Application of Multivariate Modeling for Radiation Injury Assessment: A Proof of Concept
Table 6
RC3 estimations.
CBC*
CBC-SCHEM†
Day
Estimated Response Category
95% Prediction Limit Width
95% Confidence Interval Limit Width
Day
Estimated Response Category
95%Prediction Limit Width
95%Confidence Interval Limit Width
7
1.73
1.90
0.39
7
2.26
1.76
0.63
2.56
1.89
0.32
2.54
1.70
0.39
2.93
1.90
0.30
2.72
1.88
0.33
2.78
1.90
0.60
2.85
1.67
0.54
3.17
1.89
0.63
3.20
1.68
0.57
2.09
1.88
0.33
2.48
1.99
0.44
2.57
1.89
0.33
2.34
1.68
0.39
3.16
1.93
0.37
3.00
1.84
0.42
10
3.33
1.89
0.35
10
3.16
1.68
0.41
3.31
1.89
0.39
3.22
1.68
0.36
3.50
1.90
0.53
3.48
1.68
0.54
2.24
1.90
0.35
3.08
1.68
0.93
2.78
1.90
0.39
2.67
1.68
0.37
3.18
1.89
0.37
3.06
1.68
0.35
3.48
1.90
0.36
3.47
1.68
0.37
3.48
1.90
0.38
3.43
1.72
0.37
14
3.30
1.89
0.49
14
3.22
1.67
0.46
1.90
1.90
0.37
1.96
1.69
0.33
2.70
1.90
0.37
2.54
1.68
0.36
3.14
1.90
0.36
2.99
1.68
0.36
3.43
1.89
0.37
3.13
1.67
0.41
3.77
1.89
0.38
3.76
1.74
0.36
4.02
1.89
0.46
3.79
1.68
0.46
17
2.00
1.89
0.33
17
2.28
1.69
0.37
2.66
1.90
0.35
2.46
1.68
0.35
3.06
1.90
0.34
2.86
1.69
0.34
2.97
1.89
0.33
2.81
1.67
0.32
2.86
1.89
0.52
2.86
1.68
0.47
3.41
1.89
0.49
3.38
1.67
0.45
1.91
1.89
0.28
2.87
1.72
1.13
21
2.55
1.96
0.32
21
2.40
1.73
0.39
3.10
1.90
0.34
3.27
1.68
0.58
3.28
1.90
0.32
3.45
1.68
0.37
2.87
1.90
0.46
3.04
1.68
0.74
2.25
1.94
0.87
2.51
1.70
0.86
2.14
1.92
0.33
2.40
1.80
0.37
2.73
2.01
0.38
2.48
1.77
0.39
3.06
1.94
0.33
2.77
1.73
0.37
25
3.12
1.94
0.34
25
2.99
1.74
0.34
2.65
1.94
0.75
2.59
1.72
0.68
3.15
1.93
0.55
3.09
1.70
0.52
2.21
1.92
0.48
2.83
1.70
0.84
2.68
1.93
0.36
2.69
1.70
0.54
3.07
20.6
0.33
3.42
1.85
0.53
2.32
1.94
0.55
2.37
1.72
0.55
2.97
1.99
0.69
3.07
1.76
0.63
Mean
2.86
1.91
0.42
Mean
2.90
1.71
0.48
SD
0.52
0.03
0.13
SD
0.42
0.06
0.18
SEM
0.08
0.01
0.02
SEM
0.06
0.01
0.03
The “CBC” RC3 model failed to meet the critical three assumptions of Linear-Least-Squares and was therefore NOT ACCEPTED as statistically sound for estimating H-ARS RC3. The “CBC-SCHEM” RC3 model met the critical three assumptions of Linear-Least-Squares and was therefore ACCEPTED as statistically sound for estimating H-ARS RC3.
