Table 1.

References for Choice of λ Based on Examples from FDA Briefing Documents. λ*_S is the value of λ between 0 and 1 for which the difference between the Synthesis(λ, p) and 95-95 margins would be zero, and λ*_B equals the value of λ for which the difference between the proposed Bias-adjusted and 95-95 margins would be zero.

FDA Example	Outcome Measure	β̂PS,H	v̂ar(β̂ PS,H)	v̂ar(β̂ES)	λ*B	λ*S
Amifostine	Tumor Response	0.16	0.0016	0.0007	0.33	0.43
Doripenem	Risk Difference	0.42	0.023	0.0004	0.16	0.49
Etanercept	Difference in Score	1.28	0.27	0.18	0.57	0.79
Everolimus	Risk Difference	0.25	0.00085	0.0013	0.19	0.20
Lepirudin§	Log Odds Ratio	0.40	0.04	0.011	0.59	0.98
Pemetrexed (TAX317&320)	Log Hazard Ratio	0.29	0.014	0.0097	0.56	0.76
Pemetrexed (TAX317)	Log Hazard Ratio	0.58	0.054	0.0097	0.42	0.76
Raloxifene	Log Relative Risk	0.76	0.033	0.012	0.30	0.41
Telmisartan (HOPE)	Log Hazard Ratio	0.26	0.0025	0.0018	0.28	0.33
Telmisartan (All trials)	Log Hazard Ratio	0.19	0.0034	0.0018	0.41	0.56
Febuxostat	Response Rate Difference	0.38	0.0017	0.0007	0.14	0.16
Ximelagatran (All trials)	Risk Difference	0.038	0.00013	0.088	0.59	0.59
Ximelagatran (All but EAFT)	Risk Difference	0.028	0.000055	0.088	0.51	0.52
Zoledronic acid	Risk Difference	0.13	0.0009	0.0009	0.34	0.40

^§This example comparing lepirudin versus heparin discussed at the May 2, 2000 meeting of the FDA Cardiovascular and Renal Drugs Advisory Committee Meeting was not found in the systematic search due to absense of search terms in the corresponding briefing documents.