Source: LV01p08.sas 5/7/01 Ed Stanek Table 1a. List of subject and exposure Age Person in Years Latency 1978: expos. 1942: period Obs id AGE78 EXP1 LAT1 1 1 47 0 0 2 2 9 0 0 3 3 9 0 0 4 4 33 3 0 5 5 55 1 0 6 6 64 1 0 7 7 15 0 0 8 8 42 0 0 9 9 21 0 0 10 10 47 0 0 11 11 48 0 0 12 12 26 0 0 13 13 50 1 0 14 14 18 0 0 15 15 48 0 0 16 16 27 1 0 17 17 77 1 0 18 18 47 1 0 19 19 27 0 0 20 20 26 0 0 Based on LV01p08.sas Source: LV01p08.sas 5/7/01 Ed Stanek Table 1b. List of exposure freq distribution in person years The FREQ Procedure Person*Years*expos. 1942:*EXP1 Cumulative Cumulative exp1 Frequency Percent Frequency Percent --------------------------------------------------------- 0 493 51.84 493 51.84 1 261 27.44 754 79.28 2 130 13.67 884 92.95 3 51 5.36 935 98.32 4 13 1.37 948 99.68 5 3 0.32 951 100.00 Based on LV01p08.sas Source: LV01p08.sas 5/7/01 Ed Stanek Figure 2. Example of Cox Proportional Hazards Model age mean=40 std=20 Latency=0 Effect of exposure=1 # subjects=1000 Exposure Mean=Poisson(1) * PROC PRINT DATA=pred1 The PHREG Procedure Model Information Data Set WORK.D1 Entry Time Variable age78 Age*in*1978:*AGE78 Dependent Variable survt Survival*Yrs:*SURVT Censoring Variable censor Right*censor*time:*CENSOR Censoring Value(s) 1 Ties Handling BRESLOW Summary of the Number of Event and Censored Values Percent Total Event Censored Censored 845 174 671 79.41 Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Without With Criterion Covariates Covariates -2 LOG L 1666.465 1646.173 AIC 1666.465 1648.173 SBC 1666.465 1651.332 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 20.2918 1 <.0001 Score 22.5052 1 <.0001 Wald 22.0837 1 <.0001 Analysis of Maximum Likelihood Estimates Parameter Standard Hazard Variable DF Estimate Error Chi-Square Pr > ChiSq Ratio Variable Label exp1 1 0.35966 0.07653 22.0837 <.0001 1.433 Total yrs*exposed:*EXP1 Based on LV01p08.sas