Validity of a Diagnostic Test Designated by a Single function

Abstract

We have designated the validity of a diagnostic test by a single function S taking into account the values of clinical specificity, sensitivity, and the morbidity rate. Thus, by the use of this function each possible diagnostic test can be characterized and easily differentiated from each other.When integrating function S over all x’s (morbidity rates), a parameter Is is obtained which can provide information on the total diagnostic capacity of a test because it comprises the diagnostic properties of both the positive and the negative test results (p+ and p-) in all possible samples with the morbidity rate (x) ranging from 0 to 1.p+ and p- parameters being a posteriori probabilities were used to denote the diagnostic validity,of the positive and the negative test results respectively. By these parameters diagnostic conclusions could be drawn even from unselected population samples (with x values between 0 and 1) in which sensitivity and specificity parameters have no diagnostic properties.

XXXI. On the sensitivity, specificity, and discrimination of diagnostic tests

“…Subsequent authorsl, 3,11,[13][14][15][16][17] have chosen diverse arrangements of two-letter or one-letter expressions to provide a magnificently creative array of such symbols as a, (3, a, b, R, K, p, f1(p), f2(p), p, I -p, I -b, E, Yj, p+, p-, Pij, cf, Se, In, and P[RIYj. In the absence of any national or international efforts at standardization of this statistical babel, an author newly entering the field is free to choose whatever symbols he wishes.…”

Index for Rating Predictive Accuracy of Screening Tests

“…The indices of sensitivity and specificity were calculated to measure the validity of the screening test. Sensitivity (S) is the ability of the test to give a positive result when a carcinogen is tested [3,5,17,22,23,28]. It is ,:.)…”

“…Predictive value of a positive test (P) is the probability that a substance with a positive test result is a carcinogen [3,5,17,23,28]. It is estimated by the proportion of carcinogens among the substances which are positive to the test, P = a/(a + c).…”

“…The predictive values are related to the prevalence of carcinogens (Jt) by the Bayes's Theorem [3,4,5,17,23,26,28]…”