S7 File. SAS output Glimmix outcomes under 75 years of age (quadratic analyses) The SAS System 14:33 Monday, October 19, 2015 11 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 233326 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1104079.1045 0.39631954 . 1 0 0 1095016.1411 0.02573704 . 2 0 0 1095321.4036 0.00002315 . 3 0 0 1095321.5865 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 12 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1095322 Generalized Chi-Square 232440.6 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9962 0.002917 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 1.0099 0.08238 67649 12.26 <.0001 Age -0.03548 0.000559 166E3 -63.52 <.0001 Gender 0 0.6047 0.09863 67649 6.13 <.0001 Gender 1 0 . . . . Year -0.04841 0.01361 166E3 -3.56 0.0004 YearSQ 0.001490 0.000610 166E3 2.44 0.0146 Year*Gender 0 -0.03650 0.01797 166E3 -2.03 0.0422 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.000913 0.000805 166E3 1.13 0.2567 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 166E3 4034.17 <.0001 Gender 1 67649 37.59 <.0001 Year 1 166E3 55.04 <.0001 YearSQ 1 166E3 23.37 <.0001 Year*Gender 1 166E3 4.13 0.0422 YearSQ*Gender 1 166E3 1.29 0.2567 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -1.2574 0.006908 67649 -182.02 <.0001 0.05 -1.2710 -1.2439 1 -1.5518 0.007996 67649 -194.08 <.0001 0.05 -1.5675 -1.5362 The SAS System 14:33 Monday, October 19, 2015 13 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 198758 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 958250.61524 2.00000000 . 1 0 0 980122.37867 0.04675805 . 2 0 0 983142.13085 0.00096449 . 3 0 0 983165.43979 0.00000012 . 4 0 0 983165.44131 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 14 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 983165.4 Generalized Chi-Square 199637.5 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0045 0.003186 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -2.1175 0.09687 62477 -21.86 <.0001 Age 0.02425 0.000778 136E3 31.17 <.0001 Gender 0 0.4304 0.1148 62477 3.75 0.0002 Gender 1 0 . . . . Year -0.2079 0.01562 136E3 -13.31 <.0001 YearSQ 0.007058 0.000722 136E3 9.77 <.0001 Year*Gender 0 0.02763 0.02120 136E3 1.30 0.1925 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00127 0.000963 136E3 -1.32 0.1864 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 136E3 971.28 <.0001 Gender 1 62477 14.06 0.0002 Year 1 136E3 335.08 <.0001 YearSQ 1 136E3 177.77 <.0001 Year*Gender 1 136E3 1.70 0.1925 YearSQ*Gender 1 136E3 1.75 0.1864 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -1.5070 0.008060 62477 -186.96 <.0001 0.05 -1.5228 -1.4912 1 -2.0721 0.01047 62477 -197.92 <.0001 0.05 -2.0926 -2.0516 The SAS System 14:33 Monday, October 19, 2015 15 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 171958 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 664352.42828 1.07468382 . 1 0 0 915507.90534 2.00000000 . 2 0 0 1093461.6984 1.02156616 . 3 0 0 1185785.0925 0.54476437 . 4 0 0 1212507.4217 0.16080838 . 5 0 0 1214622.6145 0.00185320 . 6 0 0 1214637.025 0.00000014 . 7 0 0 1214637.0258 0.00000000 . The SAS System 14:33 Monday, October 19, 2015 16 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 1214637 Generalized Chi-Square 176458.5 Gener. Chi-Square / DF 1.03 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0262 0.003500 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -4.7885 0.2365 57997 -20.24 <.0001 Age 0.04263 0.002389 114E3 17.85 <.0001 Gender 0 0.5998 0.2310 57997 2.60 0.0094 Gender 1 0 . . . . Year -0.2735 0.03783 114E3 -7.23 <.0001 YearSQ 0.006001 0.001928 114E3 3.11 0.0019 Year*Gender 0 0.04509 0.04777 114E3 0.94 0.3452 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00126 0.002379 114E3 -0.53 0.5958 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 114E3 318.54 <.0001 Gender 1 57997 6.74 0.0094 Year 1 114E3 110.40 <.0001 YearSQ 1 114E3 20.38 <.0001 Year*Gender 1 114E3 0.89 0.3452 YearSQ*Gender 1 114E3 0.28 0.5958 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -3.6933 0.02271 57997 -162.66 <.0001 0.05 -3.7378 -3.6488 1 -4.6578 0.03681 57997 -126.53 <.0001 0.05 -4.7299 -4.5856 The SAS System 14:33 Monday, October 19, 2015 17 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1174452.1625 1.29610451 . 1 0 0 1188182.9989 0.05995876 . 2 0 0 1189748.0431 0.00018601 . 3 0 0 1189752.3159 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 18 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1189752 Generalized Chi-Square 245722.5 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0003 0.002854 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -1.0791 0.06470 71052 -16.68 <.0001 Age 0.01015 0.000597 175E3 17.01 <.0001 Gender 0 -0.2985 0.07444 71052 -4.01 <.0001 Gender 1 0 . . . . Year 0.1726 0.01029 175E3 16.77 <.0001 YearSQ -0.00110 0.000495 175E3 -2.21 0.0269 Year*Gender 0 0.06273 0.01457 175E3 4.30 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00244 0.000698 175E3 -3.50 0.0005 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 289.28 <.0001 Gender 1 71052 16.08 <.0001 Year 1 175E3 783.48 <.0001 YearSQ 1 175E3 44.09 <.0001 Year*Gender 1 175E3 18.53 <.0001 YearSQ*Gender 1 175E3 12.24 0.0005 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 1.6402 0.007763 71052 211.28 <.0001 0.05 1.6250 1.6554 1 1.5575 0.008025 71052 194.09 <.0001 0.05 1.5418 1.5733 The SAS System 14:33 Monday, October 19, 2015 19 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 234210 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1065646.9585 0.12504222 . 1 0 0 1038950.3242 0.00380567 . 2 0 0 1039684.6056 0.00000666 . 3 0 0 1039685.2203 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 20 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1039685 Generalized Chi-Square 234220.0 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0001 0.002922 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -1.5859 0.06428 68482 -24.67 <.0001 Age 0.01116 0.000515 166E3 21.69 <.0001 Gender 0 0.4978 0.07765 68482 6.41 <.0001 Gender 1 0 . . . . Year 0.3231 0.01031 166E3 31.35 <.0001 YearSQ -0.01334 0.000473 166E3 -28.17 <.0001 Year*Gender 0 -0.09482 0.01441 166E3 -6.58 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.003918 0.000658 166E3 5.96 <.0001 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 166E3 470.48 <.0001 Gender 1 68482 41.10 <.0001 Year 1 166E3 1464.50 <.0001 YearSQ 1 166E3 1196.36 <.0001 Year*Gender 1 166E3 43.31 <.0001 YearSQ*Gender 1 166E3 35.50 <.0001 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 0.9038 0.006297 68482 143.54 <.0001 0.05 0.8915 0.9162 1 0.9445 0.006749 68482 139.95 <.0001 0.05 0.9313 0.9578 The SAS System 14:33 Monday, October 19, 2015 21 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1101624.6891 1.08173724 . 1 0 0 1349094.1111 2.00000000 . 2 0 0 1472482.4359 0.38708553 . 3 0 0 1501743.3716 0.02842941 . 4 0 0 1503230.766 0.00010323 . 5 0 0 1503234.5888 0.00000000 . The SAS System 14:33 Monday, October 19, 2015 22 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 1503235 Generalized Chi-Square 250722.5 Gener. Chi-Square / DF 1.02 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0207 0.002912 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 0.4658 0.1001 71052 4.65 <.0001 Age 0.01760 0.001043 175E3 16.88 <.0001 Gender 0 -0.2402 0.1037 71052 -2.32 0.0205 Gender 1 0 . . . . Year 0.05467 0.01621 175E3 3.37 0.0007 YearSQ 0.005624 0.000838 175E3 6.71 <.0001 Year*Gender 0 0.04177 0.02207 175E3 1.89 0.0585 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00221 0.001135 175E3 -1.94 0.0519 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 284.99 <.0001 Gender 1 71052 5.37 0.0205 Year 1 175E3 46.86 <.0001 YearSQ 1 175E3 63.39 <.0001 Year*Gender 1 175E3 3.58 0.0585 YearSQ*Gender 1 175E3 3.78 0.0519 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 3.1243 0.01462 71052 213.74 <.0001 0.05 3.0957 3.1530 1 3.2099 0.01634 71052 196.47 <.0001 0.05 3.1779 3.2419 The SAS System 14:33 Monday, October 19, 2015 23 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 238524 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1027612.4054 0.15072230 . 1 0 0 1020079.065 0.00365884 . 2 0 0 1020245.014 0.00000409 . 3 0 0 1020245.1099 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 24 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1020245 Generalized Chi-Square 238471.3 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9998 0.002895 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 1.0109 0.06907 69500 14.64 <.0001 Age -0.04086 0.000499 169E3 -81.89 <.0001 Gender 0 0.4956 0.08436 69500 5.87 <.0001 Gender 1 0 . . . . Year 0.1878 0.01108 169E3 16.95 <.0001 YearSQ -0.00411 0.000490 169E3 -8.38 <.0001 Year*Gender 0 -0.08712 0.01512 169E3 -5.76 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.003243 0.000669 169E3 4.85 <.0001 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 169E3 6705.32 <.0001 Gender 1 69500 34.51 <.0001 Year 1 169E3 363.79 <.0001 YearSQ 1 169E3 55.14 <.0001 Year*Gender 1 169E3 33.18 <.0001 YearSQ*Gender 1 169E3 23.48 <.0001 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 0.1123 0.005756 69500 19.51 <.0001 0.05 0.1010 0.1236 1 0.1727 0.006155 69500 28.06 <.0001 0.05 0.1607 0.1848 The SAS System 14:33 Monday, October 19, 2015 25 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1026376.3443 1.00768547 . 1 0 0 1345020.0678 0.58778091 . 2 0 0 1535398.8616 0.20070364 . 3 0 0 1590951.432 0.02164402 . 4 0 0 1595048.1083 0.00022740 . 5 0 0 1595071.3779 0.00000002 . 6 0 0 1595071.3788 0.00000000 . The SAS System 14:33 Monday, October 19, 2015 26 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 1595071 Generalized Chi-Square 241172.6 Gener. Chi-Square / DF 0.98 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9818 0.002801 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -0.04152 0.1098 71052 -0.38 0.7053 Age 0.01810 0.001276 175E3 14.19 <.0001 Gender 0 -0.1544 0.1043 71052 -1.48 0.1389 Gender 1 0 . . . . Year 0.3420 0.01722 175E3 19.86 <.0001 YearSQ -0.01057 0.000916 175E3 -11.54 <.0001 Year*Gender 0 0.05125 0.02380 175E3 2.15 0.0313 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00303 0.001260 175E3 -2.41 0.0160 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 201.33 <.0001 Gender 1 71052 2.19 0.1389 Year 1 175E3 954.36 <.0001 YearSQ 1 175E3 368.30 <.0001 Year*Gender 1 175E3 4.64 0.0313 YearSQ*Gender 1 175E3 5.80 0.0160 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 3.6050 0.01739 71052 207.32 <.0001 0.05 3.5709 3.6391 1 3.6251 0.01878 71052 192.98 <.0001 0.05 3.5883 3.6619 The SAS System 14:33 Monday, October 19, 2015 27 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 226287 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 973424.45828 0.17603934 . 1 0 0 965775.02965 0.00469148 . 2 0 0 965903.63564 0.00000480 . 3 0 0 965903.6774 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 28 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 965903.7 Generalized Chi-Square 226216.5 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9997 0.002972 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 1.9409 0.07119 66733 27.26 <.0001 Age -0.02911 0.000492 16E4 -59.10 <.0001 Gender 0 -0.9562 0.09414 66733 -10.16 <.0001 Gender 1 0 . . . . Year -0.02211 0.01143 16E4 -1.93 0.0531 YearSQ 0.001202 0.000504 16E4 2.39 0.0171 Year*Gender 0 0.05402 0.01668 16E4 3.24 0.0012 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00155 0.000728 16E4 -2.13 0.0334 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 16E4 3492.72 <.0001 Gender 1 66733 103.17 <.0001 Year 1 16E4 0.35 0.5568 YearSQ 1 16E4 1.38 0.2405 Year*Gender 1 16E4 10.48 0.0012 YearSQ*Gender 1 16E4 4.53 0.0334 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -0.4671 0.005981 66733 -78.11 <.0001 0.05 -0.4789 -0.4554 1 0.06194 0.006193 66733 10.00 <.0001 0.05 0.04980 0.07408 The SAS System 14:33 Monday, October 19, 2015 29 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 226287 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 978689.21905 0.23142310 . 1 0 0 969264.12644 0.00767626 . 2 0 0 969453.16115 0.00000565 . 3 0 0 969453.25903 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 30 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 969453.3 Generalized Chi-Square 226255.7 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9999 0.002973 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -1.8289 0.07361 66733 -24.85 <.0001 Age 0.02042 0.000504 16E4 40.55 <.0001 Gender 0 0.6437 0.09298 66733 6.92 <.0001 Gender 1 0 . . . . Year -0.00573 0.01187 16E4 -0.48 0.6290 YearSQ -0.00014 0.000524 16E4 -0.26 0.7945 Year*Gender 0 -0.01253 0.01655 16E4 -0.76 0.4492 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.000277 0.000726 16E4 0.38 0.7023 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 16E4 1644.52 <.0001 Gender 1 66733 47.92 <.0001 Year 1 16E4 2.10 0.1472 YearSQ 1 16E4 0.00 0.9954 Year*Gender 1 16E4 0.57 0.4492 YearSQ*Gender 1 16E4 0.15 0.7023 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -0.1313 0.005811 66733 -22.60 <.0001 0.05 -0.1427 -0.1199 1 -0.6617 0.006501 66733 -101.78 <.0001 0.05 -0.6744 -0.6489 The SAS System 14:33 Monday, October 19, 2015 31 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 226287 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1098004.5607 0.53149121 . 1 0 0 1116795.5454 0.02956436 . 2 0 0 1117949.7526 0.00012251 . 3 0 0 1117954.17 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 32 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1117954 Generalized Chi-Square 226523.8 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0011 0.002976 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -3.3595 0.1070 66733 -31.40 <.0001 Age 0.01957 0.000721 16E4 27.15 <.0001 Gender 0 0.5773 0.1309 66733 4.41 <.0001 Gender 1 0 . . . . Year 0.05780 0.01705 16E4 3.39 0.0007 YearSQ -0.00227 0.000745 16E4 -3.05 0.0023 Year*Gender 0 -0.07558 0.02324 16E4 -3.25 0.0011 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.002390 0.001016 16E4 2.35 0.0186 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 16E4 737.36 <.0001 Gender 1 66733 19.45 <.0001 Year 1 16E4 2.97 0.0850 YearSQ 1 16E4 4.50 0.0339 Year*Gender 1 16E4 10.58 0.0011 YearSQ*Gender 1 16E4 5.54 0.0186 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -1.7788 0.008234 66733 -216.02 <.0001 0.05 -1.7949 -1.7627 1 -1.7974 0.008809 66733 -204.03 <.0001 0.05 -1.8146 -1.7801 The SAS System 14:33 Monday, October 19, 2015 33 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1152270.9061 0.78186033 . 1 0 0 1320813.9108 0.70525768 . 2 0 0 1376737.3301 0.16109916 . 3 0 0 1382621.2902 0.00203230 . 4 0 0 1382684.4364 0.00000025 . 5 0 0 1382684.4437 0.00000000 . The SAS System 14:33 Monday, October 19, 2015 34 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 1382684 Generalized Chi-Square 245194.6 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9982 0.002848 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -0.5076 0.07771 71052 -6.53 <.0001 Age 0.006301 0.000853 175E3 7.39 <.0001 Gender 0 -0.2944 0.07839 71052 -3.76 0.0002 Gender 1 0 . . . . Year 0.2477 0.01209 175E3 20.50 <.0001 YearSQ -0.00242 0.000628 175E3 -3.84 0.0001 Year*Gender 0 0.06756 0.01690 175E3 4.00 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00330 0.000872 175E3 -3.78 0.0002 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 54.57 <.0001 Gender 1 71052 14.11 0.0002 Year 1 175E3 1110.09 <.0001 YearSQ 1 175E3 86.79 <.0001 Year*Gender 1 175E3 15.99 <.0001 YearSQ*Gender 1 175E3 14.28 0.0002 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 2.6231 0.01160 71052 226.07 <.0001 0.05 2.6004 2.6459 1 2.6207 0.01244 71052 210.68 <.0001 0.05 2.5963 2.6451 The SAS System 14:33 Monday, October 19, 2015 35 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_75 Response Variable OUT_CholHDL Response Distribution Binomial Link Function Logit Variance Function Default Variance Matrix Blocked By BSISID Estimation Technique MPL Degrees of Freedom Method Between-Within Class Level Information Class Levels Values BSISID 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 231030 Dimensions R-side Cov. Parameters 2 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 938181.27078 0.27472033 0.091976 1 0 2 926440.68852 0.07167768 0.000228 2 0 2 926651.64718 0.00415328 0.005692 3 0 2 926652.70748 0.00021068 2.06E-7 4 0 1 926652.71146 0.00001002 0.004076 5 0 1 926652.71145 0.00000050 0.002964 6 0 1 926652.71141 0.00000029 0.004289 7 0 1 926652.71106 0.00000043 0.006219 The SAS System 14:56 Friday, November 27, 2015 12 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Subiterations Function Change Gradient 8 0 1 926652.71155 0.00000017 0.001952 9 0 1 926652.71136 0.00000019 0.002831 10 0 1 926652.71113 0.00000028 0.004106 11 0 1 926652.71145 0.00000041 0.005954 12 0 1 926652.71098 0.00000059 0.008634 13 0 1 926652.71167 0.00000086 0.01252 14 0 1 926652.71067 0.00000001 0.012458 Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Log Pseudo-Likelihood 926652.7 Generalized Chi-Square 115620.1 Gener. Chi-Square / DF 0.50 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 0.5040 0.003870 Residual 0.5005 0.001747 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -2.9091 0.07061 68328 -41.20 <.0001 Age 0.03244 0.000698 163E3 46.50 <.0001 Gender 0 -0.2256 0.08248 68328 -2.74 0.0062 Gender 1 0 . . . . Year 0.2223 0.01002 163E3 22.19 <.0001 YearSQ -0.00763 0.000436 163E3 -17.51 <.0001 Year*Gender 0 -0.02939 0.01409 163E3 -2.09 0.0370 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.000958 0.000607 163E3 1.58 0.1143 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 163E3 2162.42 <.0001 Gender 1 68328 7.48 0.0062 Year 1 163E3 868.25 <.0001 YearSQ 1 163E3 555.26 <.0001 The SAS System 14:56 Friday, November 27, 2015 13 The GLIMMIX Procedure Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Year*Gender 1 163E3 4.35 0.0370 YearSQ*Gender 1 163E3 2.49 0.1143 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 0.1990 0.009065 68328 21.95 <.0001 0.05 0.1812 0.2168 1 0.6363 0.01006 68328 63.26 <.0001 0.05 0.6166 0.6560 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1126238.0298 0.86319363 . 1 0 0 1338889.8985 2.00000000 . 2 0 0 1431582.2221 1.23699796 . 3 0 0 1447938.4967 0.02596252 . 4 0 0 1448407.3822 0.00002586 . 5 0 0 1448407.7734 0.00000000 . The SAS System 14:33 Monday, October 19, 2015 38 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 1448408 Generalized Chi-Square 247705.0 Gener. Chi-Square / DF 1.01 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0084 0.002877 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -0.2259 0.08576 71052 -2.63 0.0085 Age 0.01382 0.000943 175E3 14.65 <.0001 Gender 0 -0.1085 0.08626 71052 -1.26 0.2086 Gender 1 0 . . . . Year 0.1628 0.01356 175E3 12.01 <.0001 YearSQ 0.001474 0.000714 175E3 2.07 0.0389 Year*Gender 0 0.03402 0.01879 175E3 1.81 0.0703 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00220 0.000981 175E3 -2.24 0.0252 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 214.55 <.0001 Gender 1 71052 1.58 0.2086 Year 1 175E3 366.28 <.0001 YearSQ 1 175E3 0.59 0.4434 Year*Gender 1 175E3 3.28 0.0703 YearSQ*Gender 1 175E3 5.01 0.0252 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 2.9011 0.01316 71052 220.41 <.0001 0.05 2.8753 2.9269 1 2.9516 0.01453 71052 203.17 <.0001 0.05 2.9231 2.9800 The SAS System 14:33 Monday, October 19, 2015 39 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 171621 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 800292.70462 0.90345866 . 1 0 0 941380.76715 0.36245344 . 2 0 0 984802.21286 0.04997493 . 3 0 0 988504.55594 0.00064696 . 4 0 0 988537.03416 0.00000008 . 5 0 0 988537.03765 0.00000000 . The SAS System 14:33 Monday, October 19, 2015 40 The GLIMMIX Procedure Convergence criterion (PCONV=1.11022E-8) satisfied. Fit Statistics -2 Res Log Pseudo-Likelihood 988537.0 Generalized Chi-Square 171342.9 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9984 0.003408 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -2.5630 0.1362 57317 -18.82 <.0001 Age 0.02008 0.001275 114E3 15.76 <.0001 Gender 0 -0.2028 0.1541 57317 -1.32 0.1880 Gender 1 0 . . . . Year -0.2364 0.02157 114E3 -10.96 <.0001 YearSQ 0.008584 0.001047 114E3 8.19 <.0001 Year*Gender 0 0.05489 0.02984 114E3 1.84 0.0658 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00198 0.001428 114E3 -1.39 0.1653 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 114E3 248.23 <.0001 Gender 1 57317 1.73 0.1880 Year 1 114E3 196.17 <.0001 YearSQ 1 114E3 113.11 <.0001 Year*Gender 1 114E3 3.38 0.0658 YearSQ*Gender 1 114E3 1.92 0.1653 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -2.7251 0.01381 57317 -197.36 <.0001 0.05 -2.7522 -2.6981 1 -2.8905 0.01587 57317 -182.10 <.0001 0.05 -2.9216 -2.8594 The SAS System 14:33 Monday, October 19, 2015 41 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1141110.2672 0.45626642 . 1 0 0 1237056.6793 0.11577394 . 2 0 0 1256947.893 0.00599338 . 3 0 0 1257512.5997 0.00000876 . 4 0 0 1257513.0385 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 42 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1257513 Generalized Chi-Square 309510.6 Gener. Chi-Square / DF 1.26 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.2600 0.003595 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 0.7104 0.07440 71052 9.55 <.0001 Age 0.01305 0.000641 175E3 20.36 <.0001 Gender 0 -0.2111 0.08720 71052 -2.42 0.0155 Gender 1 0 . . . . Year -0.6013 0.01241 175E3 -48.44 <.0001 YearSQ 0.04413 0.000615 175E3 71.71 <.0001 Year*Gender 0 0.06128 0.01721 175E3 3.56 0.0004 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00356 0.000846 175E3 -4.20 <.0001 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 414.33 <.0001 Gender 1 71052 5.86 0.0155 Year 1 175E3 4400.35 <.0001 YearSQ 1 175E3 10014.6 <.0001 Year*Gender 1 175E3 12.69 0.0004 YearSQ*Gender 1 175E3 17.67 <.0001 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 1.3611 0.009127 71052 149.13 <.0001 0.05 1.3433 1.3790 1 1.3999 0.01003 71052 139.56 <.0001 0.05 1.3803 1.4196 The SAS System 14:33 Monday, October 19, 2015 43 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 127852 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 617915.03749 0.88914503 . 1 0 0 647307.14854 0.29018304 . 2 0 0 652759.80815 0.00515593 . 3 0 0 652912.56857 0.00000190 . 4 0 0 652912.68223 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 44 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 652912.7 Generalized Chi-Square 127921.4 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0006 0.003958 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -3.4119 0.1112 52277 -30.68 <.0001 Age 0.04158 0.001115 75568 37.31 <.0001 Gender 0 0.01214 0.1180 52277 0.10 0.9180 Gender 1 0 . . . . Year -0.2559 0.01837 75568 -13.93 <.0001 YearSQ 0.01118 0.000906 75568 12.33 <.0001 Year*Gender 0 0.04635 0.02522 75568 1.84 0.0661 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00199 0.001229 75568 -1.62 0.1049 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 75568 1391.70 <.0001 Gender 1 52277 0.01 0.9180 Year 1 75568 340.71 <.0001 YearSQ 1 75568 274.52 <.0001 Year*Gender 1 75568 3.38 0.0661 YearSQ*Gender 1 75568 2.63 0.1049 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -1.8773 0.01143 52277 -164.21 <.0001 0.05 -1.8997 -1.8549 1 -2.1317 0.01342 52277 -158.89 <.0001 0.05 -2.1580 -2.1054 The SAS System 14:33 Monday, October 19, 2015 45 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1185866.0491 0.59285402 . 1 0 0 1228157.9023 0.12640747 . 2 0 0 1232252.4446 0.00220244 . 3 0 0 1232284.3302 0.00000031 . 4 0 0 1232284.3323 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 46 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1232284 Generalized Chi-Square 245084.4 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 0.9977 0.002847 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -0.8426 0.06749 71052 -12.48 <.0001 Age 0.01201 0.000647 175E3 18.55 <.0001 Gender 0 -0.02916 0.07538 71052 -0.39 0.6989 Gender 1 0 . . . . Year 0.2080 0.01076 175E3 19.32 <.0001 YearSQ -0.00379 0.000527 175E3 -7.20 <.0001 Year*Gender 0 0.01269 0.01503 175E3 0.84 0.3983 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00103 0.000730 175E3 -1.41 0.1591 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 344.08 <.0001 Gender 1 71052 0.15 0.6989 Year 1 175E3 813.15 <.0001 YearSQ 1 175E3 139.28 <.0001 Year*Gender 1 175E3 0.71 0.3983 YearSQ*Gender 1 175E3 1.98 0.1591 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 1.8553 0.008248 71052 224.92 <.0001 0.05 1.8391 1.8714 1 1.8983 0.008960 71052 211.85 <.0001 0.05 1.8807 1.9158 The SAS System 14:33 Monday, October 19, 2015 47 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1195145.6569 0.57930351 . 1 0 0 1233558.3623 0.17403340 . 2 0 0 1235889.5754 0.00135053 . 3 0 0 1235901.2645 0.00000002 . 4 0 0 1235901.2649 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 48 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1235901 Generalized Chi-Square 245664.5 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0001 0.002854 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -2.3774 0.08225 71052 -28.91 <.0001 Age 0.01344 0.000709 175E3 18.95 <.0001 Gender 0 -0.4198 0.1034 71052 -4.06 <.0001 Gender 1 0 . . . . Year -0.06368 0.01283 175E3 -4.97 <.0001 YearSQ 0.002905 0.000591 175E3 4.91 <.0001 Year*Gender 0 0.02880 0.01904 175E3 1.51 0.1303 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00031 0.000864 175E3 -0.36 0.7170 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 359.29 <.0001 Gender 1 71052 16.47 <.0001 Year 1 175E3 26.80 <.0001 YearSQ 1 175E3 40.46 <.0001 Year*Gender 1 175E3 2.29 0.1303 YearSQ*Gender 1 175E3 0.13 0.7170 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -1.9668 0.008472 71052 -232.15 <.0001 0.05 -1.9834 -1.9502 1 -1.8598 0.008634 71052 -215.41 <.0001 0.05 -1.8768 -1.8429 The SAS System 14:33 Monday, October 19, 2015 49 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1175652.1801 2.00000000 . 1 0 0 1161291.3822 0.21378197 . 2 0 0 1161628.9304 0.00024824 . 3 0 0 1161629.1051 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 50 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1161629 Generalized Chi-Square 245649.1 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0000 0.002853 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -2.5542 0.07994 71052 -31.95 <.0001 Age 0.007951 0.000598 175E3 13.30 <.0001 Gender 0 -0.04559 0.1003 71052 -0.45 0.6494 Gender 1 0 . . . . Year 0.1427 0.01274 175E3 11.20 <.0001 YearSQ -0.00685 0.000571 175E3 -12.00 <.0001 Year*Gender 0 -0.00290 0.01814 175E3 -0.16 0.8728 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00018 0.000809 175E3 -0.23 0.8204 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 177.01 <.0001 Gender 1 71052 0.21 0.6494 Year 1 175E3 242.57 <.0001 YearSQ 1 175E3 294.46 <.0001 Year*Gender 1 175E3 0.03 0.8728 YearSQ*Gender 1 175E3 0.05 0.8204 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 -1.5296 0.007264 71052 -210.57 <.0001 0.05 -1.5439 -1.5154 1 -1.4158 0.007442 71052 -190.25 <.0001 0.05 -1.4304 -1.4012 The SAS System 14:33 Monday, October 19, 2015 51 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1074510.896 0.16956497 . 1 0 0 1058230.7701 0.00708711 . 2 0 0 1058488.585 0.00000592 . 3 0 0 1058488.6614 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 52 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1058489 Generalized Chi-Square 245653.0 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0000 0.002853 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 0.8627 0.06023 71052 14.32 <.0001 Age -0.00436 0.000478 175E3 -9.11 <.0001 Gender 0 0.06606 0.07386 71052 0.89 0.3711 Gender 1 0 . . . . Year -0.01640 0.009596 175E3 -1.71 0.0874 YearSQ 0.000816 0.000437 175E3 1.87 0.0619 Year*Gender 0 0.01207 0.01356 175E3 0.89 0.3733 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00065 0.000614 175E3 -1.05 0.2915 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 82.94 <.0001 Gender 1 71052 0.80 0.3711 Year 1 175E3 2.34 0.1263 YearSQ 1 175E3 2.57 0.1091 Year*Gender 1 175E3 0.79 0.3733 YearSQ*Gender 1 175E3 1.11 0.2915 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 0.6327 0.005828 71052 108.56 <.0001 0.05 0.6212 0.6441 1 0.5236 0.006092 71052 85.94 <.0001 0.05 0.5116 0.5355 The SAS System 14:33 Monday, October 19, 2015 53 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1104772.6208 2.00000000 . 1 0 0 1084929.9406 0.00673523 . 2 0 0 1085458.144 0.00003069 . 3 0 0 1085458.5782 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 54 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1085459 Generalized Chi-Square 246121.1 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0019 0.002859 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -3.4449 0.06134 71052 -56.16 <.0001 Age 0.05376 0.000499 175E3 107.70 <.0001 Gender 0 -0.2393 0.07368 71052 -3.25 0.0012 Gender 1 0 . . . . Year 0.1639 0.009804 175E3 16.72 <.0001 YearSQ -0.00678 0.000452 175E3 -15.01 <.0001 Year*Gender 0 0.007616 0.01369 175E3 0.56 0.5780 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.000167 0.000626 175E3 0.27 0.7900 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 11599.5 <.0001 Gender 1 71052 10.55 0.0012 Year 1 175E3 599.89 <.0001 YearSQ 1 175E3 457.02 <.0001 Year*Gender 1 175E3 0.31 0.5780 YearSQ*Gender 1 175E3 0.07 0.7900 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 0.6973 0.006041 71052 115.43 <.0001 0.05 0.6855 0.7092 1 0.8114 0.006554 71052 123.80 <.0001 0.05 0.7985 0.8242 The SAS System 14:33 Monday, October 19, 2015 55 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 1 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1107563.1107 1.85082414 . 1 0 0 1085735.9366 0.70680159 . 2 0 0 1086297.1734 0.00043525 . 3 0 0 1086297.6383 0.00000000 . Convergence criterion (PCONV=1.11022E-8) satisfied. The SAS System 14:33 Monday, October 19, 2015 56 The GLIMMIX Procedure Fit Statistics -2 Res Log Pseudo-Likelihood 1086298 Generalized Chi-Square 246290.3 Gener. Chi-Square / DF 1.00 Covariance Parameter Estimates Standard Cov Parm Estimate Error Residual (VC) 1.0026 0.002861 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -3.7986 0.07340 71052 -51.75 <.0001 Age 0.02118 0.000493 175E3 42.99 <.0001 Gender 0 0.2623 0.09085 71052 2.89 0.0039 Gender 1 0 . . . . Year 0.4028 0.01188 175E3 33.90 <.0001 YearSQ -0.01170 0.000525 175E3 -22.28 <.0001 Year*Gender 0 0.001947 0.01639 175E3 0.12 0.9054 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00050 0.000725 175E3 -0.69 0.4884 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 175E3 1848.50 <.0001 Gender 1 71052 8.34 0.0039 Year 1 175E3 2425.32 <.0001 YearSQ 1 175E3 1085.02 <.0001 Year*Gender 1 175E3 0.01 0.9054 YearSQ*Gender 1 175E3 0.48 0.4884 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 0.8406 0.006220 71052 135.16 <.0001 0.05 0.8285 0.8528 1 0.6393 0.006397 71052 99.93 <.0001 0.05 0.6267 0.6518 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 234210 Dimensions R-side Cov. Parameters 2 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1684451.0779 . 13590.14 1 0 5 1681881.0371 2570.0407954 2856.132 2 0 4 1681466.8801 414.15694143 318.6647 3 0 2 1681463.4733 3.40679116 83.65737 4 0 2 1681463.2109 0.26244797 1.852075 5 0 2 1681463.2108 0.00012779 0.011062 The SAS System 13:41 Wednesday, October 14, 2015 20 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 1681463 AIC (smaller is better) 1681467 AICC (smaller is better) 1681467 BIC (smaller is better) 1681486 CAIC (smaller is better) 1681488 HQIC (smaller is better) 1681473 Generalized Chi-Square 11257339 Gener. Chi-Square / DF 48.07 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 67.1523 0.4954 Residual 48.0666 0.1692 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 58.5873 0.3355 68482 174.61 <.0001 Age -0.05463 0.003672 166E3 -14.88 <.0001 Gender 0 -2.1984 0.3707 68482 -5.93 <.0001 Gender 1 0 . . . . Year -1.1718 0.04534 166E3 -25.85 <.0001 YearSQ 0.05763 0.001975 166E3 29.18 <.0001 Year*Gender 0 0.4041 0.06325 166E3 6.39 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.01617 0.002747 166E3 -5.89 <.0001 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 166E3 221.31 <.0001 Gender 1 68482 35.16 <.0001 Year 1 166E3 938.64 <.0001 YearSQ 1 166E3 1300.67 <.0001 Year*Gender 1 166E3 40.81 <.0001 YearSQ*Gender 1 166E3 34.67 <.0001 The SAS System 13:41 Wednesday, October 14, 2015 21 The GLIMMIX Procedure Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 50.3191 0.04933 68482 1020.14 <.0001 0.05 50.2224 50.4158 1 50.1319 0.05227 68482 959.13 <.0001 0.05 50.0295 50.2343 The SAS System 13:41 Wednesday, October 14, 2015 22 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 238524 Dimensions R-side Cov. Parameters 2 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1949594.2449 . 4243.042 1 0 4 1949403.8755 190.36938094 1536.091 2 0 2 1949361.98 41.89547462 272.769 3 0 2 1949360.7401 1.23990662 22.46923 4 0 2 1949360.7318 0.00830852 0.298332 Convergence criterion (GCONV=1E-8) satisfied. The SAS System 13:41 Wednesday, October 14, 2015 23 The GLIMMIX Procedure Fit Statistics -2 Res Log Likelihood 1949361 AIC (smaller is better) 1949365 AICC (smaller is better) 1949365 BIC (smaller is better) 1949383 CAIC (smaller is better) 1949385 HQIC (smaller is better) 1949370 Generalized Chi-Square 34192301 Gener. Chi-Square / DF 143.35 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 126.16 1.0010 Residual 143.35 0.4918 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 131.23 0.5327 69500 246.32 <.0001 Age 0.3595 0.005352 169E3 67.16 <.0001 Gender 0 -5.5200 0.6078 69500 -9.08 <.0001 Gender 1 0 . . . . Year -1.7322 0.07509 169E3 -23.07 <.0001 YearSQ 0.03733 0.003284 169E3 11.37 <.0001 Year*Gender 0 0.9423 0.1048 169E3 8.99 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.03532 0.004565 169E3 -7.74 <.0001 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 169E3 4510.31 <.0001 Gender 1 69500 82.48 <.0001 Year 1 169E3 579.21 <.0001 YearSQ 1 169E3 74.22 <.0001 Year*Gender 1 169E3 80.89 <.0001 YearSQ*Gender 1 169E3 59.88 <.0001 Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 138.24 0.07114 69500 1943.07 <.0001 0.05 138.10 138.38 1 137.78 0.07533 69500 1828.99 <.0001 0.05 137.64 137.93 The SAS System 13:41 Wednesday, October 14, 2015 24 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 226287 Dimensions R-side Cov. Parameters 2 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1057778.175 . 4116.153 1 0 2 1049781.7726 7996.4023862 544.1182 2 0 4 1049389.7421 392.03053987 37.47422 3 0 2 1049387.3889 2.35318591 7.197803 4 0 2 1049387.297 0.09187888 0.123534 5 0 2 1049387.297 0.00002723 0.000416 The SAS System 13:41 Wednesday, October 14, 2015 25 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 1049387 AIC (smaller is better) 1049391 AICC (smaller is better) 1049391 BIC (smaller is better) 1049410 CAIC (smaller is better) 1049412 HQIC (smaller is better) 1049397 Generalized Chi-Square 481286.1 Gener. Chi-Square / DF 2.13 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 26.4997 0.1508 Residual 2.1269 0.007534 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 35.9364 0.1283 66733 280.12 <.0001 Age -0.09282 0.002001 16E4 -46.39 <.0001 Gender 0 -1.7566 0.1095 66733 -16.04 <.0001 Gender 1 0 . . . . Year 0.05030 0.01244 16E4 4.04 <.0001 YearSQ 0.000964 0.000519 16E4 1.86 0.0635 Year*Gender 0 0.03051 0.01716 16E4 1.78 0.0754 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00124 0.000722 16E4 -1.72 0.0846 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 16E4 2152.18 <.0001 Gender 1 66733 257.37 <.0001 Year 1 16E4 56.19 <.0001 YearSQ 1 16E4 0.90 0.3441 Year*Gender 1 16E4 3.16 0.0754 YearSQ*Gender 1 16E4 2.97 0.0846 The SAS System 13:41 Wednesday, October 14, 2015 26 The GLIMMIX Procedure Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 29.4235 0.02792 66733 1053.90 <.0001 0.05 29.3688 29.4783 1 31.0033 0.02968 66733 1044.55 <.0001 0.05 30.9451 31.0615 The SAS System 13:41 Wednesday, October 14, 2015 27 The GLIMMIX Procedure Model Information Data Set WORK.ZODIAC_45_UNDER_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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 231030 Dimensions R-side Cov. Parameters 2 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 635121.9898 . 9621.249 1 0 4 633775.90142 1346.0883743 3487.352 2 0 4 632556.84669 1219.0547325 879.9731 3 0 2 632524.5748 32.27188402 395.0849 4 0 2 632514.81568 9.75912173 29.09748 5 0 2 632514.76483 0.05084940 1.059033 6 0 3 632514.76477 0.00006725 0.002717 The SAS System 13:41 Wednesday, October 14, 2015 28 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 632514.8 AIC (smaller is better) 632518.8 AICC (smaller is better) 632518.8 BIC (smaller is better) 632537.1 CAIC (smaller is better) 632539.1 HQIC (smaller is better) 632524.4 Generalized Chi-Square 120910.0 Gener. Chi-Square / DF 0.52 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 1.0106 0.006831 Residual 0.5234 0.001843 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 6.3654 0.03856 68328 165.07 <.0001 Age -0.02479 0.000436 163E3 -56.92 <.0001 Gender 0 0.2012 0.04213 68328 4.78 <.0001 Gender 1 0 . . . . Year -0.1641 0.005124 163E3 -32.03 <.0001 YearSQ 0.005817 0.000221 163E3 26.37 <.0001 Year*Gender 0 0.01453 0.007102 163E3 2.05 0.0407 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00049 0.000305 163E3 -1.61 0.1071 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 163E3 3239.41 <.0001 Gender 1 68328 22.80 <.0001 Year 1 163E3 1943.74 <.0001 YearSQ 1 163E3 1332.09 <.0001 Year*Gender 1 163E3 4.19 0.0407 YearSQ*Gender 1 163E3 2.60 0.1071 The SAS System 13:41 Wednesday, October 14, 2015 29 The GLIMMIX Procedure Gender Least Squares Means Standard Gender Estimate Error DF t Value Pr > |t| Alpha Lower Upper 0 4.0308 0.005873 68328 686.30 <.0001 0.05 4.0193 4.0424 1 3.7281 0.006233 68328 598.09 <.0001 0.05 3.7159 3.7403 The SAS System 13:41 Wednesday, October 14, 2015 30 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_75 Response Variable Diabetesduration 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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 238144 Dimensions R-side Cov. Parameters 2 Columns in X 10 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1006997.3849 . 10458.25 1 0 2 957557.29885 49440.086074 1858.952 2 0 4 948586.0377 8971.2611478 465.7557 3 0 4 947080.91699 1505.1207149 49.58319 4 0 2 947055.93921 24.97777751 16.07227 5 0 2 947052.84209 3.09712568 0.883477 6 0 3 947052.83247 0.00961790 0.005259 7 0 3 947052.83247 0.00000011 5.654E-6 The SAS System 12:36 Sunday, October 25, 2015 13 The GLIMMIX Procedure Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 947052.8 AIC (smaller is better) 947056.8 AICC (smaller is better) 947056.8 BIC (smaller is better) 947075.2 CAIC (smaller is better) 947077.2 HQIC (smaller is better) 947062.5 Generalized Chi-Square 199768.1 Gener. Chi-Square / DF 0.84 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 28.5129 0.1566 Residual 0.8389 0.002893 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept -17.1653 0.1073 69053 -160.03 <.0001 Age 0.2141 0.001917 169E3 111.65 <.0001 Gender 0 -0.3456 0.06648 69053 -5.20 <.0001 Gender 1 0 . . . . Year 0.6116 0.006731 169E3 90.86 <.0001 YearSQ 0.007561 0.000281 169E3 26.95 <.0001 Year*Gender 0 0.03033 0.008987 169E3 3.37 0.0007 Year*Gender 1 0 . . . . YearSQ*Gender 0 -0.00158 0.000387 169E3 -4.07 <.0001 YearSQ*Gender 1 0 . . . . Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Age 1 169E3 12465.2 <.0001 Gender 1 69053 27.03 <.0001 Year 1 169E3 16866.9 <.0001 YearSQ 1 169E3 1222.21 <.0001 Year*Gender 1 169E3 11.39 0.0007 YearSQ*Gender 1 169E3 16.58 <.0001 The SAS System 12:36 Sunday, October 25, 2015 15 The GLIMMIX Procedure Model Information Data Set SYS.ZODIAC_45_UNDER_75 Response Variable Age 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 71054 not printed Gender 2 0 1 Number of Observations Read 245653 Number of Observations Used 245653 Dimensions R-side Cov. Parameters 2 Columns in X 8 Columns in Z per Subject 0 Subjects (Blocks in V) 71054 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 1347937.2313 . 39592.8 1 0 2 975255.91607 372681.31522 3238.681 2 0 6 839902.87306 135353.04301 737.2053 3 0 6 776470.26299 63432.610071 148.4285 4 0 4 762697.95276 13772.310231 49.83586 5 0 4 758180.73586 4517.2169033 9.626648 6 0 3 757875.02516 305.71069792 3.14544 7 0 3 757823.81602 51.20914546 0.992833 The SAS System 12:36 Sunday, October 25, 2015 38 The GLIMMIX Procedure Iteration History Objective Max Iteration Restarts Evaluations Function Change Gradient 8 0 3 757818.64253 5.17348615 0.086999 9 0 3 757818.60393 0.03859571 0.002163 10 0 3 757818.60391 0.00002359 1.911E-6 Convergence criterion (GCONV=1E-8) satisfied. Fit Statistics -2 Res Log Likelihood 757818.6 AIC (smaller is better) 757822.6 AICC (smaller is better) 757822.6 BIC (smaller is better) 757840.9 CAIC (smaller is better) 757842.9 HQIC (smaller is better) 757828.3 Generalized Chi-Square 35176.99 Gener. Chi-Square / DF 0.14 Covariance Parameter Estimates Standard Cov Parm Subject Estimate Error CS BSISID 102.84 0.5461 Residual 0.1432 0.000485 Solutions for Fixed Effects Standard Effect Gender Estimate Error DF t Value Pr > |t| Intercept 49.5096 0.05771 71052 857.98 <.0001 Gender 0 -0.9860 0.07919 71052 -12.45 <.0001 Gender 1 0 . . . . Year 0.9492 0.002659 175E3 357.01 <.0001 Year*Gender 0 0.01586 0.003667 175E3 4.32 <.0001 Year*Gender 1 0 . . . . YearSQ*Gender 0 0.000971 0.000109 175E3 8.92 <.0001 YearSQ*Gender 1 0.001607 0.000115 175E3 14.01 <.0001 Type III Tests of Fixed Effects Num Den Effect DF DF F Value Pr > F Gender 1 71052 155.01 <.0001 Year 1 175E3 272477 <.0001 Year*Gender 1 175E3 18.69 <.0001 YearSQ*Gender 2 175E3 137.88 <.0001