S6 File. SAS output Glimmix outcomes men and women above 75 years of age The SAS System 10:18 Tuesday, October 13, 2015 97 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_smoking Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 62099 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 231378.65396 0.88480248 0.000197 1 0 3 277495.2774 0.33756623 0.000599 2 0 3 292342.90418 0.04952533 0.000246 3 0 2 294179.22662 0.01310813 0.000118 4 0 2 294234.49508 0.00154154 6.256E-8 5 0 1 294235.47209 0.00014828 4.23E-6 6 0 1 294235.50796 0.00001400 0.000067 7 0 1 294235.51074 0.00000128 0.000415 8 0 1 294235.51134 0.00000095 0.001679 The SAS System 10:18 Tuesday, October 13, 2015 98 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 294235.50988 0.00000066 0.000189 10 0 1 294235.51091 0.00000041 0.000727 11 0 1 294235.51159 0.00000032 0.00001 12 0 1 294235.51109 0.00000031 0.000684 13 0 1 294235.51058 0.00000000 0.000679 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 294235.5 Generalized Chi-Square 10636.11 Gener. Chi-Square / DF 0.17 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.8235 0.009322 Residual 0.1713 0.001172 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 2.9646 0.3752 19128 7.90 <.0001 Age -0.07329 0.005362 42968 -13.67 <.0001 Year 0.02038 0.005651 42968 3.61 0.0003 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 42968 186.82 <.0001 Year 1 42968 13.01 0.0003 The SAS System 10:18 Tuesday, October 13, 2015 99 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_microalbuminuria Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 48983 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 219931.06925 0.10083235 0.104692 1 0 2 216367.472 0.00835725 0.000086 2 0 2 216421.9529 0.00031333 0.000257 3 0 2 216422.09983 0.00000298 0.000053 4 0 1 216422.10612 0.00000027 0.000973 5 0 1 216422.10602 0.00000022 0.000097 6 0 1 216422.10601 0.00000194 0.007563 7 0 1 216422.10586 0.00000128 0.002483 8 0 1 216422.10587 0.00000121 0.002266 The SAS System 10:18 Tuesday, October 13, 2015 102 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 216422.10602 0.00000062 0.000176 10 0 1 216422.10602 0.00000040 0.001388 11 0 1 216422.10602 0.00000000 0.001397 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 216422.1 Generalized Chi-Square 28921.76 Gener. Chi-Square / DF 0.59 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.4154 0.007706 Residual 0.5905 0.004657 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -7.8417 0.2254 16899 -34.80 <.0001 Age 0.08383 0.002891 32081 29.00 <.0001 Year -0.01090 0.004301 32081 -2.54 0.0112 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 32081 840.76 <.0001 Year 1 32081 6.43 0.0112 The SAS System 10:18 Tuesday, October 13, 2015 103 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_AlbKreat Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 296526.07625 0.09497899 0.001017 1 0 3 294718.22643 0.00872847 0.002358 2 0 2 294713.52641 0.00009715 0.000038 3 0 1 294713.09067 0.00000062 0.003313 4 0 1 294713.08824 0.00000097 0.004857 5 0 1 294713.08467 0.00000140 0.006963 6 0 1 294713.08979 0.00000200 0.009982 7 0 1 294713.08245 0.00000093 0.00204 8 0 1 294713.08587 0.00000060 0.003018 The SAS System 10:18 Tuesday, October 13, 2015 106 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 294713.08808 0.00000001 0.002915 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 294713.1 Generalized Chi-Square 47292.96 Gener. Chi-Square / DF 0.72 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.2473 0.005710 Residual 0.7176 0.004805 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 4.6144 0.1809 20267 25.51 <.0001 Age -0.06944 0.002302 45638 -30.17 <.0001 Year 0.1692 0.003187 45638 53.08 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 910.16 <.0001 Year 1 45638 2817.77 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 107 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_HbA1c Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 290578.35133 0.87266835 0.011099 1 0 3 352995.11915 0.33347102 0.004465 2 0 3 385729.56865 0.08038067 0.012262 3 0 2 391311.70707 0.01818896 0.004066 4 0 2 391280.94043 0.00040328 0.002406 5 0 1 391276.65244 0.00000050 0.000129 6 0 1 391276.63754 0.00000175 0.007527 7 0 2 391276.6236 0.00000175 1.707E-8 8 0 1 391276.63733 0.00000000 0.000025 The SAS System 10:18 Tuesday, October 13, 2015 110 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 391276.6 Generalized Chi-Square 53298.93 Gener. Chi-Square / DF 0.81 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.1112 0.005937 Residual 0.8087 0.006027 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.6696 0.3144 20267 11.67 <.0001 Age -0.03826 0.003889 45638 -9.84 <.0001 Year 0.2001 0.004564 45638 43.84 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 96.77 <.0001 Year 1 45638 1921.56 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 111 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_SBP Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 276359.34934 0.97305733 0.005751 1 0 3 354101.56304 0.50356586 0.043648 2 0 4 400870.6637 0.19901306 0.000303 3 0 3 413864.50081 0.07070400 0.001883 4 0 2 414118.95764 0.00422159 0.0004 5 0 2 414085.28367 0.00003496 0.000165 6 0 2 414085.13446 0.00000134 0.003341 7 0 1 414085.12888 0.00000016 0.003071 8 0 1 414085.12776 0.00000130 0.002769 The SAS System 10:18 Tuesday, October 13, 2015 114 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 414085.11614 0.00000118 0.002523 10 0 1 414085.12673 0.00000000 0.00256 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 414085.1 Generalized Chi-Square 50953.55 Gener. Chi-Square / DF 0.77 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.1838 0.007059 Residual 0.7731 0.005851 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 5.2389 0.3916 20267 13.38 <.0001 Age -0.04835 0.004851 45638 -9.97 <.0001 Year 0.1703 0.005695 45638 29.90 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 99.35 <.0001 Year 1 45638 894.30 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 115 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable Out_Obesity Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 56887 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 207256.31561 0.15195872 0.001613 1 0 2 203748.40504 0.00733365 0.000984 2 0 2 203830.17084 0.00019072 0.000635 3 0 1 203829.65067 0.00000131 0.000017 4 0 1 203829.6598 0.00000002 4.931E-6 5 0 0 203829.65967 0.00000000 3.89E-6 Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 10:18 Tuesday, October 13, 2015 116 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 203829.7 Generalized Chi-Square 13493.94 Gener. Chi-Square / DF 0.24 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7658 0.009276 Residual 0.2372 0.001695 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.8988 0.1971 17869 19.78 <.0001 Age -0.05808 0.002763 39015 -21.02 <.0001 Year 0.01858 0.003179 39015 5.85 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 39015 441.76 <.0001 Year 1 39015 34.18 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 117 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable Out_Overweight Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 56887 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 220709.73877 0.18495255 0.003271 1 0 2 218968.74376 0.00397897 0.000014 2 0 2 218982.10259 0.00001783 0.00006 3 0 1 218982.0974 0.00000006 0.000195 4 0 1 218982.09741 0.00000127 0.006085 5 0 2 218982.09731 0.00000123 0.00003 6 0 1 218982.09741 0.00000019 0.00091 7 0 1 218982.09742 0.00000060 0.002022 8 0 1 218982.09737 0.00000132 0.004494 The SAS System 10:18 Tuesday, October 13, 2015 118 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 218982.09748 0.00000087 0.000211 10 0 1 218982.09741 0.00000001 0.000142 11 0 1 218982.09741 0.00000093 0.004441 12 0 1 218982.09733 0.00000000 0.004441 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 218982.1 Generalized Chi-Square 22697.42 Gener. Chi-Square / DF 0.40 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.6104 0.008353 Residual 0.3990 0.002844 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -0.5504 0.1874 17869 -2.94 0.0033 Age 0.004329 0.002532 39015 1.71 0.0873 Year -0.01329 0.003245 39015 -4.10 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 39015 2.92 0.0873 Year 1 39015 16.78 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 119 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_BMI Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 56887 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 229167.01641 0.53803480 0.004751 1 0 2 225040.29277 0.04181167 0.000576 2 0 2 225157.19981 0.00035612 0.000037 3 0 1 225156.84474 0.00000086 0.000434 4 0 1 225156.84572 0.00000161 0.004797 5 0 2 225156.84614 0.00000147 2.051E-9 6 0 0 225156.84573 0.00000001 2.993E-6 The SAS System 10:18 Tuesday, October 13, 2015 120 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 225156.8 Generalized Chi-Square 15252.51 Gener. Chi-Square / DF 0.27 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7367 0.009171 Residual 0.2681 0.001920 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -6.8529 0.2160 17869 -31.73 <.0001 Age 0.06937 0.002953 39015 23.49 <.0001 Year -0.00177 0.003699 39015 -0.48 0.6332 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 39015 551.65 <.0001 Year 1 39015 0.23 0.6332 The SAS System 10:18 Tuesday, October 13, 2015 121 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_BMI Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 286980.33919 0.33898393 0.00024 1 0 4 302794.84853 0.05505536 0.000147 2 0 2 304630.13041 0.00774586 0.010911 3 0 2 304540.99769 0.00034655 0.000312 4 0 2 304534.53102 0.00001822 0.000011 5 0 1 304534.15526 0.00000062 0.005193 6 0 1 304534.13661 0.00000004 0.005515 7 0 2 304534.13551 0.00000107 0.000265 8 0 1 304534.1287 0.00000020 0.000883 The SAS System 10:18 Tuesday, October 13, 2015 122 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 304534.12984 0.00000024 0.000423 10 0 1 304534.12835 0.00000040 0.001765 11 0 1 304534.13085 0.00000170 0.007501 12 0 1 304534.12015 0.00000001 0.007408 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 304534.1 Generalized Chi-Square 30412.56 Gener. Chi-Square / DF 0.46 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.4722 0.007488 Residual 0.4615 0.003171 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 6.5324 0.2203 20267 29.66 <.0001 Age -0.09078 0.002882 45638 -31.50 <.0001 Year 0.2044 0.003597 45638 56.81 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 992.31 <.0001 Year 1 45638 3227.52 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 123 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_CholHDL Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 59737 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 248200.35184 0.14072523 0.003468 1 0 2 242353.95524 0.00733082 0.008435 2 0 2 242496.00926 0.00015800 2.749E-8 3 0 1 242496.09471 0.00000088 0.00036 4 0 1 242496.09596 0.00000028 0.001346 5 0 1 242496.09615 0.00000014 0.000547 6 0 1 242496.09607 0.00000047 0.002204 7 0 1 242496.09579 0.00000049 0.00067 8 0 1 242496.09608 0.00000006 0.000334 The SAS System 10:18 Tuesday, October 13, 2015 124 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 242496.09605 0.00000001 0.000287 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 242496.1 Generalized Chi-Square 27494.41 Gener. Chi-Square / DF 0.46 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.5392 0.007716 Residual 0.4603 0.003218 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.8368 0.1892 19011 9.71 <.0001 Age -0.02103 0.002512 40723 -8.37 <.0001 Year 0.05175 0.003314 40723 15.62 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 40723 70.13 <.0001 Year 1 40723 243.86 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 125 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_CholHDL Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 299644.12919 0.61506737 0.000263 1 0 2 339233.10024 0.13041975 0.002155 2 0 2 353271.10622 0.01057918 0.00189 3 0 2 354275.38621 0.00113722 1.669E-6 4 0 2 354273.22955 0.00002909 0.000223 5 0 1 354273.1041 0.00000010 0.005089 6 0 1 354273.10618 0.00000000 0.0052 The SAS System 10:18 Tuesday, October 13, 2015 126 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 354273.1 Generalized Chi-Square 50390.37 Gener. Chi-Square / DF 0.76 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.1807 0.005752 Residual 0.7646 0.005326 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 6.4006 0.2471 20267 25.90 <.0001 Age -0.08378 0.003093 45638 -27.09 <.0001 Year 0.2229 0.003958 45638 56.31 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 733.73 <.0001 Year 1 45638 3170.89 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 127 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_FeetExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 34071 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 149920.40827 2.00000000 0.022881 1 0 2 146695.09352 0.21320004 0.000038 2 0 2 146777.90332 0.00469667 3.065E-8 3 0 2 146777.83464 0.00005750 1.053E-9 4 0 1 146777.83243 0.00000231 0.000531 5 0 1 146777.83241 0.00001049 0.001897 6 0 1 146777.8325 0.00000023 0.00192 7 0 1 146777.83249 0.00000001 0.001919 The SAS System 10:18 Tuesday, October 13, 2015 132 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 146777.8 Generalized Chi-Square 18455.80 Gener. Chi-Square / DF 0.54 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.4632 0.009087 Residual 0.5417 0.005436 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -7.0021 0.2426 14431 -28.86 <.0001 Age 0.07396 0.003079 19637 24.02 <.0001 Year 0.000777 0.004877 19637 0.16 0.8735 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 19637 577.08 <.0001 Year 1 19637 0.03 0.8735 The SAS System 10:18 Tuesday, October 13, 2015 133 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_Insulin Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 263913.19994 0.34191140 0.000269 1 0 2 262593.73398 0.04092494 0.000383 2 0 2 261330.78919 0.00513155 0.000079 3 0 2 261511.70585 0.00055440 1.33E-7 4 0 1 261491.39226 0.00006117 4.645E-6 5 0 2 261493.62523 0.00000674 0.00003 6 0 1 261493.37912 0.00000069 0.000617 7 0 1 261493.40595 0.00000013 0.000192 8 0 1 261493.40334 0.00000001 0.00018 The SAS System 10:18 Tuesday, October 13, 2015 136 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 261493.4 Generalized Chi-Square 14007.47 Gener. Chi-Square / DF 0.21 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.8502 0.009493 Residual 0.2125 0.001409 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -6.4417 0.2081 20267 -30.96 <.0001 Age 0.04979 0.002858 45638 17.42 <.0001 Year 0.06971 0.003667 45638 19.01 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 303.49 <.0001 Year 1 45638 361.42 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 137 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_Diet Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 268903.40577 0.35949112 0.003602 1 0 2 263683.04051 0.00942397 0.003156 2 0 2 263760.07836 0.00026305 0.000013 3 0 1 263759.61561 0.00001328 0.002176 4 0 1 263759.58702 0.00000081 0.000531 5 0 1 263759.58566 0.00000018 0.000116 6 0 1 263759.58559 0.00000039 0.001285 7 0 1 263759.58564 0.00000044 0.000288 8 0 1 263759.58559 0.00000003 0.000173 The SAS System 10:18 Tuesday, October 13, 2015 138 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 263759.58557 0.00000059 0.001944 10 0 1 263759.58567 0.00000066 0.000435 11 0 1 263759.58559 0.00000003 0.000329 12 0 1 263759.58557 0.00000000 0.000329 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 263759.6 Generalized Chi-Square 17006.58 Gener. Chi-Square / DF 0.26 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7532 0.008713 Residual 0.2580 0.001709 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 0.5379 0.2111 20267 2.55 0.0108 Age -0.02085 0.002899 45638 -7.19 <.0001 Year -0.02056 0.003264 45638 -6.30 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 51.72 <.0001 Year 1 45638 39.67 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 139 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_DMoralMed Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 238137.92262 0.18584919 0.001089 1 0 3 236486.95551 0.00453861 3.184E-6 2 0 2 236497.45117 0.00004055 0.000023 3 0 1 236497.4273 0.00000012 0.000422 4 0 1 236497.42758 0.00000153 0.005269 5 0 2 236497.42709 0.00000141 4.986E-9 6 0 0 236497.42754 0.00000000 1.571E-6 The SAS System 10:18 Tuesday, October 13, 2015 140 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 236497.4 Generalized Chi-Square 17688.16 Gener. Chi-Square / DF 0.27 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7428 0.008647 Residual 0.2684 0.001776 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.9213 0.1664 20267 11.54 <.0001 Age -0.01670 0.002276 45638 -7.34 <.0001 Year -0.01932 0.002708 45638 -7.14 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 53.85 <.0001 Year 1 45638 50.91 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 141 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_AntihypertensiveMed Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 277137.18601 0.47285284 0.000064 1 0 2 271679.98958 0.02549372 0.00561 2 0 2 271698.54292 0.00060379 0.000014 3 0 2 271696.13459 0.00005288 0.000115 4 0 1 271695.96459 0.00000316 0.002463 5 0 1 271695.95371 0.00000116 0.002699 6 0 1 271695.95561 0.00000035 0.001156 7 0 1 271695.95491 0.00000059 0.001359 8 0 1 271695.95366 0.00000068 0.001576 The SAS System 10:18 Tuesday, October 13, 2015 142 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 271695.95505 0.00000079 0.001827 10 0 1 271695.95343 0.00000020 0.000942 11 0 1 271695.95381 0.00000000 0.000943 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 271696.0 Generalized Chi-Square 22108.01 Gener. Chi-Square / DF 0.34 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.6661 0.008471 Residual 0.3355 0.002250 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -0.9567 0.2055 20267 -4.65 <.0001 Age 0.01430 0.002763 45638 5.18 <.0001 Year 0.08847 0.003125 45638 28.31 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 26.80 <.0001 Year 1 45638 801.29 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 143 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_LipidLoweringMed Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 246353.18033 0.62630648 0.013147 1 0 2 244365.76416 0.00920313 0.002427 2 0 2 244407.97355 0.00049991 6.55E-6 3 0 1 244408.3175 0.00000814 0.001774 4 0 1 244408.3099 0.00000086 0.000852 5 0 1 244408.30984 0.00000033 0.000395 6 0 1 244408.31002 0.00000015 0.000182 7 0 1 244408.30993 0.00000070 0.002511 8 0 1 244408.31036 0.00000096 0.001148 The SAS System 10:18 Tuesday, October 13, 2015 144 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 244408.30979 0.00000044 0.000531 10 0 1 244408.31004 0.00000001 0.000538 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 244408.3 Generalized Chi-Square 18139.85 Gener. Chi-Square / DF 0.28 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7252 0.008529 Residual 0.2752 0.001824 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 6.1665 0.1854 20267 33.26 <.0001 Age -0.1057 0.002596 45638 -40.73 <.0001 Year 0.1943 0.003228 45638 60.19 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 1658.82 <.0001 Year 1 45638 3623.26 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 145 The SAS System 10:00 Thursday, October 22, 2015 21 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_Smoking Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 301180.19935 0.86635334 . 1 0 0 362220.3628 0.32247317 . 2 0 0 390623.72926 0.05016006 . 3 0 0 394869.44974 0.00107706 . 4 0 0 394947.49392 0.00000041 . 5 0 0 394947.51985 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 14 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 394947.5 Generalized Chi-Square 69982.24 Gener. Chi-Square / DF 1.06 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0619 0.005850 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.5346 0.2923 20267 12.09 <.0001 Age -0.03607 0.003569 45638 -10.11 <.0001 Year 0.1872 0.004381 45638 42.74 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 102.12 <.0001 Year 1 45638 1826.46 <.0001 The SAS System 09:52 Thursday, October 15, 2015 17 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_macroalbuminuria Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 39053 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 15 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 167017.93072 0.98367826 . 1 0 0 214708.48357 0.52647023 . 2 0 0 241769.4634 0.14576429 . 3 0 0 249852.38452 0.00968105 . 4 0 0 250451.29035 0.00004605 . 5 0 0 250454.20628 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 18 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 250454.2 Generalized Chi-Square 38897.52 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9961 0.007129 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -10.0529 0.4784 15006 -21.02 <.0001 Age 0.09605 0.005775 24044 16.63 <.0001 Year -0.09188 0.009241 24044 -9.94 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 24044 276.64 <.0001 Year 1 24044 98.85 <.0001 The SAS System 09:52 Thursday, October 15, 2015 19 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_HbA1c Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 62688 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 280730.166 0.14677837 . 1 0 0 274410.66197 0.00604085 . 2 0 0 274561.0638 0.00000839 . 3 0 0 274561.16096 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 20 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 274561.2 Generalized Chi-Square 62696.08 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0002 0.005649 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.6040 0.1498 19587 10.71 <.0001 Age -0.01397 0.001817 43098 -7.69 <.0001 Year 0.02529 0.002837 43098 8.91 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 43098 59.14 <.0001 Year 1 43098 79.46 <.0001 The SAS System 09:52 Thursday, October 15, 2015 21 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_SBP Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 63808 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 275561.57448 0.16565618 . 1 0 0 273102.75251 0.00501856 . 2 0 0 273144.45558 0.00000391 . 3 0 0 273144.48132 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 22 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 273144.5 Generalized Chi-Square 63768.83 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9994 0.005596 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -1.0926 0.1429 19825 -7.65 <.0001 Age -0.00457 0.001725 43980 -2.65 0.0080 Year 0.08267 0.002857 43980 28.93 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 43980 7.03 0.0080 Year 1 43980 837.18 <.0001 The SAS System 09:52 Thursday, October 15, 2015 23 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable OUT_EyeExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 43849 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 14 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 212254.57722 0.67974774 . 1 0 0 229316.73007 0.10303331 . 2 0 0 231671.13517 0.00216893 . 3 0 0 231714.01993 0.00000079 . 4 0 0 231714.03471 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 24 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 231714.0 Generalized Chi-Square 43871.85 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0006 0.006758 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -4.1485 0.2810 15636 -14.76 <.0001 Age 0.03388 0.003371 28210 10.05 <.0001 Year -0.06157 0.006062 28210 -10.16 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28210 100.99 <.0001 Year 1 28210 103.17 <.0001 The SAS System 09:52 Thursday, October 15, 2015 25 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_EyeExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 304367.33668 0.13106269 . 1 0 0 302947.73136 0.00126273 . 2 0 0 302976.41877 0.00000023 . 3 0 0 302976.42433 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 26 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 302976.4 Generalized Chi-Square 65266.42 Gener. Chi-Square / DF 0.99 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9903 0.005455 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.1703 0.1573 20267 7.44 <.0001 Age -0.03680 0.001922 45638 -19.14 <.0001 Year 0.2288 0.003012 45638 75.97 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 366.49 <.0001 Year 1 45638 5771.31 <.0001 The SAS System 09:52 Thursday, October 15, 2015 27 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable PRO_FeetExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Number of Observations Read 65908 Number of Observations Used 65908 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 316463.15653 0.41838243 . 1 0 0 325302.02212 0.01621252 . 2 0 0 326056.75875 0.00007332 . 3 0 0 326059.52652 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 28 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 326059.5 Generalized Chi-Square 65989.78 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0013 0.005516 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.3471 0.1836 20267 18.23 <.0001 Age -0.04071 0.002236 45638 -18.21 <.0001 Year 0.1362 0.003121 45638 43.64 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 45638 331.46 <.0001 Year 1 45638 1904.62 <.0001 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable HbA1c Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Gender 1 1 Number of Observations Read 65908 Number of Observations Used 62688 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 439698.62508 . 2147.355 1 0 3 439489.76391 208.86116757 898.3547 2 0 4 439341.80057 147.96334446 383.6976 3 0 2 439325.13393 16.66663315 108.7624 4 0 2 439323.4095 1.72443439 9.926577 5 0 2 439323.39552 0.01397498 0.289077 6 0 3 439323.39551 0.00001182 0.000742 The SAS System 10:00 Thursday, October 22, 2015 22 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 439323.4 AIC (smaller is better) 439327.4 AICC (smaller is better) 439327.4 BIC (smaller is better) 439343.2 CAIC (smaller is better) 439345.2 HQIC (smaller is better) 439332.6 Generalized Chi-Square 2467273 Gener. Chi-Square / DF 39.36 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 57.7400 0.7832 Residual 39.3599 0.2698 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 41.9813 0.8387 19587 50.05 <.0001 Age 0.1030 0.01120 43098 9.20 <.0001 Year 0.03159 0.01425 43098 2.22 0.0266 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 43098 84.58 <.0001 Year 1 43098 4.92 0.0266 The SAS System 10:00 Thursday, October 22, 2015 23 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable HbA1c Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Gender 1 1 Number of Observations Read 65908 Number of Observations Used 62688 Dimensions R-side Cov. Parameters 2 Columns in X 4 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 439099.23484 . 2104.992 1 0 3 438894.26866 204.96618727 880.3882 2 0 4 438755.72284 138.54581713 387.1197 3 0 2 438738.75054 16.97229558 106.1392 4 0 2 438737.11945 1.63109113 9.726278 5 0 2 438737.10611 0.01334022 0.275202 6 0 3 438737.1061 0.00001066 0.00069 The SAS System 10:00 Thursday, October 22, 2015 24 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 438737.1 AIC (smaller is better) 438741.1 AICC (smaller is better) 438741.1 BIC (smaller is better) 438756.9 CAIC (smaller is better) 438758.9 HQIC (smaller is better) 438746.3 Generalized Chi-Square 2444778 Gener. Chi-Square / DF 39.00 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 57.1295 0.7741 Residual 39.0016 0.2672 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 49.8402 0.8941 19587 55.74 <.0001 Age 0.08775 0.01116 43097 7.86 <.0001 Year -1.2999 0.05620 43097 -23.13 <.0001 YearSQ 0.05987 0.002446 43097 24.48 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 43097 61.79 <.0001 Year 1 43097 534.93 <.0001 YearSQ 1 43097 599.32 <.0001 The SAS System 10:00 Thursday, October 22, 2015 25 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable SBP Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Gender 1 1 Number of Observations Read 65908 Number of Observations Used 63808 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 541338.5013 . 136.4415 1 0 4 541337.66119 0.84010975 4.603251 2 0 2 541337.66021 0.00098470 0.094424 Convergence criterion (GCONV=1E-8) satisfied. The SAS System 10:00 Thursday, October 22, 2015 26 The GLIMMIX Procedure Fit Statistics -2 Res Log Likelihood 541337.7 AIC (smaller is better) 541341.7 AICC (smaller is better) 541341.7 BIC (smaller is better) 541357.5 CAIC (smaller is better) 541359.5 HQIC (smaller is better) 541346.8 Generalized Chi-Square 12780849 Gener. Chi-Square / DF 200.31 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 148.63 2.3417 Residual 200.31 1.3380 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 155.88 1.5592 19825 99.97 <.0001 Age -0.00732 0.02023 43980 -0.36 0.7174 Year -0.9629 0.02727 43980 -35.31 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 43980 0.13 0.7174 Year 1 43980 1246.92 <.0001 The SAS System 10:00 Thursday, October 22, 2015 27 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable BMI Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Gender 1 1 Number of Observations Read 65908 Number of Observations Used 56887 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 271561.73472 . 1112.983 1 0 4 270067.87068 1493.8640409 183.2283 2 0 4 269971.06684 96.80384439 10.56473 3 0 2 269970.6618 0.40503711 1.969802 4 0 2 269970.64698 0.01481550 0.027248 5 0 2 269970.64698 0.00000285 0.000072 The SAS System 10:00 Thursday, October 22, 2015 28 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 269970.6 AIC (smaller is better) 269974.6 AICC (smaller is better) 269974.6 BIC (smaller is better) 269990.5 CAIC (smaller is better) 269992.5 HQIC (smaller is better) 269979.8 Generalized Chi-Square 140462.2 Gener. Chi-Square / DF 2.47 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 22.6718 0.2532 Residual 2.4693 0.01769 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 42.3072 0.4376 17869 96.69 <.0001 Age -0.1739 0.006293 39015 -27.64 <.0001 Year 0.04567 0.006845 39015 6.67 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 39015 764.14 <.0001 Year 1 39015 44.51 <.0001 The SAS System 10:00 Thursday, October 22, 2015 29 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_FEMALE_OVER_75 Response Variable CholHDL Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 20268 not printed Gender 1 1 Number of Observations Read 65908 Number of Observations Used 59737 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 20268 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 153759.48472 . 1395.736 1 0 3 153395.42199 364.06273503 135.5653 2 0 2 153389.72086 5.70112776 46.86923 3 0 2 153388.99018 0.73067633 2.681173 4 0 2 153388.98785 0.00233689 0.04947 5 0 2 153388.98785 0.00000081 0.000051 The SAS System 10:00 Thursday, October 22, 2015 30 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 153389.0 AIC (smaller is better) 153393.0 AICC (smaller is better) 153393.0 BIC (smaller is better) 153408.8 CAIC (smaller is better) 153410.8 HQIC (smaller is better) 153398.2 Generalized Chi-Square 24219.20 Gener. Chi-Square / DF 0.41 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.9882 0.01216 Residual 0.4055 0.002841 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 2.9940 0.1033 19011 28.97 <.0001 Age 0.01277 0.001418 40723 9.00 <.0001 Year -0.03315 0.001748 40723 -18.97 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 40723 81.08 <.0001 Year 1 40723 359.86 <.0001 The SAS System 10:00 Thursday, October 22, 2015 31 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_smoking Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 39590 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 157075.54369 0.64962781 0.000044 1 0 2 166266.69292 0.09229658 0.000198 2 0 2 167272.89185 0.01005212 0.000655 3 0 2 167301.12009 0.00062975 1.177E-7 4 0 2 167301.78191 0.00003287 0.000018 5 0 1 167301.80751 0.00000167 0.001643 6 0 1 167301.80927 0.00000117 0.000072 7 0 0 167301.80879 . 0.000063 8 0 7 167301.80879 2.00000000 0.000049 The SAS System 10:18 Tuesday, October 13, 2015 146 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 167301.80879 0.00000106 0.00159 10 0 1 167301.80932 0.00000115 0.00019 11 0 1 167301.80875 0.00000040 0.000437 12 0 1 167301.80894 0.00000000 0.000435 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 167301.8 Generalized Chi-Square 7048.35 Gener. Chi-Square / DF 0.18 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.8172 0.01126 Residual 0.1780 0.001548 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.7156 0.3625 13058 4.73 <.0001 Age -0.04060 0.005139 26529 -7.90 <.0001 Year -0.03804 0.005328 26529 -7.14 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 26529 62.42 <.0001 Year 1 26529 50.99 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 147 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_Smoking Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 183202.43002 0.75590327 0.000038 1 0 4 215733.18693 0.19732734 0.000646 2 0 4 227687.32457 0.06690881 0.000086 3 0 2 228261.30105 0.01067037 0.008509 4 0 2 228149.13305 0.00082673 5.528E-6 5 0 2 228139.36676 0.00006735 0.00205 6 0 2 228138.56387 0.00000469 1.71E-10 7 0 1 228138.50309 0.00000032 0.00074 8 0 1 228138.49959 0.00000110 0.002199 The SAS System 10:18 Tuesday, October 13, 2015 148 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 228138.49265 0.00000065 0.000355 10 0 1 228138.49648 0.00000015 0.000012 11 0 0 228138.49745 0.00000000 1.79E-6 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 228138.5 Generalized Chi-Square 22622.91 Gener. Chi-Square / DF 0.54 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.4014 0.01038 Residual 0.5406 0.005120 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 2.3314 0.4167 13718 5.59 <.0001 Age -0.02547 0.005383 28129 -4.73 <.0001 Year 0.1884 0.005902 28129 31.92 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 22.39 <.0001 Year 1 28129 1019.09 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 149 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_macroalbuminuria Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 22294 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 12 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 86372.903857 0.59080999 0.000126 1 0 4 93871.357062 0.07211310 0.000078 2 0 2 95179.916194 0.02034464 0.003035 3 0 2 95178.459831 0.00324534 1.061E-8 4 0 1 95171.95486 0.00044950 1.411E-6 5 0 0 95171.011832 0.00006205 2.893E-6 6 0 0 95170.881296 0.00000856 3.646E-6 7 0 0 95170.863276 0.00000118 3.759E-6 8 0 0 95170.86079 0.00000016 3.775E-6 The SAS System 10:18 Tuesday, October 13, 2015 152 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 0 95170.860447 0.00000002 3.777E-6 10 0 0 95170.860399 0.00000000 3.777E-6 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 95170.86 Generalized Chi-Square 2746.66 Gener. Chi-Square / DF 0.12 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.8612 0.01391 Residual 0.1232 0.001551 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -9.1255 0.4212 9395 -21.66 <.0001 Age 0.09453 0.005867 12896 16.11 <.0001 Year -0.04264 0.006983 12896 -6.11 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 12896 259.56 <.0001 Year 1 12896 37.28 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 153 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_AlbKreat Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 191944.48495 0.20385583 0.000573 1 0 3 193353.34797 0.02821167 0.000281 2 0 2 193349.10144 0.00092589 0.001797 3 0 2 193345.99369 0.00001980 0.000026 4 0 1 193345.93189 0.00000007 0.00239 5 0 1 193345.93173 0.00000067 0.000704 6 0 1 193345.92978 0.00000019 0.000197 7 0 1 193345.93032 0.00000053 0.002266 8 0 1 193345.9288 0.00000062 0.000649 The SAS System 10:18 Tuesday, October 13, 2015 154 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 193345.9306 0.00000174 0.007479 10 0 1 193345.92556 0.00000205 0.002143 11 0 1 193345.93151 0.00000000 0.002076 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 193345.9 Generalized Chi-Square 29546.51 Gener. Chi-Square / DF 0.71 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.2553 0.007647 Residual 0.7061 0.006155 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.4830 0.2598 13718 13.40 <.0001 Age -0.05543 0.003326 28129 -16.67 <.0001 Year 0.1836 0.004362 28129 42.08 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 277.83 <.0001 Year 1 28129 1770.84 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 155 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_HbA1c Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 182926.56617 0.88478570 0.001303 1 0 3 224442.5668 0.33875097 0.011983 2 0 3 245838.70503 0.16536692 0.003333 3 0 3 249025.59557 0.03327851 0.000281 4 0 2 248863.66817 0.00044135 0.000789 5 0 2 248859.99994 0.00000455 0.000771 6 0 1 248859.95261 0.00000003 0.000441 7 0 1 248859.95237 0.00000021 0.00004 8 0 1 248859.95365 0.00000219 0.004808 The SAS System 10:18 Tuesday, October 13, 2015 158 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 248859.9404 0.00000239 0.000446 10 0 1 248859.95488 0.00000024 0.000046 11 0 1 248859.95339 0.00000250 0.005499 12 0 2 248859.96855 0.00000250 2.282E-9 13 0 1 248859.95338 0.00000257 0.005656 14 0 2 248859.96898 0.00000281 0.000519 15 0 1 248859.95195 0.00000000 0.000571 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 248860.0 Generalized Chi-Square 31540.40 Gener. Chi-Square / DF 0.75 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.1532 0.009108 Residual 0.7537 0.007710 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 2.4825 0.4526 13718 5.48 <.0001 Age -0.02420 0.005690 28129 -4.25 <.0001 Year 0.2010 0.006237 28129 32.24 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 18.09 <.0001 Year 1 28129 1039.19 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 159 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_SBP Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 40697 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 168585.86657 0.18420967 0.000247 1 0 2 168107.60741 0.00548771 0.001294 2 0 1 168116.31202 0.00006257 1.19E-6 3 0 1 168116.26943 0.00000040 0.000164 4 0 1 168116.26959 0.00000036 0.001991 5 0 1 168116.26955 0.00000043 0.00059 6 0 1 168116.26959 0.00000128 0.007055 7 0 2 168116.26948 0.00000153 0.002069 8 0 1 168116.26961 0.00000045 0.000613 The SAS System 10:18 Tuesday, October 13, 2015 160 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 168116.26957 0.00000013 0.000182 10 0 1 168116.26959 0.00000040 0.002173 11 0 1 168116.26955 0.00000005 0.001891 12 0 1 168116.26956 0.00000000 0.001867 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 168116.3 Generalized Chi-Square 28034.29 Gener. Chi-Square / DF 0.69 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.3167 0.007420 Residual 0.6889 0.005801 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -2.1688 0.2232 13441 -9.72 <.0001 Age 0.01426 0.002869 27253 4.97 <.0001 Year 0.07071 0.004000 27253 17.68 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 27253 24.71 <.0001 Year 1 27253 312.45 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 161 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_SBP Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 170905.01215 0.98586337 0.00066 1 0 4 222574.88691 0.53372669 0.018249 2 0 4 254217.2269 0.18043627 0.034239 3 0 2 263828.57082 0.06553800 0.183293 4 0 2 264260.33323 0.00548620 0.000181 5 0 2 264237.7979 0.00014327 0.000242 6 0 2 264237.22205 0.00000408 7.011E-9 7 0 1 264237.20604 0.00000103 0.002633 8 0 1 264237.20001 0.00000079 0.000309 The SAS System 10:18 Tuesday, October 13, 2015 162 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 264237.20484 0.00000001 0.000285 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 264237.2 Generalized Chi-Square 28831.72 Gener. Chi-Square / DF 0.69 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.2512 0.009349 Residual 0.6890 0.006630 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.5695 0.5762 13718 6.19 <.0001 Age -0.02851 0.007293 28129 -3.91 <.0001 Year 0.1759 0.007892 28129 22.29 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 15.28 <.0001 Year 1 28129 496.82 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 163 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable Out_Obesity Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 37696 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 150155.65971 0.19759067 0.003232 1 0 2 146648.17741 0.00721957 0.000727 2 0 2 146736.77017 0.00023910 2.078E-7 3 0 1 146736.94656 0.00000119 0.000024 4 0 1 146736.94604 0.00000014 0.000247 5 0 1 146736.94604 0.00000079 0.001356 6 0 1 146736.94608 0.00000073 0.000113 7 0 1 146736.94604 0.00000036 0.000619 8 0 1 146736.94606 0.00000000 0.000619 The SAS System 10:18 Tuesday, October 13, 2015 164 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 146736.9 Generalized Chi-Square 9192.39 Gener. Chi-Square / DF 0.24 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7608 0.01114 Residual 0.2439 0.002170 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 2.6173 0.2946 12504 8.89 <.0001 Age -0.05307 0.004179 25189 -12.70 <.0001 Year 0.03133 0.004766 25189 6.57 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 25189 161.26 <.0001 Year 1 25189 43.22 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 165 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable Out_Overweight Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 37696 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 143009.85941 0.20343510 1.557E-6 1 0 2 142965.26132 0.00352813 2.429E-6 2 0 1 142965.41795 0.00001511 0.000933 3 0 1 142965.41746 0.00000055 0.000756 4 0 1 142965.41749 0.00000044 0.00061 5 0 1 142965.41747 0.00000036 0.000493 6 0 1 142965.41748 0.00000000 0.000493 The SAS System 10:18 Tuesday, October 13, 2015 166 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 142965.4 Generalized Chi-Square 14058.53 Gener. Chi-Square / DF 0.37 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.6357 0.01028 Residual 0.3730 0.003311 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.1937 0.2401 12504 4.97 <.0001 Age -0.01245 0.003283 25189 -3.79 0.0001 Year -0.01565 0.004077 25189 -3.84 0.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 25189 14.38 0.0001 Year 1 25189 14.73 0.0001 The SAS System 10:18 Tuesday, October 13, 2015 167 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_BMI Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 37696 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 149417.06928 0.26633776 0.002577 1 0 2 145419.20114 0.02830149 0.000238 2 0 2 145524.37081 0.00026324 0.000028 3 0 1 145524.37741 0.00000506 0.000794 4 0 1 145524.37692 0.00000053 0.000039 5 0 1 145524.3769 0.00000000 0.000038 Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 10:18 Tuesday, October 13, 2015 168 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 145524.4 Generalized Chi-Square 10355.11 Gener. Chi-Square / DF 0.27 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7309 0.01097 Residual 0.2747 0.002447 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -5.8748 0.2626 12504 -22.37 <.0001 Age 0.06142 0.003636 25189 16.89 <.0001 Year -0.00341 0.004450 25189 -0.77 0.4439 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 25189 285.34 <.0001 Year 1 25189 0.59 0.4439 The SAS System 10:18 Tuesday, October 13, 2015 169 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_BMI Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 183852.41305 0.46854182 0.000755 1 0 4 200454.79201 0.11170015 0.000799 2 0 2 203428.42863 0.02468655 0.0192 3 0 2 203246.79054 0.00180560 0.000018 4 0 2 203223.15943 0.00011021 0.000082 5 0 1 203221.56824 0.00000720 3.607E-6 6 0 1 203221.46289 0.00000020 0.001513 7 0 1 203221.45836 0.00000160 0.003432 8 0 1 203221.44975 0.00000030 0.002413 The SAS System 10:18 Tuesday, October 13, 2015 170 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 203221.45113 0.00000244 0.005217 10 0 1 203221.46387 0.00000135 0.000891 11 0 1 203221.45711 0.00000000 0.000943 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 203221.5 Generalized Chi-Square 19161.82 Gener. Chi-Square / DF 0.46 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.4550 0.009468 Residual 0.4579 0.004109 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 4.6011 0.3178 13718 14.48 <.0001 Age -0.06795 0.004172 28129 -16.29 <.0001 Year 0.2270 0.004995 28129 45.44 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 265.27 <.0001 Year 1 28129 2064.94 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 171 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_CholHDL Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 38721 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 157363.36688 0.15992530 0.014712 1 0 2 154516.07183 0.01395984 0.003241 2 0 2 154566.45332 0.00015538 1.378E-8 3 0 1 154566.56295 0.00000197 0.001791 4 0 1 154566.56402 0.00000051 0.000128 5 0 1 154566.56415 0.00000000 0.000128 Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 10:18 Tuesday, October 13, 2015 172 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 154566.6 Generalized Chi-Square 17122.92 Gener. Chi-Square / DF 0.44 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.5583 0.009551 Residual 0.4422 0.003885 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 0.4373 0.2448 13012 1.79 0.0741 Age -0.00666 0.003286 25706 -2.03 0.0427 Year 0.05535 0.004175 25706 13.26 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 25706 4.11 0.0427 Year 1 25706 175.79 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 173 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_CholHDL Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 188970.182 0.69199078 0.003413 1 0 3 218264.21201 0.16564590 0.000674 2 0 3 229798.06237 0.02625656 0.000251 3 0 2 230785.9289 0.00307735 0.000204 4 0 2 230777.20933 0.00006249 0.000056 5 0 2 230776.9487 0.00000221 0.001971 6 0 1 230776.9397 0.00000068 0.000651 7 0 1 230776.94315 0.00000249 0.008264 8 0 2 230776.93071 0.00000315 0.003033 The SAS System 10:18 Tuesday, October 13, 2015 174 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 230776.94644 0.00000033 0.001815 10 0 1 230776.94478 0.00000093 0.001511 11 0 1 230776.94014 0.00000000 0.001501 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 230776.9 Generalized Chi-Square 30987.38 Gener. Chi-Square / DF 0.74 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.1935 0.007897 Residual 0.7405 0.006799 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 5.1002 0.3593 13718 14.20 <.0001 Age -0.06697 0.004535 28129 -14.77 <.0001 Year 0.2234 0.005462 28129 40.90 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 218.11 <.0001 Year 1 28129 1672.47 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 175 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_FeetExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 22532 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 14 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 4 97507.025013 0.20939481 0.000107 1 0 2 95334.761097 0.03467174 0.003771 2 0 2 95398.448872 0.00094872 0.000163 3 0 2 95398.373224 0.00000356 1.77E-10 4 0 1 95398.367577 0.00000179 0.001846 5 0 1 95398.367455 0.00000402 0.002655 6 0 1 95398.36783 0.00000229 0.000066 7 0 1 95398.367594 0.00000107 0.001111 8 0 1 95398.367515 0.00000242 0.001598 The SAS System 10:18 Tuesday, October 13, 2015 180 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 95398.367741 0.00000003 0.001616 10 0 1 95398.367726 0.00000098 0.000518 11 0 1 95398.367637 0.00000112 0.000728 12 0 1 95398.367543 0.00000001 0.000714 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 95398.37 Generalized Chi-Square 11358.43 Gener. Chi-Square / DF 0.50 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.5043 0.01132 Residual 0.5042 0.006342 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -6.8289 0.3184 10084 -21.45 <.0001 Age 0.07592 0.004090 12445 18.56 <.0001 Year -0.00797 0.006226 12445 -1.28 0.2007 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 12445 344.58 <.0001 Year 1 12445 1.64 0.2007 The SAS System 10:18 Tuesday, October 13, 2015 181 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_FeetExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 195114.58851 0.39913987 0.000133 1 0 2 201726.50364 0.01725434 0.069822 2 0 2 202189.22746 0.00118205 0.000036 3 0 2 202180.10195 0.00004403 8.429E-9 4 0 2 202179.75121 0.00000149 0.000027 5 0 1 202179.7394 0.00000004 0.000167 6 0 1 202179.73915 0.00000041 0.001571 7 0 1 202179.73776 0.00000037 6.924E-6 8 0 0 202179.739 0.00000000 7.634E-7 The SAS System 10:18 Tuesday, October 13, 2015 182 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 202179.7 Generalized Chi-Square 28253.76 Gener. Chi-Square / DF 0.68 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.2842 0.008192 Residual 0.6752 0.005982 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.6822 0.2980 13718 5.65 <.0001 Age -0.02194 0.003810 28129 -5.76 <.0001 Year 0.1458 0.004632 28129 31.48 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 33.17 <.0001 Year 1 28129 990.82 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 183 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_Insulin Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 170225.27995 0.44711840 0.000286 1 0 2 173904.20655 0.05447038 0.000081 2 0 2 172663.5763 0.00777389 0.000187 3 0 2 172818.22346 0.00097029 0.000422 4 0 1 172798.24624 0.00012334 5.536E-6 5 0 2 172800.77401 0.00001564 6.83E-10 6 0 1 172800.45329 0.00000197 0.000175 7 0 1 172800.49392 0.00000023 0.000012 8 0 1 172800.48889 0.00000027 0.000467 The SAS System 10:18 Tuesday, October 13, 2015 184 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 172800.48965 0.00000113 0.001518 10 0 1 172800.48862 0.00000090 0.000069 11 0 1 172800.49002 0.00000003 0.000043 12 0 1 172800.48932 0.00000096 0.001648 13 0 1 172800.48875 0.00000030 0.001113 14 0 1 172800.48958 0.00000081 0.000311 15 0 1 172800.48987 0.00000075 0.001007 16 0 1 172800.48885 0.00000003 0.001002 17 0 1 172800.48941 0.00000018 0.000676 18 0 1 172800.48946 0.00000000 0.000672 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 172800.5 Generalized Chi-Square 8680.34 Gener. Chi-Square / DF 0.21 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.8629 0.01171 Residual 0.2074 0.001751 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -7.2435 0.2905 13718 -24.93 <.0001 Age 0.05612 0.004053 28129 13.85 <.0001 Year 0.07871 0.005257 28129 14.97 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 191.80 <.0001 Year 1 28129 224.17 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 185 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_Diet Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 169001.75224 0.39169755 0.000608 1 0 2 165919.45762 0.01240292 0.000331 2 0 2 165934.66892 0.00028193 2.462E-7 3 0 1 165934.85097 0.00000215 8.371E-8 4 0 1 165934.848 0.00000111 0.002242 5 0 1 165934.84793 0.00000114 0.00011 6 0 1 165934.84787 0.00000030 0.000517 7 0 1 165934.84785 0.00000023 0.000034 8 0 1 165934.84784 0.00000000 0.000035 The SAS System 10:18 Tuesday, October 13, 2015 186 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 165934.8 Generalized Chi-Square 9564.45 Gener. Chi-Square / DF 0.23 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7838 0.01087 Residual 0.2286 0.001929 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 0.8931 0.2820 13718 3.17 0.0015 Age -0.02305 0.003935 28129 -5.86 <.0001 Year -0.03688 0.004250 28129 -8.68 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 34.30 <.0001 Year 1 28129 75.32 <.0001 The SAS System 10:18 Tuesday, October 13, 2015 187 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_DMoralMed Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 2 152869.15883 0.17919969 0.000496 1 0 2 150971.60534 0.01068818 0.000012 2 0 2 150992.59463 0.00009954 0.000011 3 0 1 150992.58995 0.00000061 0.000055 4 0 1 150992.5901 0.00000020 0.000405 5 0 1 150992.59004 0.00000014 0.000089 6 0 1 150992.59008 0.00000030 0.000605 7 0 1 150992.59017 0.00000051 0.000565 8 0 1 150992.59002 0.00000021 0.000079 The SAS System 10:18 Tuesday, October 13, 2015 188 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 150992.59008 0.00000000 0.000079 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 150992.6 Generalized Chi-Square 10670.12 Gener. Chi-Square / DF 0.25 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7521 0.01063 Residual 0.2550 0.002151 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.6896 0.2231 13718 7.57 <.0001 Age -0.01405 0.003096 28129 -4.54 <.0001 Year -0.00795 0.003607 28129 -2.20 0.0275 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 20.60 <.0001 Year 1 28129 4.86 0.0275 The SAS System 10:18 Tuesday, October 13, 2015 189 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_LipidLoweringMed Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 3 157643.76492 0.11111306 0.004017 1 0 3 155525.15337 0.01538304 0.000311 2 0 2 155579.3807 0.00035795 5.038E-7 3 0 1 155579.25222 0.00000221 0.000229 4 0 1 155579.25691 0.00000003 0.000257 5 0 1 155579.25671 0.00000010 0.000016 6 0 1 155579.25678 0.00000053 0.001211 7 0 1 155579.25706 0.00000050 0.000048 8 0 1 155579.25679 0.00000149 0.003425 The SAS System 10:18 Tuesday, October 13, 2015 192 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 9 0 1 155579.25599 0.00000147 0.000024 10 0 0 155579.25677 0.00000000 7.514E-6 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 155579.3 Generalized Chi-Square 10941.17 Gener. Chi-Square / DF 0.26 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.7459 0.01066 Residual 0.2615 0.002211 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 5.2305 0.2419 13718 21.62 <.0001 Age -0.09267 0.003409 28129 -27.18 <.0001 Year 0.1997 0.004120 28129 48.47 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 738.82 <.0001 Year 1 28129 2349.33 <.0001 The SAS System 09:52 Thursday, October 15, 2015 1 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_microalbuminuria Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 32100 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 15 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 140719.35849 0.15913243 . 1 0 0 138594.47613 0.00558520 . 2 0 0 138638.14857 0.00000383 . 3 0 0 138638.17145 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 2 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 138638.2 Generalized Chi-Square 32094.86 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9999 0.007893 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -4.6899 0.2304 11572 -20.36 <.0001 Age 0.05808 0.002820 20525 20.60 <.0001 Year -0.03923 0.004394 20525 -8.93 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 20525 424.23 <.0001 Year 1 20525 79.73 <.0001 The SAS System 09:52 Thursday, October 15, 2015 5 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_HbA1c Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 39993 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 179808.94283 0.14116118 . 1 0 0 175640.31843 0.00555748 . 2 0 0 175744.06833 0.00000807 . 3 0 0 175744.14293 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 6 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 175744.1 Generalized Chi-Square 40003.25 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0003 0.007074 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 2.2838 0.2068 13278 11.05 <.0001 Age -0.02327 0.002528 26712 -9.21 <.0001 Year 0.03128 0.003774 26712 8.29 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 26712 84.74 <.0001 Year 1 26712 68.72 <.0001 The SAS System 09:52 Thursday, October 15, 2015 7 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_EyeExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 29179 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 13 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 139697.40129 0.78031561 . 1 0 0 156720.14373 0.18892577 . 2 0 0 160005.3965 0.00889498 . 3 0 0 160114.34278 0.00001287 . 4 0 0 160114.47194 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 8 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 160114.5 Generalized Chi-Square 29202.01 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0009 0.008287 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept -4.0994 0.4219 10883 -9.72 <.0001 Age 0.02642 0.005092 18293 5.19 <.0001 Year -0.03646 0.009553 18293 -3.82 0.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 18293 26.92 <.0001 Year 1 18293 14.57 0.0001 The SAS System 09:52 Thursday, October 15, 2015 9 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable PRO_EyeExamination Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 195546.40555 0.20659701 . 1 0 0 198251.89497 0.00369055 . 2 0 0 198355.47714 0.00000322 . 3 0 0 198355.58399 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 10 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 198355.6 Generalized Chi-Square 42423.30 Gener. Chi-Square / DF 1.01 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0138 0.007008 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 0.2166 0.2298 13718 0.94 0.3458 Age -0.03314 0.002830 28129 -11.71 <.0001 Year 0.2855 0.004344 28129 65.72 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 137.16 <.0001 Year 1 28129 4319.69 <.0001 The SAS System 09:52 Thursday, October 15, 2015 11 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable OUT_AntihypertensiveMed Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Residual PL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Number of Observations Read 41850 Number of Observations Used 41850 Dimensions R-side Cov. Parameters 1 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique None Parameters 0 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 0 0 0 198338.53471 0.22208409 . 1 0 0 194848.45373 0.00755154 . 2 0 0 194922.40953 0.00000662 . 3 0 0 194922.45061 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 09:52 Thursday, October 15, 2015 12 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 194922.5 Generalized Chi-Square 41910.10 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0015 0.006924 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 1.1722 0.2295 13718 5.11 <.0001 Age -0.00856 0.002816 28129 -3.04 0.0024 Year 0.06690 0.003891 28129 17.19 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 28129 9.24 0.0024 Year 1 28129 295.57 <.0001 The SAS System 09:52 Thursday, October 15, 2015 13 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable HbA1c Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Gender 1 0 Number of Observations Read 41850 Number of Observations Used 39993 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 282664.6042 . 998.3384 1 0 3 282568.12524 96.47896319 277.4913 2 0 2 282554.84368 13.28156381 73.852 3 0 2 282553.94526 0.89841468 8.122946 4 0 2 282553.93468 0.01058368 0.208977 5 0 2 282553.93467 0.00000704 0.000573 The SAS System 10:00 Thursday, October 22, 2015 32 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 282553.9 AIC (smaller is better) 282557.9 AICC (smaller is better) 282557.9 BIC (smaller is better) 282573.0 CAIC (smaller is better) 282575.0 HQIC (smaller is better) 282563.0 Generalized Chi-Square 1714570 Gener. Chi-Square / DF 42.87 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 52.2084 0.9048 Residual 42.8750 0.3719 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 38.2408 1.1108 13278 34.42 <.0001 Age 0.1579 0.01485 26712 10.63 <.0001 Year -0.02413 0.01881 26712 -1.28 0.1996 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 26712 113.05 <.0001 Year 1 26712 1.65 0.1996 The SAS System 10:00 Thursday, October 22, 2015 33 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable HbA1c Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Gender 1 0 Number of Observations Read 41850 Number of Observations Used 39993 Dimensions R-side Cov. Parameters 2 Columns in X 4 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 282400.11518 . 989.8886 1 0 3 282305.05281 95.06236447 274.352 2 0 2 282292.1144 12.93841732 72.3916 3 0 2 282291.25376 0.86064230 7.838183 4 0 2 282291.24392 0.00983049 0.197194 5 0 2 282291.24392 0.00000625 0.000521 The SAS System 10:00 Thursday, October 22, 2015 34 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 282291.2 AIC (smaller is better) 282295.2 AICC (smaller is better) 282295.2 BIC (smaller is better) 282310.3 CAIC (smaller is better) 282312.3 HQIC (smaller is better) 282300.3 Generalized Chi-Square 1703171 Gener. Chi-Square / DF 42.59 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 51.8335 0.8978 Residual 42.5910 0.3694 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 46.5259 1.2152 13278 38.29 <.0001 Age 0.1405 0.01484 26711 9.47 <.0001 Year -1.3500 0.08239 26711 -16.38 <.0001 YearSQ 0.05804 0.003512 26711 16.53 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 26711 89.60 <.0001 Year 1 26711 268.47 <.0001 YearSQ 1 26711 273.08 <.0001 The SAS System 10:00 Thursday, October 22, 2015 35 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable SBP Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Gender 1 0 Number of Observations Read 41850 Number of Observations Used 40697 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 342133.24072 . 851.469 1 0 3 342096.88093 36.35979110 378.6602 2 0 2 342082.11152 14.76940465 74.62697 3 0 2 342081.58944 0.52208615 8.672136 4 0 2 342081.5825 0.00693323 0.173628 5 0 2 342081.5825 0.00000279 0.000394 The SAS System 10:00 Thursday, October 22, 2015 36 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 342081.6 AIC (smaller is better) 342085.6 AICC (smaller is better) 342085.6 BIC (smaller is better) 342100.6 CAIC (smaller is better) 342102.6 HQIC (smaller is better) 342090.6 Generalized Chi-Square 7237350 Gener. Chi-Square / DF 177.85 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 152.47 2.8832 Residual 177.85 1.5186 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 168.09 2.0455 13441 82.17 <.0001 Age -0.2297 0.02694 27253 -8.53 <.0001 Year -0.7700 0.03499 27253 -22.00 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 27253 72.68 <.0001 Year 1 27253 484.18 <.0001 The SAS System 10:00 Thursday, October 22, 2015 37 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable BMI Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Gender 1 0 Number of Observations Read 41850 Number of Observations Used 37696 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 163933.13411 . 658.213 1 0 4 163201.87828 731.25582988 126.4324 2 0 4 163144.89997 56.97831004 2.800462 3 0 2 163144.86466 0.03530819 0.493634 4 0 2 163144.86354 0.00112587 0.002449 5 0 2 163144.86354 0.00000003 2.126E-6 The SAS System 10:00 Thursday, October 22, 2015 38 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 163144.9 AIC (smaller is better) 163148.9 AICC (smaller is better) 163148.9 BIC (smaller is better) 163163.9 CAIC (smaller is better) 163165.9 HQIC (smaller is better) 163153.9 Generalized Chi-Square 60770.41 Gener. Chi-Square / DF 1.61 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 13.3070 0.1791 Residual 1.6122 0.01438 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 37.1486 0.4336 12504 85.68 <.0001 Age -0.1271 0.006298 25189 -20.17 <.0001 Year 0.03587 0.006842 25189 5.24 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 25189 407.02 <.0001 Year 1 25189 27.48 <.0001 The SAS System 10:00 Thursday, October 22, 2015 39 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_MALE_OVER_75 Response Variable CholHDL Response Distribution Gaussian Link Function Identity Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique Restricted Maximum Likelihood Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 13719 not printed Gender 1 0 Number of Observations Read 41850 Number of Observations Used 38721 Dimensions R-side Cov. Parameters 2 Columns in X 3 Columns in Z per Subject 0 Subjects (Blocks in V) 13719 Max Obs per Subject 16 Optimization Information Optimization Technique Dual Quasi-Newton Parameters in Optimization 1 Lower Boundaries 0 Upper Boundaries 0 Fixed Effects Profiled Residual Variance Profiled Starting From Data Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 0 0 4 101801.37285 . 836.2781 1 0 2 101648.37031 153.00253836 278.2454 2 0 2 101608.4089 39.96141151 97.16833 3 0 2 101603.94957 4.45932730 20.20693 4 0 2 101603.7666 0.18296928 1.150416 5 0 2 101603.76599 0.00060434 0.012732 6 0 3 101603.76599 0.00000007 8.135E-6 The SAS System 10:00 Thursday, October 22, 2015 40 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 101603.8 AIC (smaller is better) 101607.8 AICC (smaller is better) 101607.8 BIC (smaller is better) 101622.8 CAIC (smaller is better) 101624.8 HQIC (smaller is better) 101612.8 Generalized Chi-Square 16479.47 Gener. Chi-Square / DF 0.43 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.9861 0.01492 Residual 0.4256 0.003757 Solutions for Fixed Effects Standard Effect Estimate Error DF t Value Pr > |t| Intercept 3.8556 0.1369 13012 28.17 <.0001 Age 0.004020 0.001891 25706 2.13 0.0335 Year -0.03595 0.002295 25706 -15.66 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 25706 4.52 0.0335 Year 1 25706 245.26 <.0001