Table 3. Sensitivity Analysis Inverse Probability Weighted Models.

Outcome	Adjusted risk, mean (SE)		Risk reduction, early NIT adjusted risk − control adjusted risk, mean (95% CI)	P value	NNT absolute risk reduction
	No early NIT (control) (n = 62 876)	Early NIT (treated) (n = 16 164)
Death or MIa	0.005 (0.0003)	0.003 (0.0005)	−0.002 (−0.003 to −0.001)b	.001	1/0.002 = 500b
Deathc	0.005 (0.0003)	0.003 (0.0004)	−0.002 (−0.003 to −0.001)b	<.001	1/0.002 = 500b
Acute MIa	0.003 (0.0002)	0.002 (0.0004)	−0.001 (−0.002 to 0.0003)	.22	NAd
Coronary revascularizationa	0.003 (0.0002)	0.003 (0.0004)	0.001 (−0.0002 to 0.002)	.13	NAd
MACEa	0.005 (0.0003)	0.003 (0.0005)	−0.002 (−0.003 to −0.001)b	.001	1/0.002 = 500b

Abbreviations: CAD, coronary artery disease; MACE, major adverse cardiovascular events; MI, myocardial infarction; NIT, noninvasive cardiac testing; NNT, number needed to treat.

^a Doubly robust inverse probability weighting model models with regression adjustment for age, sex, race, smoking, body mass index (calculated as weight in kilograms divided by height in meters squared), self and family history of CAD, initial troponin, antidiabetic medication, anticoagulant medication, antihyperlipidemia medication, antihypertension medication, and Elixhauser comorbidities.

^b Statistically significant differences.

^c Estimate based on inverse probability weighting model without regression adjustment since one or more covariates perfectly predicted death.

^d Difference in event rates are not statistically significant at α = 0.05 and the 95% CI contains zero.