As shown in Table 3b, in males, all four density measures presented significant diagnostic accuracy for classifying AUD diagnosis, with FHDrPSND–AUDMaxSxCnt having the highest AUC. However, among dichotomous measures, FHMother–AUDDx failed to be significantly different from 50% probability at classifying AUD diagnosis. In males, other dichotomous measures, although significant, nevertheless presented diagnostic accuracy closer to 50% probability/chance, indicating poor diagnostic accuracy. Whereas, in females, all dichotomous and density measures presented significant diagnostic accuracy for classifying AUD diagnosis, albeit the dichotomous measures were again closer to 50% probability/chance, indicating poor diagnostic accuracy. Among density measures, FHDrPSND–AUDMaxSxCnt yielded the highest AUC and among dichotomous measures, FHFather–AUDDx, FHParent–AUDDx, and FHFirst–AUDDx presented equally high AUC. Overall, in both males and females, the density measures present consistently significant and higher diagnostic accuracy at classifying DSM-5 AUD diagnosis unlike dichotomous measures, which present high variability and poor diagnostic accuracy.
