Table 2.

Comparison of the estimated causal odds ratio using the trend-in-trend design and the cohort study method. The population is stratified into five subgroups for the trend-in-trend algorithm. Confounders are sampled independently for each calendar period. CPE indicates cumulative probability of exposure, SD standard deviation.

	Number of Unmeasured Confounders	C-statistic of the CPE Model	Trend-in-Trend Odds Ratio		Cohort Study Odds Ratio
			Mean (% bias)	SD*	Mean (% bias)	SD*
2.5	0	0.62	2.46 (−1.6)	0.0198	2.50 (0.0)	0.0101
2.5	2	0.57	2.45 (−2.0)	0.0207	4.79 (91.6)	0.0104
2.5	4	0.41	2.41 (−3.6)	0.0212	4.91 (96.4)	0.0104
2.0	0	0.62	2.03 (−1.5)	0.0184	2.01 (0.5)	0.0091
2.0	2	0.57	1.94 (−3.0)	0.0191	4.28 (114)	0.0089
2.0	4	0.41	1.93 (−3.5)	0.0177	4.32 (116)	0.0090
1.5	0	0.62	1.53 (2.0)	0.0124	1.51 (0.7)	0.0092
1.5	2	0.57	1.53 (2.0)	0.0132	3.24 (116)	0.0088
1.5	4	0.41	1.46 (−2.7)	0.0129	3.38 (125)	0.0094
1.0	0	0.62	1.02 (2.0)	0.0098	0.99 (−1.0)	0.0087
1.0	2	0.57	1.03 (3.0)	0.0111	2.11 (111)	0.0083
1.0	4	0.41	0.97 (−3.0)	0.0112	2.27 (127)	0.0084