In drug and alcohol studies performance tests for laboratory applications must meet the usual metric requirements of stability, reliability, and sensitivity. Sensitivity is the ability to detect perceived substances at low levels. However, when such tests are applied for use as interlock devices to prevent persons from driving who may have had alcohol and drug problems previously, emphasis can shift from sensitivity to a concern for specificity (i.e., minimization of false positives). From a series of experiments using a set of performance tests, and by combining data using extended baseline, size of the decrement, and multiple cutoffs, a computer program has been developed which can be used to tune these false positives to very low levels (0-3 percent). The simulation program can be used with any performance battery and application for research into interlock devices will be discussed. (A)
Samenvatting