Accuracy of gestalt perception of acute chest pain in predicting coronary artery disease

Table 2 Accuracy of the 3 classifications of chest pain according to medical judgment

	Sensitivity	Specificity	Positive LR	Negative LR	Positive PV	Negative PV
Classification 1 (2-level)
Typical (vs atypical)	48% (40%-55%)	66% (59%-73%)	1.4 (0.65-2.0)	0.79 (0.62-1.02)	56% (48%-64%)	58% (51%-65%)
Classification 2 (3-level)1
Angina (vs undefined/no-angina)	42% (34%-50%)	69% (62%-76%)	1.35 (0.89-2.1)		56% (47%-65%)
No-angina (vs undefined/angina)	76% (69%-83%)	36% (29%-43%)		0.67 (0.40-1.1)		62% (53%-72%)
Classification 3 (4-level)1
Definitely no-angina (vs the other 3 groups)		27% (20%-34%)		0.63 (0.32-1.15)		70% (59%-81%)

Numbers in parenthesis are 95%CI.

¹Because only no-angina distinguished itself from the other 3 classifications, only specificity, negative LR and negative PV were calculated. LR: Likelihood ratio; PV: Predictive value.