Title: Odds Ratio as Measure of Strength of Association in Diagnostic Screening Test
Abstract:I here propose an odds ratio based on false rates as a measure of the strength of association between state of nature or condition in a population and test results in diagnostic screening tests.The me...I here propose an odds ratio based on false rates as a measure of the strength of association between state of nature or condition in a population and test results in diagnostic screening tests.The method provides an estimate for the proposed odds ratio that depends only on the estimated sensitivity and specificity of the test in the event that the prevalence rate is not known.The proposed method unlike the traditional odds ratio provides estimates of not only the proposed false rates based odds ratio-type measure of association, but also alternative sample estimates of its associated standard deviation and test statistic for significance that intrinsically and structurally partials out, that is, does not include in its formulation the number of subjects in the sample known or believed to actually have the condition in nature but test negative or actually do not have the condition in nature but test positive to the condition in the screening test.The proposed method given that the prevalence rate of the condition in the population is known, provides sample estimates of the false positive rate, false negative rate and their odds as well as the proportion of the population expected to test positive to the condition in the screening test which are additional useful information to guide policy formulation and implementation over and above the traditional odds ratio method.Modified estimates of the standard deviation and test statistic for the proposed measure that adjust for the fact that some sample observations in a screening test are not known and cannot therefore validly be used in traditional odds ratio estimation method are provided.The proposed method which is shown to provide more information and to be at least as efficient as the traditional odds ratio method is illustrated with some sample data.Read More