As shown in Table 4b, in Whites, all dichotomous and density measures presented significant diagnostic accuracy at classifying AUD diagnosis. However, AUC of density measures were greater compared to dichotomous measures which were closer to 50% probability/chance. Further, FHDrPSND–AUDMaxSxCnt and FHFather–AUDDx yielded the highest AUC among density and dichotomous measures respectively. On the other hand, in Blacks, except for FHFirst–AUDDx, all other dichotomous measures failed to be significantly different from random chance at classifying AUD diagnosis. In other words, those measures were completely uninformative. On the contrary, in Blacks, all density measures presented significant diagnostic accuracy for classifying AUD diagnosis FHDrPSND–AUDMaxSxCnt presented the highest AUC. Therefore, in both Whites and Blacks, density measures presented consistent and significantly greater diagnostic accuracy for classifying AUD diagnosis, unlike dichotomous measures, which had poor diagnostic accuracy.
