NOTE: Copyright (c) 1999-2000 by SAS Institute Inc., Cary, NC, USA. NOTE: SAS (r) Proprietary Software Release 8.1 (TS1M0) Licensed to NEW YORK UNIVERSITY, Site 0011830001. NOTE: This session is executing on the WIN_98 platform. NOTE: SAS initialization used: real time 14.88 seconds 1 DATA NEW; 2 *This step reads data with days as lines; 3 infile 'dayanx.txt'; 4 input id day sex dianx; 5 diaryday=day-1; 6 If day < 32; 7 TITLE1 'G89.2247 Lecture 12 Analyses'; NOTE: The infile 'dayanx.txt' is: File Name=C:\My Documents\My SAS Files\V8\dayanx.txt, RECFM=V,LRECL=256 NOTE: 2380 records were read from the infile 'dayanx.txt'. The minimum record length was 7. The maximum record length was 9. NOTE: The data set WORK.NEW has 2108 observations and 5 variables. NOTE: DATA statement used: real time 1.08 seconds 8 proc genmod data=new descending; 9 class id; 10 model dianx=diaryday sex diaryday*sex /dist=bin; 11 Repeated subject=ID /Type=ar(1); 12 Title2 'GEE marginal repeated measures analysis with interaction'; 13 run; NOTE: PROC GENMOD is modeling the probability that dianx='1'. NOTE: Algorithm converged. NOTE: The scale parameter was held fixed. NOTE: Algorithm converged. NOTE: There were 2108 observations read from the data set WORK.NEW. NOTE: PROCEDURE GENMOD used: real time 3.78 seconds 14 proc genmod data=new descending; 15 class id; 16 model dianx=diaryday sex /dist=bin; 17 Repeated subject=ID /Type=Ar(1); 18 Title2 'GEE marginal repeated measures analysis, no interaction'; 19 run; NOTE: PROC GENMOD is modeling the probability that dianx='1'. NOTE: Algorithm converged. NOTE: The scale parameter was held fixed. NOTE: Algorithm converged. NOTE: There were 2108 observations read from the data set WORK.NEW. NOTE: PROCEDURE GENMOD used: real time 0.54 seconds 20 Data new2; 21 *This step creates file with days as variables, persons as lines; 22 array anx[31] a1-a31; 23 array dday[31] d1-d31; 24 do i=1 to 31 until (last.id); 25 set new; 26 by id; 27 anx[i]=dianx; 28 dday[i]=day; 29 end; 30 *Now we tag days beyond first incidence of high anxiety 31 cl1-cl32 are binary indicators of history of anxiety; 32 array clos[32] cl1-cl32 ; 33 do i=1 to 31; 34 clos[i]=0; 35 end; 36 do i=1 to 30; 37 if (anx[i]=1) or (clos[i]=1) then clos[i+1]=1; 38 end; 39 *The next lines create variables that are used in traditional survival analysis; 40 do i=1 to 31; 41 if clos[i]=0 then anxtime=i; 42 end; 43 censor=0; 44 if cl31=0 then censor=1; 45 drop i dianx day; 46 run; NOTE: There were 2108 observations read from the data set WORK.NEW. NOTE: The data set WORK.NEW2 has 68 observations and 99 variables. NOTE: DATA statement used: real time 0.26 seconds 47 data new3; 48 *This step creates file with tagged days as lines; 49 set new2; 50 array anx[31] a1-a31; 51 array dday[31] d1-d31; 52 array clos[32] cl1-cl32; 53 past=0; 54 do i=1 to 31; 55 day=dday[i]; 56 dianx=anx[i]; 57 past=clos[i]; 58 output; 59 end; 60 drop i a1-a31 d1-d31 cl1-cl32; NOTE: There were 68 observations read from the data set WORK.NEW2. NOTE: The data set WORK.NEW3 has 2108 observations and 8 variables. NOTE: DATA statement used: real time 0.05 seconds 61 data new4; 62 *This creates dummy codes for day and deletes days beyond incident days; 63 set new3; 64 diaryday=day-15; 65 array dumday[31] dm1-dm31; 66 do i=1 to 31; 67 dumday[i]=0; 68 end; 69 dumday[day]=1; 70 if past=0; 71 drop past i; 72 run; NOTE: There were 2108 observations read from the data set WORK.NEW3. NOTE: The data set WORK.NEW4 has 1412 observations and 38 variables. NOTE: DATA statement used: real time 0.11 seconds 73 proc logist descending; 74 model dianx= dm1-dm31 sex /noint; 75 title2 'discrete time hazard model with 31 days'; 76 run; NOTE: PROC LOGISTIC is modeling the probability that dianx=1. WARNING: There is possibly a quasi-complete separation of data points. The maximum likelihood estimate may not exist. WARNING: The LOGISTIC procedure continues in spite of the above warning. Results shown are based on the last maximum likelihood iteration. Validity of the model fit is questionable. NOTE: There were 1412 observations read from the data set WORK.NEW4. NOTE: PROCEDURE LOGIST used: real time 1.70 seconds 77 data new5; 78 *This step creates dummy codes for 8 (4day) periods; 79 set new3; 80 diaryday=day; 81 array dumday[8] dm1-dm8; 82 do i=1 to 8; 83 dumday[i]=0; 84 end; 85 pairday=ceil(day/4); 86 dumday[pairday]=1; 87 if past=0; 88 drop past i pairday; 89 run; NOTE: There were 2108 observations read from the data set WORK.NEW3. NOTE: The data set WORK.NEW5 has 1412 observations and 15 variables. NOTE: DATA statement used: real time 0.72 seconds 90 proc logist descending ; 91 model dianx= dm1-dm8 sex /noint; 92 Title2 'hazard model with 8 periods, no interaction'; 93 run; NOTE: PROC LOGISTIC is modeling the probability that dianx=1. NOTE: Convergence criterion (GCONV=1E-8) satisfied. NOTE: There were 1412 observations read from the data set WORK.NEW5. NOTE: PROCEDURE LOGIST used: real time 0.16 seconds 94 proc logist descending; 95 model dianx= dm1-dm8 sex sex*dm2 sex*dm3 sex*dm4 96 sex*dm5 sex*dm6 sex*dm7 sex*dm8/noint; 97 Title2 'hazard model with 8 periods, interaction'; 98 run; NOTE: PROC LOGISTIC is modeling the probability that dianx=1. WARNING: There is possibly a quasi-complete separation of data points. The maximum likelihood estimate may not exist. WARNING: The LOGISTIC procedure continues in spite of the above warning. Results shown are based on the last maximum likelihood iteration. Validity of the model fit is questionable. NOTE: There were 1412 observations read from the data set WORK.NEW5. NOTE: PROCEDURE LOGIST used: real time 0.27 seconds 99 data survive; 100 set new2; 101 keep id sex anxtime censor; 102 run; NOTE: There were 68 observations read from the data set WORK.NEW2. NOTE: The data set WORK.SURVIVE has 68 observations and 4 variables. NOTE: DATA statement used: real time 0.00 seconds 103 proc sort; by sex; NOTE: There were 68 observations read from the data set WORK.SURVIVE. NOTE: The data set WORK.SURVIVE has 68 observations and 4 variables. NOTE: PROCEDURE SORT used: real time 0.04 seconds 104 proc means; by sex; var anxtime censor; 105 title2 'analysis of time to anxiety and censoring by gender'; 106 run; NOTE: There were 68 observations read from the data set WORK.SURVIVE. NOTE: PROCEDURE MEANS used: real time 0.37 seconds 107 proc phreg; 108 model ANXTIME * CENSOR (1) = SEX; 109 title2 'Cox proportional hazards model'; 110 run; NOTE: Convergence criterion (GCONV=1E-8) satisfied. NOTE: There were 68 observations read from the data set WORK.SURVIVE. NOTE: PROCEDURE PHREG used: real time 0.54 seconds 111 proc lifetest method=pl OUTSURV=work._surv; 112 time ANXTIME * CENSOR (1); 113 test SEX; 114 title2 'Product limit life table analysis'; 115 run; NOTE: There were 68 observations read from the data set WORK.SURVIVE. NOTE: The data set WORK._SURV has 45 observations and 5 variables. NOTE: PROCEDURE LIFETEST used: real time 0.54 seconds 116 data work._surv; set work._surv; 117 if survival > 0 then _lsurv = -log(survival); 118 if _lsurv > 0 then _llsurv = log(_lsurv); 119 run; NOTE: There were 45 observations read from the data set WORK._SURV. NOTE: The data set WORK._SURV has 45 observations and 7 variables. NOTE: DATA statement used: real time 0.22 seconds 120 ** Survival plots **; 121 goptions reset=symbol; 122 goptions ftext=SWISS ctext=BLACK htext=1 cells; 123 proc gplot data=work._surv ; 124 label anxtime = 'Survival Time'; 125 axis2 minor=none major=(number=6) 126 label=(angle=90 'Survival Distribution Function'); 127 symbol1 i=stepj c=BLUE l=1 width=1; 128 plot survival * anxtime=1 / 129 description="SDF of anxtime"; 130 run; NOTE: 25 observation(s) contained a MISSING value for the SURVIVAL * anxtime request. G89.2247 Lecture 12 Analyses 1 GEE marginal repeated measures analysis with interaction 09:10 Wednesday, April 11, 2001 The GENMOD Procedure Model Information Data Set WORK.NEW Distribution Binomial Link Function Logit Dependent Variable dianx Observations Used 2108 Probability Modeled Pr( dianx = 1 ) Class Level Information Class Levels Values id 68 1 2 4 5 6 7 9 10 11 12 13 14 15 16 18 19 22 27 28 29 30 32 35 36 37 38 40 41 42 43 44 45 47 49 50 52 55 56 57 59 60 62 64 65 66 67 68 70 71 72 73 74 77 78 79 80 81 84 85 88 90 91 93 94 95 97 98 99 Response Profile Ordered Ordered Level Value Count 1 1 275 2 0 1833 Parameter Information Parameter Effect Prm1 Intercept Prm2 diaryday Prm3 sex Prm4 diaryday*sex Criteria For Assessing Goodness Of Fit Criterion DF Value Value/DF Deviance 2104 1504.0408 0.7148 Scaled Deviance 2104 1504.0408 0.7148 Pearson Chi-Square 2104 2237.0850 1.0633 Scaled Pearson X2 2104 2237.0850 1.0633 Log Likelihood -752.0204 G89.2247 Lecture 12 Analyses 2 GEE marginal repeated measures analysis with interaction 09:10 Wednesday, April 11, 2001 The GENMOD Procedure Algorithm converged. Analysis Of Initial Parameter Estimates Standard Wald 95% Confidence Chi- Parameter DF Estimate Error Limits Square Pr > ChiSq Intercept 1 -2.5601 0.2360 -3.0226 -2.0976 117.70 <.0001 diaryday 1 0.0617 0.0118 0.0385 0.0849 27.19 <.0001 sex 1 -1.2471 0.3447 -1.9227 -0.5715 13.09 0.0003 diaryday*sex 1 0.0328 0.0165 0.0004 0.0652 3.94 0.0470 Scale 0 1.0000 0.0000 1.0000 1.0000 NOTE: The scale parameter was held fixed. GEE Model Information Correlation Structure AR(1) Subject Effect id (68 levels) Number of Clusters 68 Correlation Matrix Dimension 31 Maximum Cluster Size 31 Minimum Cluster Size 31 Algorithm converged. Analysis Of GEE Parameter Estimates Empirical Standard Error Estimates Standard 95% Confidence Parameter Estimate Error Limits Z Pr > |Z| Intercept -2.5868 0.4526 -3.4738 -1.6997 -5.72 <.0001 diaryday 0.0643 0.0134 0.0381 0.0905 4.82 <.0001 sex -1.2507 0.7300 -2.6814 0.1800 -1.71 0.0866 diaryday*sex 0.0345 0.0238 -0.0122 0.0812 1.45 0.1480 G89.2247 Lecture 12 Analyses 3 GEE marginal repeated measures analysis, no interaction 09:10 Wednesday, April 11, 2001 The GENMOD Procedure Model Information Data Set WORK.NEW Distribution Binomial Link Function Logit Dependent Variable dianx Observations Used 2108 Probability Modeled Pr( dianx = 1 ) Class Level Information Class Levels Values id 68 1 2 4 5 6 7 9 10 11 12 13 14 15 16 18 19 22 27 28 29 30 32 35 36 37 38 40 41 42 43 44 45 47 49 50 52 55 56 57 59 60 62 64 65 66 67 68 70 71 72 73 74 77 78 79 80 81 84 85 88 90 91 93 94 95 97 98 99 Response Profile Ordered Ordered Level Value Count 1 1 275 2 0 1833 Parameter Information Parameter Effect Prm1 Intercept Prm2 diaryday Prm3 sex Criteria For Assessing Goodness Of Fit Criterion DF Value Value/DF Deviance 2105 1507.9906 0.7164 Scaled Deviance 2105 1507.9906 0.7164 Pearson Chi-Square 2105 2160.0907 1.0262 Scaled Pearson X2 2105 2160.0907 1.0262 Log Likelihood -753.9953 G89.2247 Lecture 12 Analyses 4 GEE marginal repeated measures analysis, no interaction 09:10 Wednesday, April 11, 2001 The GENMOD Procedure Algorithm converged. Analysis Of Initial Parameter Estimates Standard Wald 95% Confidence Chi- Parameter DF Estimate Error Limits Square Pr > ChiSq Intercept 1 -2.8844 0.1851 -3.2473 -2.5215 242.72 <.0001 diaryday 1 0.0792 0.0082 0.0631 0.0954 92.92 <.0001 sex 1 -0.6181 0.1349 -0.8825 -0.3537 20.99 <.0001 Scale 0 1.0000 0.0000 1.0000 1.0000 NOTE: The scale parameter was held fixed. GEE Model Information Correlation Structure AR(1) Subject Effect id (68 levels) Number of Clusters 68 Correlation Matrix Dimension 31 Maximum Cluster Size 31 Minimum Cluster Size 31 Algorithm converged. Analysis Of GEE Parameter Estimates Empirical Standard Error Estimates Standard 95% Confidence Parameter Estimate Error Limits Z Pr > |Z| Intercept -2.9410 0.4668 -3.8560 -2.0261 -6.30 <.0001 diaryday 0.0831 0.0119 0.0597 0.1065 6.96 <.0001 sex -0.5709 0.4294 -1.4125 0.2706 -1.33 0.1836 G89.2247 Lecture 12 Analyses 5 discrete time hazard model with 31 days 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure Model Information Data Set WORK.NEW4 Response Variable dianx Number of Response Levels 2 Number of Observations 1412 Link Function Logit Optimization Technique Fisher's scoring Response Profile Ordered Total Value dianx Frequency 1 1 47 2 0 1365 Model Convergence Status Quasi-complete separation of data points detected. WARNING: The maximum likelihood estimate may not exist. WARNING: The LOGISTIC procedure continues in spite of the above warning. Results shown are based on the last maximum likelihood iteration. Validity of the model fit is questionable. Model Fit Statistics Without With Criterion Covariates Covariates AIC 1957.448 411.443 SC 1957.448 579.531 -2 Log L 1957.448 347.443 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 1610.0048 32 <.0001 Score 1238.8609 32 <.0001 Wald 335.6574 32 <.0001 G89.2247 Lecture 12 Analyses 6 discrete time hazard model with 31 days 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure WARNING: The validity of the model fit is questionable. Analysis of Maximum Likelihood Estimates Standard Parameter DF Estimate Error Chi-Square Pr > ChiSq dm1 1 -2.3878 0.5057 22.2993 <.0001 dm2 1 -3.9808 1.0275 15.0094 0.0001 dm3 1 -3.2560 0.7456 19.0721 <.0001 dm4 1 -2.4882 0.5573 19.9334 <.0001 dm5 1 -2.4110 0.5603 18.5158 <.0001 dm6 1 -16.0494 457.2 0.0012 0.9720 dm7 1 -3.7794 1.0310 13.4379 0.0002 dm8 1 -16.0462 461.7 0.0012 0.9723 dm9 1 -3.7564 1.0322 13.2449 0.0003 dm10 1 -3.7328 1.0334 13.0481 0.0003 dm11 1 -3.7136 1.0332 12.9190 0.0003 dm12 1 -16.0456 475.9 0.0011 0.9731 dm13 1 -16.0456 475.9 0.0011 0.9731 dm14 1 -16.0456 475.9 0.0011 0.9731 dm15 1 -16.0456 475.9 0.0011 0.9731 dm16 1 -3.6939 1.0330 12.7879 0.0003 dm17 1 -2.5253 0.6362 15.7563 <.0001 dm18 1 -16.0411 497.1 0.0010 0.9743 dm19 1 -2.1411 0.5698 14.1192 0.0002 dm20 1 -16.0357 521.3 0.0009 0.9755 dm21 1 -3.4973 1.0383 11.3466 0.0008 dm22 1 -16.0311 528.0 0.0009 0.9758 dm23 1 -3.4667 1.0401 11.1090 0.0009 dm24 1 -2.7207 0.7633 12.7035 0.0004 dm25 1 -3.3891 1.0397 10.6255 0.0011 dm26 1 -16.0374 557.3 0.0008 0.9770 dm27 1 -2.6385 0.7629 11.9616 0.0005 dm28 1 -16.0411 574.0 0.0008 0.9777 dm29 1 -2.1411 0.6453 11.0083 0.0009 dm30 1 -1.4535 0.5355 7.3666 0.0066 dm31 1 -1.5070 0.5844 6.6505 0.0099 sex 1 -0.2280 0.3226 0.4997 0.4796 Odds Ratio Estimates Point 95% Wald Effect Estimate Confidence Limits dm1 0.092 0.034 0.247 dm2 0.019 0.002 0.140 dm3 0.039 0.009 0.166 G89.2247 Lecture 12 Analyses 7 discrete time hazard model with 31 days 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure WARNING: The validity of the model fit is questionable. Odds Ratio Estimates Point 95% Wald Effect Estimate Confidence Limits dm4 0.083 0.028 0.248 dm5 0.090 0.030 0.269 dm6 <0.001 <0.001 >999.999 dm7 0.023 0.003 0.172 dm8 <0.001 <0.001 >999.999 dm9 0.023 0.003 0.177 dm10 0.024 0.003 0.181 dm11 0.024 0.003 0.185 dm12 <0.001 <0.001 >999.999 dm13 <0.001 <0.001 >999.999 dm14 <0.001 <0.001 >999.999 dm15 <0.001 <0.001 >999.999 dm16 0.025 0.003 0.188 dm17 0.080 0.023 0.278 dm18 <0.001 <0.001 >999.999 dm19 0.118 0.038 0.359 dm20 <0.001 <0.001 >999.999 dm21 0.030 0.004 0.232 dm22 <0.001 <0.001 >999.999 dm23 0.031 0.004 0.240 dm24 0.066 0.015 0.294 dm25 0.034 0.004 0.259 dm26 <0.001 <0.001 >999.999 dm27 0.071 0.016 0.319 dm28 <0.001 <0.001 >999.999 dm29 0.118 0.033 0.416 dm30 0.234 0.082 0.668 dm31 0.222 0.070 0.697 sex 0.796 0.423 1.498 Association of Predicted Probabilities and Observed Responses Percent Concordant 79.6 Somers' D 0.622 Percent Discordant 17.4 Gamma 0.642 Percent Tied 3.1 Tau-a 0.040 Pairs 64155 c 0.811 G89.2247 Lecture 12 Analyses 8 hazard model with 8 periods, no interaction 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure Model Information Data Set WORK.NEW5 Response Variable dianx Number of Response Levels 2 Number of Observations 1412 Link Function Logit Optimization Technique Fisher's scoring Response Profile Ordered Total Value dianx Frequency 1 1 47 2 0 1365 Model Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates AIC 1957.448 400.016 SC 1957.448 447.291 -2 Log L 1957.448 382.016 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 1575.4313 9 <.0001 Score 1235.3546 9 <.0001 Wald 445.6609 9 <.0001 G89.2247 Lecture 12 Analyses 9 hazard model with 8 periods, no interaction 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure Analysis of Maximum Likelihood Estimates Standard Parameter DF Estimate Error Chi-Square Pr > ChiSq dm1 1 -2.8509 0.3559 64.1547 <.0001 dm2 1 -3.5629 0.4977 51.2486 <.0001 dm3 1 -4.0152 0.6192 42.0480 <.0001 dm4 1 -5.0941 1.0244 24.7268 <.0001 dm5 1 -3.0139 0.4439 46.1021 <.0001 dm6 1 -3.4660 0.5564 38.8062 <.0001 dm7 1 -3.6478 0.6246 34.1068 <.0001 dm8 1 -1.6870 0.3771 20.0185 <.0001 sex 1 -0.2312 0.3190 0.5250 0.4687 Odds Ratio Estimates Point 95% Wald Effect Estimate Confidence Limits dm1 0.058 0.029 0.116 dm2 0.028 0.011 0.075 dm3 0.018 0.005 0.061 dm4 0.006 <0.001 0.046 dm5 0.049 0.021 0.117 dm6 0.031 0.010 0.093 dm7 0.026 0.008 0.089 dm8 0.185 0.088 0.388 sex 0.794 0.425 1.483 Association of Predicted Probabilities and Observed Responses Percent Concordant 68.0 Somers' D 0.429 Percent Discordant 25.1 Gamma 0.460 Percent Tied 6.8 Tau-a 0.028 Pairs 64155 c 0.714 G89.2247 Lecture 12 Analyses 10 hazard model with 8 periods, interaction 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure Model Information Data Set WORK.NEW5 Response Variable dianx Number of Response Levels 2 Number of Observations 1412 Link Function Logit Optimization Technique Fisher's scoring Response Profile Ordered Total Value dianx Frequency 1 1 47 2 0 1365 Model Convergence Status Quasi-complete separation of data points detected. WARNING: The maximum likelihood estimate may not exist. WARNING: The LOGISTIC procedure continues in spite of the above warning. Results shown are based on the last maximum likelihood iteration. Validity of the model fit is questionable. Model Fit Statistics Without With Criterion Covariates Covariates AIC 1957.448 406.384 SC 1957.448 490.428 -2 Log L 1957.448 374.384 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 1583.0634 16 <.0001 Score 1236.4094 16 <.0001 Wald 421.3291 16 <.0001 G89.2247 Lecture 12 Analyses 11 hazard model with 8 periods, interaction 09:10 Wednesday, April 11, 2001 The LOGISTIC Procedure WARNING: The validity of the model fit is questionable. Analysis of Maximum Likelihood Estimates Standard Parameter DF Estimate Error Chi-Square Pr > ChiSq dm1 1 -2.5649 0.4237 36.6542 <.0001 dm2 1 -3.4340 0.7184 22.8475 <.0001 dm3 1 -4.0254 1.0089 15.9192 <.0001 dm4 1 -4.0073 1.0090 15.7720 <.0001 dm5 1 -2.6856 0.5967 20.2559 <.0001 dm6 1 -3.5835 1.0138 12.4945 0.0004 dm7 1 -15.2029 333.5 0.0021 0.9636 dm8 1 -2.4849 0.7360 11.3996 0.0007 sex 1 -0.7370 0.5935 1.5420 0.2143 dm2*sex 1 0.2998 1.0994 0.0744 0.7851 dm3*sex 1 0.5211 1.3702 0.1446 0.7037 dm4*sex 1 -10.4585 171.6 0.0037 0.9514 dm5*sex 1 -0.0114 0.9831 0.0001 0.9908 dm6*sex 1 0.6656 1.3123 0.2573 0.6120 dm7*sex 1 12.4334 333.5 0.0014 0.9703 dm8*sex 1 1.5733 1.0066 2.4430 0.1180 Odds Ratio Estimates Point 95% Wald Effect Estimate Confidence Limits dm1 0.077 0.034 0.176 Association of Predicted Probabilities and Observed Responses Percent Concordant 68.9 Somers' D 0.453 Percent Discordant 23.5 Gamma 0.490 Percent Tied 7.6 Tau-a 0.029 Pairs 64155 c 0.727 G89.2247 Lecture 12 Analyses 12 analysis of time to anxiety and censoring by gender 09:10 Wednesday, April 11, 2001 -------------------------------------------- sex=0 --------------------------------------------- The MEANS Procedure Variable N Mean Std Dev Minimum Maximum ャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャ anxtime 23 17.6956522 12.2490622 1.0000000 31.0000000 censor 23 0.3478261 0.4869848 0 1.0000000 ャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャ -------------------------------------------- sex=1 --------------------------------------------- Variable N Mean Std Dev Minimum Maximum ャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャ anxtime 45 22.3333333 11.0925856 1.0000000 31.0000000 censor 45 0.3777778 0.4903101 0 1.0000000 ャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャャ G89.2247 Lecture 12 Analyses 13 Cox proportional hazards model 09:10 Wednesday, April 11, 2001 The PHREG Procedure Model Information Data Set WORK.SURVIVE Dependent Variable anxtime Censoring Variable censor Censoring Value(s) 1 Ties Handling BRESLOW Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 68 43 25 36.76 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 330.073 329.444 AIC 330.073 331.444 SBC 330.073 333.206 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 0.6283 1 0.4280 Score 0.6497 1 0.4202 Wald 0.6463 1 0.4215 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio sex 1 -0.25785 0.32075 0.6463 0.4215 0.773 G89.2247 Lecture 12 Analyses 14 Product limit life table analysis 09:10 Wednesday, April 11, 2001 The LIFETEST Procedure Product-Limit Survival Estimates Survival Standard Number Number anxtime Survival Failure Error Failed Left 0.0000 1.0000 0 0 0 68 1.0000 . . . 1 67 1.0000 . . . 2 66 1.0000 . . . 3 65 1.0000 . . . 4 64 1.0000 0.9265 0.0735 0.0317 5 63 2.0000 0.9118 0.0882 0.0344 6 62 3.0000 . . . 7 61 3.0000 0.8824 0.1176 0.0391 8 60 4.0000 . . . 9 59 4.0000 . . . 10 58 4.0000 . . . 11 57 4.0000 0.8235 0.1765 0.0462 12 56 5.0000 . . . 13 55 5.0000 . . . 14 54 5.0000 . . . 15 53 5.0000 0.7647 0.2353 0.0514 16 52 7.0000 0.7500 0.2500 0.0525 17 51 9.0000 0.7353 0.2647 0.0535 18 50 10.0000 0.7206 0.2794 0.0544 19 49 11.0000 0.7059 0.2941 0.0553 20 48 16.0000 0.6912 0.3088 0.0560 21 47 17.0000 . . . 22 46 17.0000 . . . 23 45 17.0000 0.6471 0.3529 0.0580 24 44 19.0000 . . . 25 43 19.0000 . . . 26 42 19.0000 . . . 27 41 19.0000 0.5882 0.4118 0.0597 28 40 21.0000 0.5735 0.4265 0.0600 29 39 23.0000 0.5588 0.4412 0.0602 30 38 24.0000 . . . 31 37 24.0000 0.5294 0.4706 0.0605 32 36 25.0000 0.5147 0.4853 0.0606 33 35 27.0000 . . . 34 34 27.0000 0.4853 0.5147 0.0606 35 33 29.0000 . . . 36 32 29.0000 . . . 37 31 29.0000 0.4412 0.5588 0.0602 38 30 30.0000 . . . 39 29 30.0000 . . . 40 28 30.0000 . . . 41 27 G89.2247 Lecture 12 Analyses 15 Product limit life table analysis 09:10 Wednesday, April 11, 2001 The LIFETEST Procedure Product-Limit Survival Estimates Survival Standard Number Number anxtime Survival Failure Error Failed Left 30.0000 . . . 42 26 30.0000 0.3676 0.6324 0.0585 43 25 31.0000* . . . 43 24 31.0000* . . . 43 23 31.0000* . . . 43 22 31.0000* . . . 43 21 31.0000* . . . 43 20 31.0000* . . . 43 19 31.0000* . . . 43 18 31.0000* . . . 43 17 31.0000* . . . 43 16 31.0000* . . . 43 15 31.0000* . . . 43 14 31.0000* . . . 43 13 31.0000* . . . 43 12 31.0000* . . . 43 11 31.0000* . . . 43 10 31.0000* . . . 43 9 31.0000* . . . 43 8 31.0000* . . . 43 7 31.0000* . . . 43 6 31.0000* . . . 43 5 31.0000* . . . 43 4 31.0000* . . . 43 3 31.0000* . . . 43 2 31.0000* . . . 43 1 31.0000* . . . 43 0 NOTE: The marked survival times are censored observations. Summary Statistics for Time Variable anxtime Quartile Estimates Point 95% Confidence Interval Percent Estimate [Lower Upper) 75 . . . 50 27.0000 19.0000 30.0000 25 8.0000 4.0000 19.0000 G89.2247 Lecture 12 Analyses 16 Product limit life table analysis 09:10 Wednesday, April 11, 2001 The LIFETEST Procedure Mean Standard Error 20.3971 1.3756 NOTE: The mean survival time and its standard error were underestimated because the largest observation was censored and the estimation was restricted to the largest event time. Summary of the Number of Censored and Uncensored Values Percent Total Failed Censored Censored 68 43 25 36.76 Rank Tests for the Association of anxtime with Covariates Univariate Chi-Squares for the Wilcoxon Test Test Standard Pr > Variable Statistic Deviation Chi-Square Chi-Square sex 2.5217 2.1305 1.4009 0.2366 Covariance Matrix for the Wilcoxon Statistics Variable sex sex 4.53921 Forward Stepwise Sequence of Chi-Squares for the Wilcoxon Test Pr > Chi-Square Pr > Variable DF Chi-Square Chi-Square Increment Increment sex 1 1.4009 0.2366 1.4009 0.2366 Univariate Chi-Squares for the Log-Rank Test Test Standard Pr > Variable Statistic Deviation Chi-Square Chi-Square sex 2.4003 2.9780 0.6497 0.4202 G89.2247 Lecture 12 Analyses 17 Product limit life table analysis 09:10 Wednesday, April 11, 2001 The LIFETEST Procedure Covariance Matrix for the Log-Rank Statistics Variable sex sex 8.86823 Forward Stepwise Sequence of Chi-Squares for the Log-Rank Test Pr > Chi-Square Pr > Variable DF Chi-Square Chi-Square Increment Increment sex 1 0.6497 0.4202 0.6497 0.4202