Random effects proportional odds regression model for global impression scale
Arthritis Clinical Trial

The GLIMMIX Procedure

Model Information
Data Set WORK.ARTHRITIS
Response Variable y
Response Distribution Multinomial (ordered)
Link Function Cumulative Logit
Variance Function Default
Variance Matrix Blocked By id
Estimation Technique Maximum Likelihood
Likelihood Approximation Gauss-Hermite Quadrature
Degrees of Freedom Method Containment

Number of Observations Read 1194
Number of Observations Used 1194

Response Profile
Ordered
Value
y Total
Frequency
1 1 86
2 2 332
3 3 478
4 4 240
5 5 58
The GLIMMIX procedure is modeling the probabilities of levels of y having lower Ordered Values in the Response Profile table.

Dimensions
G-side Cov. Parameters 3
Columns in X 7
Columns in Z per Subject 2
Subjects (Blocks in V) 303
Max Obs per Subject 4

Optimization Information
Optimization Technique Dual Quasi-Newton
Parameters in Optimization 10
Lower Boundaries 2
Upper Boundaries 0
Fixed Effects Not Profiled
Starting From GLM estimates
Quadrature Points 30

Iteration History
Iteration Restarts Evaluations Objective
Function
Change Max
Gradient
0 0 4 3047.1334157 . 147.364
1 0 5 3042.023412 5.11000369 138.0259
2 0 5 3035.8819458 6.14146622 137.0961
3 0 5 3033.8780696 2.00387618 138.7859
4 0 5 3031.509098 2.36897166 137.5461
5 0 5 3028.2238329 3.28526508 133.6531
6 0 5 3026.5473423 1.67649054 132.9108
7 0 5 3023.8437136 2.70362870 131.3992
8 0 5 3022.2062355 1.63747816 130.5198
9 0 5 3019.6676259 2.53860956 128.8923
10 0 5 3017.2990126 2.36861331 127.5443
11 0 13 3016.3496584 0.94935421 127.0196
12 0 17 3016.2521494 0.09750897 126.9578
13 0 17 3016.2402195 0.01192987 126.951
14 0 25 3016.2400969 0.00012262 126.9509
15 0 17 3016.2400527 0.00004418 126.9509
16 0 21 3016.2400523 0.00000045 126.9509
17 0 15 3016.2400365 0.00001579 126.9509
18 0 15 3016.2400332 0.00000331 126.9509
19 1 51 3016.2400332 0.00000001 126.9509
20 1 15 3016.2399718 0.00006144 126.9507
21 1 3 2966.6540393 49.58593248 82.45886
22 1 2 2921.8282867 44.82575257 156.6188
23 1 3 2915.5877034 6.24058329 138.8605
24 1 2 2897.3110791 18.27662431 92.39018
25 1 2 2883.0272388 14.28384035 47.78544
26 1 3 2874.5617221 8.46551668 46.67652
27 1 3 2873.5118953 1.04982677 47.13933
28 1 3 2873.2541993 0.25769605 46.47516
29 1 2 2872.8280388 0.42616050 43.59488
30 1 3 2872.7270072 0.10103152 41.22958
31 1 3 2872.6910649 0.03594237 41.83902
32 1 4 2872.5589275 0.13213735 38.01682
33 1 4 2871.2643768 1.29455073 15.9771
34 1 3 2871.1170942 0.14728261 15.16233
35 1 3 2871.0975481 0.01954603 13.85224
36 1 3 2871.0962749 0.00127323 13.76206
37 1 3 2871.0962339 0.00004100 13.88421
38 1 3 2871.0962329 0.00000101 13.88789
39 1 4 2871.0962321 0.00000083 13.88983
40 1 4 2871.096229 0.00000301 13.88234
41 1 6 2871.0960383 0.00019073 13.9745
42 1 2 2871.0958219 0.00021644 13.87216
43 1 2 2871.0957717 0.00005018 13.90266
44 1 4 2871.0954025 0.00036926 13.85584
45 1 8 2870.4699851 0.62541733 8.126829
46 1 3 2870.2810078 0.18897729 10.60846
47 1 3 2870.2729554 0.00805244 10.48051
48 1 6 2869.2271257 1.04582971 5.688627
49 1 3 2869.0240612 0.20306452 3.295322
50 1 3 2868.9990763 0.02498482 2.217486
51 1 4 2868.8763929 0.12268348 0.763875
52 1 3 2868.8705403 0.00585261 0.096985
53 1 3 2868.8704971 0.00004311 0.004328
54 1 3 2868.8704971 0.00000007 0.000227
55 1 3 2868.8704971 0.00000000 5.573E-6

Convergence criterion (GCONV=1E-12) satisfied.

Fit Statistics
-2 Log Likelihood 2868.87
AIC (smaller is better) 2888.87
AICC (smaller is better) 2889.06
BIC (smaller is better) 2926.01
CAIC (smaller is better) 2936.01
HQIC (smaller is better) 2903.73

Fit Statistics for Conditional Distribution
-2 log L(y | r. effects) 1961.62

Estimated G Matrix
Effect Row Col1 Col2
Intercept 1 4.1424 -0.05161
sqrtmonth 2 -0.05161 0.3243

Covariance Parameter Estimates
Cov Parm Subject Estimate Standard Error
UN(1,1) id 4.1424 0.8549
UN(2,1) id -0.05161 0.2626
UN(2,2) id 0.3243 0.1617

Solutions for Fixed Effects
Effect y Estimate Standard Error DF t Value Pr > |t|
Intercept 1 -5.2945 0.3289 301 -16.10 <.0001
Intercept 2 -2.0085 0.2483 301 -8.09 <.0001
Intercept 3 1.0753 0.2382 301 4.51 <.0001
Intercept 4 3.8683 0.3006 301 12.87 <.0001
trt   0.1140 0.3184 586 0.36 0.7205
sqrtmonth   0.4000 0.1018 300 3.93 0.0001
trt*sqrtmonth   0.4005 0.1432 586 2.80 0.0053

Type III Tests of Fixed Effects
Effect Num DF Den DF F Value Pr > F
trt 1 586 0.13 0.7205
sqrtmonth 1 300 15.44 0.0001
trt*sqrtmonth 1 586 7.82 0.0053