acceptable level of performance required for TGC is achieved by only a few hospital glucose meter systems.^6,7

FIGURE 2.

Quality control glucose values and acceptable manufacturer tolerance limits. The x axes display the high-level (255-345 mg/dL) (A) and low-level (25-55 mg/dL) (B) test groups, each reflecting glucose results from n = 10 test strips. The y axes illustrate the glucose QC results. Groups were tested using 2 glucose meters. All raw glucose values plotted lie within the manufacturer-accepted ranges denoted by the black horizontal lines. Numbers next to points indicate overlapping QC values. Analysis of variance performed for groups showed P < 0.001 for both high- and low-level QC.

Successfully meeting QC requirements represents a fundamental criterion for the daily use of POC devices. Therefore, our third desire is that hospitals adopt more than 1 QC level for glucose meters and that QC tolerances be tighter. Use of at least 2 QC levels will lead to more consistent bedside performance, whereas improved use of QC constraints will induce a realization of improved consistency on a proactive basis. By analogy, QC tolerances and the number of levels used for other POC analytes should follow this same consistency model.

Conclusions

The global POCT field is in need of whole-blood standards, harmonization among methods, and improved QC. Granting these 3 wishes will facilitate common sense consistency among measurement procedures performed at the point of care and, in our opinion, will ultimately improve diagnoses, treatment decisions, and patient outcomes.