Table 1.

Statistical measures summarizing and comparing models A and B. The two risk categories, high and low risk are defined as risk above or below 20%. The four risk categories are defined by risk thresholds at 5%, 10% and 20%. Overall, 10% of subjects in the population have an event. Reclassification percent and net reclassification index compare each of models A and B with no model.

	Model A	Model B
	No risk categories
AUC	0.601	0.705
R-squared	0.012	0.058

	Two risk categories (high and low)
Proportion of subjects
classified as high risk
All subjects	0.009	0.098
Case subjects	0.020	0.258
Control subjects	0.008	0.081
Reclassification percent	0.9%	9.8%
Net reclassification index	0.012	0.177

	Four risk categories
Proportion of subjects
classified as high risk
All subjects	0.009	0.098
Case subjects	0.020	0.258
Control subjects	0.008	0.081
Proportion of subjects
classified as med-high risk
All subjects	0.403	0.275
Case subjects	0.519	0.382
Control subjects	0.390	0.263
Proportion of subjects
classified as med-low risk
All subjects	0.569	0.376
Case subjects	0.453	0.275
Control Subjects	0.582	0.387
Proportion of subjects
classified as low risk
All subjects	0.019	0.251
Case subjects	0.009	0.085
Control Subjects	0.020	0.269
Reclassification percent	59.7%	72.5%
Net reclassification index	0.15	0.47