A SCHEME FOR THE COMPARISON OF QUANTITATIVE METHODS

Journal/Book: THE AMERICAN JOURNAL OF CLINICAL PATHOLOGY - Vol. 43 No. 6 1965 Printed in U.S.A.. 1964/1965;

Abstract: ROY N. BARNETT M.D. Division of Laboratories Norwalk Hospital Norwalk Connecticut Received September 12 1964 *The statistical consultant was Dr. W. J. Youden of the National Bureau of Standards. The scheme described below was developed as part of the program of the Standards Committee of the College of American Pathologists in response to the need for a statistical method by which new diagnostic products could be evaluated. It was found to be so satisfactory for this purpose that it is presented here in a form which permits a more general application. Although the discussion and examples are chosen to reflect a clinical laboratory orientation the scheme is equally applicable to many nonclinical types of quantitative analyses. In brief the scheme uses a number of simple statistical tools by means of which a quantitative method of analysis (test method) can be evaluated in a single laboratory by comparison with a reference method. The reference method may be the one currently employed in the laboratory or a different one thought to be more suitable for some reason. APPLICABILITY The scheme is applicable to most substances present in blond or body fluids. The material for which the analysis is performed must be such that it can be prepared in a reasonably pure form and can be defined quite specifically. For example glucose urea cholesterol and calcium are suitable. Enzyme methods can be compared satisfactorily only if pH of the reaction mixture substrate direction of reaction expression of units etc. are identical and this is rarely the case. If the methodology itself causes different substances to be given the same name as for example protein fractions isolated by different technics the scheme may not be applicable; an the other hand use of the scheme may be useful in defining the problem precisely. schö

Zurück | Weiter