Table 2.

Diagnostic accuracy results of the comparison between algorithm and all other detection methods in the development and validation studies.

Method			True positive, n		False positive, n		True negative, n		False negative, n		Sensitivity, %		Specificity, %		Positive predictive value, %		Negative predictive value, %
Development study, development data set (n=120)
	Algorithm	11		10		99		0		100		91		52		100
	Manual GTT	9		0		109		2		82		100		100		98
	Incident reporting	7		0		109		4		64		100		100		96
Development study, testing data set (n=120)
	Algorithm	8		9		101		2		80		92		47		98
	Manual GTT	9		0		110		0		100		100		100		100
	Incident reporting	7		0		110		3		70		100		100		97
Validation study (n=298)
	Algorithm	15		17		266		0		100		94		47		100
	Manual chart review	14		0		283		1		93		100		100		99
	ANQb measure	5		0		283		10		33		100		100		97

^aGTT: Global Trigger Tool.

^bANQ: Swiss National Association For Quality Development in Hospitals and Clinics.