Table 2.

The physician-alone performance and human-algorithm integration (HIA) performance of the physician participants (n=34); the Wilcoxon signed-rank test was used to compare the physician-alone and the HAI performance.

Measures	Physician-alone performance	HAI performance	P valuea
Sensitivity, median (IQR)	0.95 (0.90-1.00)	0.99 (0.96-1.00)	<.001
Specificity, median (IQR)	0.90 (0.82-0.94)	0.95 (0.90-0.98)	<.001
Accuracy, median (IQR)	0.90 (0.88-0.94)	0.96 (0.93-0.98)	<.001
Human-algorithm agreement, κ, median (IQR)	0.69 (0.63-0.74)	0.80 (0.76-0.82)	<.001

^aAll P values are statistically significant.