Table 5.

Results of the comparison of the models for predicting the treatment effect on disability progression from the treatment effect on relapse rate.

		Absolute discrepancy	$w_{{\hat{Y}}_{2j}}/w_{Y_{2j}}$	$w_{{\hat{\mu }}_{2j}^{\mathit{CM}}}/w_{{\hat{\mu }}_{2j}^{\mathit{FEMR}}}$
Model	Prior	Median (range)	Median (range)	Median (range)
FEMR		0.16 (0.01, 1.16)	1.02 (1.00, 1.21)
REMR	I	0.15 (0.01, 1.15)	1.07 (1.00, 1.54)	1.96 (1.36, 2.56)
REMR	II	0.16 (0.01, 1.15)	1.07 (1.01, 1.52)	1.95 (1.34, 2.53)
REMR	III	0.15 (0.01, 1.15)	1.07 (1.01, 1.51)	1.91 (1.36, 2.43)
REMR	IV	0.16 (0.01, 1.15)	1.07 (1.01, 1.51)	1.93 (1.37, 2.58)
Daniels & Hughes	I	0.16 (0.01, 1.15)	1.11 (1.02, 1.50)	2.44 (1.65, 5.14)
Daniels & Hughes	II	0.17 (0.02, 1.16)	1.11 (1.02, 1.56)	2.28 (1.62, 5.78)
Daniels & Hughes	III	0.16 (0.01, 1.15)	1.11 (1.02, 1.59)	2.43 (1.61, 5.15)
Daniels & Hughes	IV	0.16 (0.01, 1.16)	1.11 (1.02, 1.45)	2.43 (1.51, 5.11)
BRMA PNF	I	0.14 (0.02, 1.23)	1.16 (1.02, 1.83)	2.95 (1.95, 4.85)
BRMA PNF	II	0.16 (0.01, 1.23)	1.18 (1.02, 1.73)	2.88 (2.02, 4.68)
BRMA PNF	III	0.15 (0.00, 1.23)	1.11 (1.02, 1.52)	2.26 (1.45, 4.48)
BRMA PNF	IV	0.15 (0.01, 1.24)	1.17 (1.02, 1.86)	2.90 (1.74, 4.92)
BRMA	Wishart A	0.16 (0.00, 1.24)	1.78 (1.10, 4.27)	7.00 (3.48, 10.07)
BRMA	Wishart B	0.13 (0.00, 1.23)	1.23 (1.03, 1.95)	3.28 (2.09, 5.60)

CM: current model in each row.