Comparing methods for identifying patients with heart failure using electronic data sources

Table 3 Replication of definitions #1 - #4 in validation set of patients (n = 100)*

#	Heart failure criteria based on claims and laboratory data	Sensitivity (%)	Specificity (%)	*AUC*
1	≥2 heart failure encounters OR any hospital discharge diagnosis of heart failure OR ≥ 1 BNP level ≥ 200 pg/ml	78.46	68.57	0.735
2	≥2 heart failure encounters OR any hospital discharge diagnosis of heart failure	70.77	74.29	0.725
3	≥2 heart failure encounters OR ≥ 1 BNP level ≥ 200 pg/ml	63.08	82.86	0.730
4	≥2 heart failure encounters OR *any hospital discharge diagnosis of heart failure OR*≥ 1 BNP ≥ 500 pg/ml	70.77	71.43	0.711

*Calculated sensitivity, specificity, and area under curve for each claim signature is compared to the Framingham definition of heart failure (i.e., meeting either two major criteria or one major and two minor criteria).

ISSN: 1472-6963