Classification rates using the VM training dataset for training and VM testing dataset for testing.

PPV = positive predictive value, NPV = negative predictive value, NA = not applicable, LR = logistic regression. Logistic [0.5] indicates the threshold used for determining model calls in the logistic regression analysis. In this case, all individuals with probabilities ≥0.5 were given a model call of “high-risk”. For MDR 20.6% of the testing dataset could not be classified. The first entry in the table represents the classification rate considering only the individuals that could be classified in the denominator. The number in parentheses gives the classification rate considering the entire testing dataset as the denominator. For example, using MDR, 71 individuals who were actually cases could be classified and of those 51/71 = 71.8% were correctly classified as “high-risk”. Considering all cases that were tested, 51/87 = 58.6% were correctly classified. For the consensus of methods, individuals were called high-risk only if two or more methods classified them as high-risk.