HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Description of Analytical Methods and Results

Manuscripts describing the development and evaluation of the performance of methods and instruments should discuss linearity, imprecision, analytical specificity, recovery, lower limit of detection, comparability with other analytical methods, lower limit of quantification and reference interval(s). Some clinical data are usually needed.

Document the analytical advantages of the new or modified method over existing methods.

Calibration curves and linearity. Data for these studies should be subjected to linear regression analysis (if a linear response is obtained) and should include the slope, intercept, r², standard deviation of residuals, and the standard deviations of the slope and intercept.

Standard deviations of repeated points may be included.

In preparing nonlinear calibration curves, authors may use any objective, statistically valid method but must specify the method used (see, e.g., Ref. (1)).

Imprecision. Studies must include estimates of "within-run" and "total" standard deviations (1). Each should be determined at low, normal, and above-normal concentrations with use of specimens that are in an appropriate biological matrix.

One method of estimating both within-run and total standard deviations is the analysis of variance experiment described in NCCLS EP5-T (2), which calls for two replicates per specimen per run and two runs per day for 20 days. This permits separate estimation of between-day and between-run, within-day standard deviations, as well as within-run and total standard deviations.

For acceptable alternatives that include only one run per day, see the cited document.

Indicators of Accuracy ("Trueness"). Accuracy (or “trueness” in the recent nomenclature) of a new method can be estimated by (a) analyses of certified Reference Materials by the new method or (b) comparisons of results of a new method with results of a Reference Method. These are the only accepted approaches to trueness. When neither is available, other evidence relevant to the ability of the method to measure the analyte (measurand) is needed. Recovery studies involve analyses after known amounts of analyte are added to the biological fluid on which the determination will be performed. Recovery of added analyte should be calculated [(final concentration – initial concentration)/added concentration], not the observed final concentration as a proportion of expected final concentration.

Interference studies should be performed to assess the effects of common interferents, including lipid particles, hemoglobin, bilirubin, and components of uremic plasma. Exogenous materials, such as ingredients of blood collection containers (tubes) and commonly used or commonly coadministered drugs that might interfere with the determination, should also be tested for interferences. Selection of materials to test should be guided by an understanding of the chemistry and physics of the measuring system. Thus chemicals that are structurally similar to the analyte should be tested to assess the selectivity of the method. (The term “selectivity” is preferred over specificity; selectivity can be quantified.) In characterizing non-spectrophotometric methods, chemicals that may interfere in the detection system should be studied more intensively than chemicals that are historically important for interference in spectrophotometric methods.

Comparison-of-methods studies should compare results by the new or proposed method with those by a reference-quality method or other generally accepted analytical method for which assay performance is documented (3, 4).

It is desirable to test 100 to 200 different samples from patients who have been selected to include a wide variety of pathologic conditions and to present a range of values for the analyte that includes those likely to be encountered in routine application.

For a table of the required number of samples, see Linnet (5).

If regression analysis is used for statistical evaluation of the data, supply slopes and intercepts (and their standard deviations) and standard deviations of residuals (Sy|x, often called standard errors of estimates). Unbiased (e.g., Deming) regression is typically required (6). A program to perform Deming regression is available on-line as a supplement from this journal (7).

The correlation coefficient has limited utility. Residuals plots [e.g., Bland-Altman (8, 9)] are often useful. On the horizontal axis, plot the mean of results by the two studied methods, no the result of one method.

Analytical sensitivity and detection limit. These terms are commonly confused. The International Union of Pure and Applied Chemistry defines analytical sensitivity as the ability of an analytical procedure to produce a change in signal for a defined change of the quantity This is often visualized as the slope of the calibration curve.

Detection limit (or limit of detection) is defined as the lowest concentration or quantity of an analyte that can be detected with a stated reasonable uncertainty for a given analytical procedure. The operational definition of this limit must be supplied by the author: e.g., the concentration at a signal-to-noise ratio of 10 or the concentration corresponding to a signal 3 SD above the mean for a calibrator that is free of analyte. This is also called the limit of the blank.

Analytical quality. Results obtained for the performance characteristics should be compared objectively with well-documented quality specifications, e.g., published data on the state of the art, performance required by regulatory bodies such as CLIA 88, or recommendations documented by expert professional groups (10).

Reference interval (normal range). Depending on the conclusions of the accuracy studies, modification of an accepted reference interval may be indicated. Description of the reference interval study should include details about sampling; selection of subjects, including their number, age, and sex distribution; the statistical method for summarizing the results (11); and other factors that would influence the values obtained.

Chromatograms. Chromatograms from gas-liquid and liquid chromatography should usually be presented so that readers can see the efficiency of the separation and observe the resolution from interferents in the matrix. Similar images are often needed for electrophoretic separations.

Enzyme activities. Enzyme activities may be expressed in international units (U) or katals. Temperature and other key assay features must be described in the text or by reference to a published method.

When first mentioned in the text, enzymes (whether measured by activity or mass assays) must be numbered (EC no.) in accordance with the recommendations of the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology on the Nomenclature and Classification of Enzymes (12).

References

1. Linnet K, Boyd JC. Selection and analytical evaluation of methods – with statistical techniques. In: Burtis CA, Ashwood ER, Bruns DE. Tietz textbook of clinical chemistry and molecular diagnostics. 4th ed. St. Louis: Saunders, 2006: 353-407.
2. NCCLS Tentative Guideline EP5-T. User evaluation of precision performance of clinical chemistry devices. Wayne, PA: National Committee for Clinical Laboratory Standards, June 1984.
3. Carey RN, Garber CC. Evaluation of methods. In: Kaplan LA, Pesce AJ, eds. Clinical chemistry. Theory, practice and correlation, 2nd ed. St. Louis: CV Mosby, 1989:290–310.
4. Koch DO, Peters T Jr. Selection and evaluation of methods. In: Burtis CA, Ashwood ER, eds. Tietz textbook of clinical chemistry, 2nd ed. Philadelphia: WB Saunders, 1994:508–25.
5. Linnet K. Necessary sample size for method comparison studies based on regression analysis. Clin Chem 1999;45:882–94. [Abstract/Full Text]
6. Linnet K. Evaluation of regression procedures for methods comparison studies. Clin Chem 1993;39:424–32. [Full Text]
7. Martin RF. General Deming regression for estimating systematic bias and its confidence interval in method-comparison studies. Data supplement. GDR: Executable program for general Deming regression calculations and graphics. [Abstract/Full Text/Data Supplement].
8. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307–10.
9. Altman DG, Bland JM. Commentary on quantifying agreement between two methods of measurement. Clin Chem 2002;48:801-2.[Extract/Full Text]
10. Fraser CG, Petersen PH. Analytical performance characteristics should be judged against objective quality specifications. Clin Chem 1999;45:321–3. [Extract/Full Text]
11. Solberg HE. Establishment and use of reference values. In: Burtis CA, Ashwood ER, eds. Tietz textbook of clinical chemistry, 4th ed. St. Louis: Saunders, 2006:425–48.
12. International Union of Biochemistry and Molecular Biology, Nomenclature Committee. Enzyme nomenclature 1992. San Diego: Academic Press, 1992:862pp.