Eur J Phys Rehabil Med 2024 Feb 28;60(1):145–153
Table V. —Example of E-value calculation and interpretation in quantitative bias analysis. [31, 32]
Example of E-value calculation Description Observed treatment-outcome association from primary analysis (after adjusting for measured confounders) Risk ratio (RR) = 3.8 E-value calculation using formula E-value = RR + √[RR×(RR‒1)] = 3.8+√[3.8×(3.8‒1] = 7.1 Interpretation • An unmeasured confounder would need to be associated with both the treatment and outcome by a risk ratio of 7.1-fold each (above and beyond measured confounders) to fully explain away the observed risk ratio of RR=3.8
• An unmeasured confounder weakly associated with treatment and outcome (i.e., less than a risk ratio of 7.1-fold each) would not be enough to fully explain away the observed risk ratio of RR=3.8
• Therefore, considerably strong confounding associations (with treatment and outcome) would be needed to fully explain away the observed treatment-outcome association of RR=3.8