AMFIT Version 2.07 Nov 1997 May 4, 1998 Ganbarimashoo! 16:48:40 Copyright (C) HiroSoft International 1986-96 Written by Dale L. Preston and Donald A. Pierce Sun-4 Version Workspace size - 8000000 < af-tabav.txt log log-tabav.txt @ ! output to log file ! af-tabav.txt ! Oak Ridge Mortality Study: X-10 Y-12 Evers ! setup for amfit-- ERR Main Effects Model ! see epicure User's Guide and Command Summary ! run amfit and at the prompt input this file, i.e. ! < af-tabav.txt ! ! see variables.txt ( same as Appendix Table AIII ORNL-6785 ) ! for definitions of variable NAMES age F XG IG B S L y1 e1 D PY @ ! INPUT gdata.txt @ ! read data from gdata.txt Input from gdata.txt 4230 records read 4230 records used 0 records rejected Workspace for 300 variables. 13 are currently defined. Up to 287 new variables can be created TRAN LAGE= log(age/52.5) @ ! log age for power law model TRAN D = D/100000 @ ! D= dose in Sv TRAN C= py/1000 ; BRATE=e1/C; LBR=log(BRATE) @ TRAN ageS= (age-52.5)/100 @ TRAN A = (age-12.5)/5 @ ! A is age expressed as a factor LEVELS F 3 XG 10 IG 3 B 5 S 2 L 2 A 15 @ ! SELECT XG < 10 @ ! remove first ! to select low dose data CASES y1 @ ! observed events pyr C @ ! person-years ( 1000s ) !______________________________________________________________ ! Fit Main Effects Model ! Internal Baseline Based on Power Law For Age ! With Linear Excess Relative Risk (ERR) !_____________________________________________________________ NOSTRATA @ Workspace for 300 variables. 19 are currently defined. Up to 281 new variables can be created FITOPT S nointercept @ LOGLINEAR 0 LAGE B S L IG F @ ! baseline rates ! LINEAR 1 D=0 @ FIT @ 6 Iterations Piece-wise exponential regression model Product additive excess model {T0*(1 + T1 + T2 + ...)} Y1 is used for cases C is used for person years Parameter Summary Table # Name Estimate Std.Err. Test Stat. P value -- ----------------------- ------------ ---------- ----------- ------- Log-linear term 0 1 B_1 ..................... 0.5520 0.1315 4.199 < 0.001 2 B_2 ..................... 0.7507 0.1010 7.436 < 0.001 3 B_3 ..................... 0.7529 0.09194 8.189 < 0.001 4 B_4 ..................... 0.5907 0.1021 5.784 < 0.001 5 B_5 ..................... 0.09033 0.1809 0.4994 > 0.5 7 S_2 ..................... -0.4192 0.08328 -5.033 < 0.001 9 L_2 ..................... 0.1108 0.09052 1.224 0.221 11 IG_2 .................... 0.09002 0.07292 1.235 0.217 12 IG_3 .................... 0.01874 0.08714 0.2150 > 0.5 14 F_2 ..................... 0.1280 0.08610 1.486 0.137 15 F_3 ..................... 0.04666 0.07778 0.5999 > 0.5 16 LAGE .................... 5.219 0.2164 24.12 < 0.001 Linear term 1 17 D ....................... 0 Fixed 2.499 0.012 Records used = 4230 Deviance = 2030.72 df = 4218 Pearson Chi2 = 142093 NULL @ ! Rererence model for likelihood ratio test LINEAR 1 +D @ SCORE @ ! Score test for Excess Relative Risk The score statistic is 6.2459 with df = 1 (P = 0.0124) ! ERR cannot be less than -1/4.617 ( max dose) LINEAR 1 D @ PARA 17 <20 > -0.216 @ FIT @LRT @ ! Likelihood ratio test 4 Iterations Piece-wise exponential regression model Product additive excess model {T0*(1 + T1 + T2 + ...)} Y1 is used for cases C is used for person years Parameter Summary Table # Name Estimate Std.Err. Test Stat. P value -- ----------------------- ------------ ---------- ----------- ------- Log-linear term 0 1 B_1 ..................... 0.5380 0.1317 4.084 < 0.001 2 B_2 ..................... 0.7308 0.1016 7.195 < 0.001 3 B_3 ..................... 0.7279 0.09290 7.834 < 0.001 4 B_4 ..................... 0.5652 0.1031 5.483 < 0.001 5 B_5 ..................... 0.07170 0.1812 0.3956 > 0.5 7 S_2 ..................... -0.4097 0.08338 -4.913 < 0.001 9 L_2 ..................... 0.1181 0.09063 1.303 0.193 11 IG_2 .................... 0.05757 0.07497 0.7679 0.443 12 IG_3 .................... 0.02600 0.08728 0.2979 > 0.5 14 F_2 ..................... 0.1558 0.08724 1.786 0.074 15 F_3 ..................... 0.06145 0.07807 0.7870 0.431 16 LAGE .................... 5.204 0.2166 24.02 < 0.001 Linear term 1 17 D ....................... 1.515 0.8194 1.849 0.064 Records used = 4230 Deviance = 2025.15 df = 4217 Pearson Chi2 = 124420 LR statistic = 5.570 df = 1 P = 0.0183 BOUND 17 @ ! 95% likelihood-based CI for ERR Likelihood bound for parameter 17 D MLE 1.515 97.50% lower bound 0.17515 97.50% upper bound 3.5874 !______________________________________________________________ ! Fit Main Effects Model ! Use External Rates-- see epicure user's guide Chapter 7 ! Below use LBR = LOG(BRATE) as "offset" ! With Linear Excess Relative Risk (ERR) !_____________________________________________________________ LOGLINEAR 0 LBR=1 ageS B S L IG F @ ! baseline rates LINEAR 1 D=0 @ FIT @ 4 Iterations Piece-wise exponential regression model Product additive excess model {T0*(1 + T1 + T2 + ...)} Y1 is used for cases C is used for person years Parameter Summary Table # Name Estimate Std.Err. Test Stat. P value -- ----------------------- ------------ ---------- ----------- ------- Log-linear term 0 1 B_1 ..................... -0.07873 0.1341 -0.5871 > 0.5 2 B_2 ..................... -0.02431 0.1029 -0.2363 > 0.5 3 B_3 ..................... -0.07727 0.09281 -0.8325 0.405 4 B_4 ..................... -0.2251 0.1017 -2.213 0.027 5 B_5 ..................... -0.6441 0.1783 -3.613 < 0.001 7 S_2 ..................... -0.4226 0.08326 -5.075 < 0.001 9 L_2 ..................... 0.1056 0.09046 1.167 0.243 11 IG_2 .................... 0.07411 0.07276 1.018 0.308 12 IG_3 .................... 0.02170 0.08705 0.2493 > 0.5 14 F_2 ..................... 0.1260 0.08608 1.464 0.143 15 F_3 ..................... 0.04453 0.07776 0.5727 > 0.5 16 LBR ..................... 1.000 Fixed -0.4597 > 0.5 17 AGES .................... -0.2486 0.3818 -0.6512 > 0.5 Linear term 1 18 D ....................... 0 Fixed 2.458 0.014 Records used = 4230 Deviance = 2020.05 df = 4218 Pearson Chi2 = 186359 NULL @ ! Rererence model for likelihood ratio test LINEAR 1 +D @ SCORE @ ! Score test for Excess Relative Risk The score statistic is 6.0395 with df = 1 (P = 0.0140) LINEAR 1 D @ PARA 18 <20 > -0.216 @ FIT @LRT @ ! Likelihood ratio test 4 Iterations Piece-wise exponential regression model Product additive excess model {T0*(1 + T1 + T2 + ...)} Y1 is used for cases C is used for person years Parameter Summary Table # Name Estimate Std.Err. Test Stat. P value -- ----------------------- ------------ ---------- ----------- ------- Log-linear term 0 1 B_1 ..................... -0.09218 0.1344 -0.6860 0.493 2 B_2 ..................... -0.04345 0.1035 -0.4200 > 0.5 3 B_3 ..................... -0.1013 0.09374 -1.080 0.280 4 B_4 ..................... -0.2495 0.1027 -2.431 0.015 5 B_5 ..................... -0.6618 0.1786 -3.705 < 0.001 7 S_2 ..................... -0.4135 0.08336 -4.960 < 0.001 9 L_2 ..................... 0.1125 0.09057 1.242 0.214 11 IG_2 .................... 0.04277 0.07479 0.5719 > 0.5 12 IG_3 .................... 0.02873 0.08719 0.3296 > 0.5 14 F_2 ..................... 0.1528 0.08719 1.753 0.080 15 F_3 ..................... 0.05877 0.07805 0.7530 0.451 16 LBR ..................... 1.000 Fixed -0.4962 > 0.5 17 AGES .................... -0.2731 0.3823 -0.7143 0.475 Linear term 1 18 D ....................... 1.449 0.8049 1.801 0.072 Records used = 4230 Deviance = 2014.75 df = 4217 Pearson Chi2 = 163948 LR statistic = 5.299 df = 1 P = 0.0213 BOUND 18 @ ! 95% likelihood-based CI for ERR Likelihood bound for parameter 18 D MLE 1.449 97.50% lower bound 0.14589 97.50% upper bound 3.4782 END @ Total elapsed time: 0:05:01