Table 4.

Sensitivity analysis to determine effect of an unmeasured confounder on the observed effect size of radiation therapy^*

Prevalence of unmeasured confounder in patients who did not have the outcome, % (range)		Unadjusted association between unmeasured confounder and outcome	Effect size of radiation therapy after accounting for unmeasured confounder, hazard ratio
			Systematic error			Systematic + random error
RT	No RT	Hazard ratio	2.50%	50%	97.50%	2.50%	50%	97.50%
5(0–10)	70(60–80)	1.92	0.24	0.32	0.44	0.17	0.32	0.60
5(0–10)	50(40–60)	1.75	0.27	0.32	0.40	0.20	0.32	0.54
5(0–10)	85(80–90)	2.07	0.20	0.33	0.51	0.16	0.33	0.66
15(10–20)	70(60–80)	3.23	0.46	0.66	0.92	0.34	0.66	1.22
15(10–20)	50(40–60)	3.14	0.40	0.51	0.66	0.30	0.52	0.89
15(10–20)	85(80–90)	3.32	0.59	0.86	1.21	0.43	0.86	1.60

^*The format of this table is adapted from Smith et al.³⁸