ABSTRACT
Accuracy in drug screening can be expressed as the probability that a positive (or negative) result is correct. This probability can be calculated from Bayes’ theorem if the performance characteristics of the test and the prevalence of the drug in the subject population are known. The probability that the test result is correct is called the predictive value of the test. The predictive value associated with a test result can be increased by sequential independent tests. To meet the requirements for independence, these tests must be performed on a second aliquot of the sample and must be based on a different physical or chemical property of the drug. The predictive value of a positive result from a single immunoassay screen on random urine specimens from a general workplace population has been reported to range from 0.10 to 0.70 (3). This is often not sufficient evidence for forensic applications.
