The QUANTLIFE Procedure (Experimental)

Example 76.1 Primary Biliary Cirrhosis Study

This example illustrates how to detect varying covariate effects on survival time with quantile regression analysis. Consider a study of primary biliary cirrhosis, a rare but fatal chronic liver disease discussed by Fleming and Harrington (1991). Researchers followed 418 patients with this disease, of whom 161 died during the study.

The data set contains the following variables:

  • Time, follow-up time, in years

  • Status, event indicator with value 1 for death time and value 0 for censored time

  • Age, age in years from birth to study registration

  • Albumin, serum albumin level, in gm/dl

  • Bilirubin, serum bilirubin level, in mg/dl

  • Edema, edema presence

  • Protime, prothrombin time, in seconds

The following statements create the data set PBC that is used in this example:

data pbc;
   input Time Status Age Albumin Bilirubin Edema Protime @@;
   label Time="Follow-up Time in Days";
   logAlbumin   = log(Albumin);
   logBilirubin = log(Bilirubin); 
   logProtime   = log(Protime);
   datalines;
  400 1 58.7652 2.60 14.5 1.0 12.2 4500 0 56.4463 4.14  1.1 0.0 10.6
 1012 1 70.0726 3.48  1.4 0.5 12.0 1925 1 54.7406 2.54  1.8 0.5 10.3
 1504 0 38.1054 3.53  3.4 0.0 10.9 2503 1 66.2587 3.98  0.8 0.0 11.0
 1832 0 55.5346 4.09  1.0 0.0  9.7 2466 1 53.0568 4.00  0.3 0.0 11.0
 2400 1 42.5079 3.08  3.2 0.0 11.0   51 1 70.5599 2.74 12.6 1.0 11.5
 3762 1 53.7139 4.16  1.4 0.0 12.0  304 1 59.1376 3.52  3.6 0.0 13.6

   ... more lines ...   

  989 0 35.0000 3.23  0.7 0.0 10.8  681 1 67.0000 2.96  1.2 0.0 10.9
 1103 0 39.0000 3.83  0.9 0.0 11.2 1055 0 57.0000 3.42  1.6 0.0  9.9
  691 0 58.0000 3.75  0.8 0.0 10.4  976 0 53.0000 3.29  0.7 0.0 10.6
;

The next statements fit a linear model for the log of survival time of the PBC patients with the covariates logBilirubin, logProtime, logAlbumin, Age, and Edema.

ods graphics on;
proc quantlife data=pbc log method=na plot=(quantplot survival) seed=1268;
   model Time*Status(0)=logBilirubin logProtime logAlbumin Age Edema
                        /quantile=(.1 .2 .3 .4 .5 .6 .75);
run;

You use the QUANTILE= option to specify a set of quantiles of interest for comparing quantile-specific covariate effects. The METHOD= option specifies the Nelson-Aalen method for estimating the regression parameters.

The QUANTLIFE procedure provides resampling methods for computing confidence limits for the parameters; see the section Confidence Interval for details. By default, the repetition number is 200. You can request a different number of repetitions with the NREP= option. You can also use the SEED= option to specify the seed for generating random numbers so that you can later reproduce the results.

Output 76.1.1 displays model information and information about censoring in the data. Out of 418 observations, 257 are censored.

Output 76.1.1: Model Information

The QUANTLIFE Procedure

Model Information
Data Set WORK.PBC
Dependent Variable Log(Time)
Censoring Variable Status
Censoring Value(s) 0
Number of Observations 418
Method Nelson-Aalen
Replications 200
Seed for Random Number Generator 1268

Summary of the Number of Event and Censored
Values
Total Event Censored Percent
Censored
418 161 257 61.48


Output 76.1.2 provides the parameter estimates. Each quantile level has a set of parameter estimates and confidence limits.

Output 76.1.2: Parameter Estimates at Different Quantiles

Parameter Estimates
Quantile Parameter DF Estimate Standard
Error
95% Confidence Limits t Value Pr > |t|
0.1000 Intercept 1 14.8030 4.0967 6.7736 22.8325 3.61 0.0003
  logBilirubin 1 -0.4488 0.1485 -0.7398 -0.1578 -3.02 0.0027
  logProtime 1 -3.6378 1.4560 -6.4915 -0.7841 -2.50 0.0129
  logAlbumin 1 1.9286 0.9756 0.0165 3.8408 1.98 0.0487
  Age 1 -0.0244 0.0107 -0.0455 -0.00334 -2.27 0.0237
  Edema 1 -1.0712 0.6688 -2.3820 0.2396 -1.60 0.1100
0.2000 Intercept 1 15.1800 2.6664 9.9540 20.4060 5.69 <.0001
  logBilirubin 1 -0.6532 0.0886 -0.8268 -0.4796 -7.37 <.0001
  logProtime 1 -3.3273 0.9401 -5.1699 -1.4847 -3.54 0.0004
  logAlbumin 1 1.6842 0.6888 0.3343 3.0342 2.45 0.0149
  Age 1 -0.0291 0.00687 -0.0425 -0.0156 -4.23 <.0001
  Edema 1 -0.7265 0.3179 -1.3497 -0.1034 -2.29 0.0228
0.3000 Intercept 1 13.2382 2.5296 8.2804 18.1961 5.23 <.0001
  logBilirubin 1 -0.6013 0.0762 -0.7506 -0.4521 -7.90 <.0001
  logProtime 1 -2.5816 0.8907 -4.3273 -0.8359 -2.90 0.0039
  logAlbumin 1 1.7246 0.7142 0.3248 3.1245 2.41 0.0162
  Age 1 -0.0244 0.00716 -0.0385 -0.0104 -3.41 0.0007
  Edema 1 -0.8577 0.2763 -1.3992 -0.3163 -3.10 0.0020
0.4000 Intercept 1 13.4716 3.0874 7.4204 19.5228 4.36 <.0001
  logBilirubin 1 -0.6047 0.0846 -0.7705 -0.4389 -7.15 <.0001
  logProtime 1 -2.1632 1.1726 -4.4615 0.1351 -1.84 0.0658
  logAlbumin 1 0.9819 0.7191 -0.4274 2.3912 1.37 0.1728
  Age 1 -0.0255 0.00681 -0.0389 -0.0122 -3.74 0.0002
  Edema 1 -1.0589 0.3104 -1.6672 -0.4506 -3.41 0.0007
0.5000 Intercept 1 10.9205 2.8047 5.4235 16.4175 3.89 0.0001
  logBilirubin 1 -0.5315 0.0904 -0.7087 -0.3543 -5.88 <.0001
  logProtime 1 -1.2222 1.2142 -3.6020 1.1577 -1.01 0.3148
  logAlbumin 1 1.5700 0.6284 0.3383 2.8016 2.50 0.0129
  Age 1 -0.0318 0.00883 -0.0491 -0.0145 -3.60 0.0004
  Edema 1 -0.7316 0.3743 -1.4653 0.00202 -1.95 0.0513
0.6000 Intercept 1 11.2381 2.6294 6.0846 16.3917 4.27 <.0001
  logBilirubin 1 -0.5701 0.0852 -0.7370 -0.4031 -6.69 <.0001
  logProtime 1 -1.3508 1.1402 -3.5856 0.8840 -1.18 0.2368
  logAlbumin 1 1.3704 0.5091 0.3726 2.3682 2.69 0.0074
  Age 1 -0.0226 0.0109 -0.0440 -0.00111 -2.06 0.0399
  Edema 1 -0.5141 0.3088 -1.1193 0.0912 -1.66 0.0968
0.7500 Intercept 1 10.0954 3.1893 3.8445 16.3463 3.17 0.0017
  logBilirubin 1 -0.6366 0.1071 -0.8466 -0.4267 -5.94 <.0001
  logProtime 1 -0.9670 1.2343 -3.3862 1.4521 -0.78 0.4338
  logAlbumin 1 1.8148 0.5883 0.6618 2.9678 3.08 0.0022
  Age 1 -0.0203 0.0156 -0.0509 0.0102 -1.30 0.1931
  Edema 1 -0.3529 0.3120 -0.9644 0.2586 -1.13 0.2587


For comparison, the following statements use the LIFEREG procedure to fit a Weibull distribution to the data. The LIFEREG procedure fits an accelerated failure time model, which assumes that the effect of independent variables is multiplicative on the event time.

proc lifereg data=pbc;
   model Time*Status(0)=logBilirubin logProtime logAlbumin Age Edema;
run;

Output 76.1.3 provides the parameter estimates that are computed by PROC LIFEREG.

Output 76.1.3: Parameter Estimates From PROC LIFEREG

The LIFEREG Procedure

Analysis of Maximum Likelihood Parameter Estimates
Parameter DF Estimate Standard Error 95% Confidence Limits Chi-Square Pr > ChiSq
Intercept 1 12.2155 1.4539 9.3658 15.0651 70.59 <.0001
logBilirubin 1 -0.5770 0.0556 -0.6861 -0.4680 107.55 <.0001
logProtime 1 -1.7565 0.5248 -2.7850 -0.7280 11.20 0.0008
logAlbumin 1 1.6694 0.4276 0.8313 2.5074 15.24 <.0001
Age 1 -0.0265 0.0053 -0.0368 -0.0162 25.35 <.0001
Edema 1 -0.6303 0.1805 -0.9842 -0.2764 12.19 0.0005
Scale 1 0.6807 0.0430 0.6014 0.7704    
Weibull Shape 1 1.4691 0.0928 1.2980 1.6628    


The p-value for logProtime is very small. For this same variable, the p-values that result from the quantile regression analysis are 0.3148 for the 0.5th quantile and 0.4338 for the 0.75th quantile, and the p-values are much smaller for the lower quantiles. Apparently, the effect of this covariate depends on which side of the response distribution is being modeled.

The PLOT=QUANTPLOT option in the PROC QUANTLIFE statement requests the quantile process plots in Output 76.1.4 and Output 76.1.5. These displays plot the estimated regression parameter against the quantile level. You can use these plots to compare quantile-specific covariate effects. If the curve is not constant, it can indicate heterogeneity in the data. The interpretation of the regression coefficients at a given quantile is similar to classical regression analysis. That is, the coefficient from a given covariate indicates the effect on log(Time) of a unit change in that covariate, assuming that the other covariates are fixed.

Output 76.1.4: Quantile Processes with 95$\% $ Confidence Bands

Quantile Processes with 95% Confidence Bands


Output 76.1.5: Quantile Processes with 95$\% $ Confidence Bands

Quantile Processes with 95% Confidence Bands


In the first panel, you can see that the effect of logProtime has a negative effect over the lower quantiles, which diminishes in magnitude at the median and upper quantiles. This insight would be missed with the accelerated failure model.