Table 3. Predictive Performance of the DOAC Score Compared to HAS-BLED.

Presented is the predictive performance of the DOAC Score compared to HAS-BLED in the development, refinement, and validation cohorts. RE-LY was the development cohort. GARFIELD-AF was used to refine the prediction model. COMBINE-AF and RAMQ were the validation cohorts. C-statistics were compared at one-year. In each cohort, the DOAC Score had stronger predictive performance than the HAS-BLED score.

	N	Number of Events	DOAC Score, C-Statistic	HAS-BLED, C-Statistic	P-value
RE-LY	5,684	386	0.72 (0.71-0.74)	0.60 (0.58-0.62)	<0.001
GARFIELD-AF	12,296	131	0.71 (0.67-0.75)	0.66 (0.62-0.71)	0.025
COMBINE-AF	25,586	692	0.67 (0.64-0.69)	0.63 (0.61-0.65)	<0.001
RAMQ	11,945	258	0.65 (0.61-0.68)	0.58 (0.55-0.62)	<0.001