BookmarkSubscribeRSS Feed
esr_79
Calcite | Level 5

Hi, I'm working on my dissertation and need to obtain individual survival estimates for everyone in my cohort.  I've played around with several output commands but haven't figured out exactly what I need (i.e. baseline, xbeta, etc.). Does anyone have any advice or suggestions? One other helpful bit of information is that none of my covariates are time dependent.

Any help or advice is appreciated.

5 REPLIES 5
Reeza
Super User

If you want what the model estimates try looking at the outsurv= data set or basically the survival estimates.

esr_79
Calcite | Level 5

Thank you for your suggestion. From what I understand, the outsurv statement accompanies PROC LIFETEST, not PHREG. I need to obtain the probability of death, adjusted by covariates from my dataset. With the BASELINE statement for PHREG I have tried the SURVIVAL = _ALL_ option but my output appeared to be each covariate's mean by days of observation with survival function estimates, standard error and upper and lower confidence bounds for the survival estimate.

Reeza
Super User

Then use the appropriate method for proc phreg to obtain the survival estimates.

From the doc:

output out=estimates survival;

OUTPUT Statement

OUTPUT <OUT=SAS-data-set> < keyword=name ...keyword=name> </options> ;

The OUTPUT statement creates a new SAS data set containing statistics calculated for each observation. These can include the estimated linear predictor () and its standard error, survival distribution estimates, residuals, and influence statistics. In addition, this data set includes the time variable, the explanatory variables listed in the MODEL statement, the censoring variable (if specified), and the BY, STRATA, FREQ, and ID variables (if specified).

esr_79
Calcite | Level 5

Thank you for the response. I've explored the output function, too. My baseline statement is as follows:

    

     baseline out = dataset survival = surv;

After looking at my output dataset, it appears that each covariate's values are mean values -- which isn't what I expected. Is this something I should be concerned with?  With the survival function estimate by length of observation, I could go back in my original data set and merge in my estimates with observations that accumulate that length of observation, correct?  I think I was thrown by expecting to see in my output dataset each patient identifier with respective 1's and 0's according to their covariate value and a survival estimate at the end.  What's going on with the mean covariate values in the output dataset? Do you know anything about that?

Again, thank you for your responses.

Reeza
Super User

Use the output statement NOT the baseline statement.

Here's a worked example from the SAS documentation. Your output dataset of interest is surv_data and the survival estimate is in a variable called surv_estimate. Read the doc for further information and the examples are also extremely helpful.

data Myeloma;

   input Time VStatus LogBUN HGB Platelet Age LogWBC Frac

         LogPBM Protein SCalc;

   label Time='Survival Time'

         VStatus='0=Alive 1=Dead';

   datalines;

1.25  1  2.2175   9.4  1  67  3.6628  1  1.9542  12  10

1.25  1  1.9395  12.0  1  38  3.9868  1  1.9542  20  18

2.00  1  1.5185   9.8  1  81  3.8751  1  2.0000   2  15

2.00  1  1.7482  11.3  0  75  3.8062  1  1.2553   0  12

2.00  1  1.3010   5.1  0  57  3.7243  1  2.0000   3   9

3.00  1  1.5441   6.7  1  46  4.4757  0  1.9345  12  10

5.00  1  2.2355  10.1  1  50  4.9542  1  1.6628   4   9

5.00  1  1.6812   6.5  1  74  3.7324  0  1.7324   5   9

6.00  1  1.3617   9.0  1  77  3.5441  0  1.4624   1   8

6.00  1  2.1139  10.2  0  70  3.5441  1  1.3617   1   8

6.00  1  1.1139   9.7  1  60  3.5185  1  1.3979   0  10

6.00  1  1.4150  10.4  1  67  3.9294  1  1.6902   0   8

7.00  1  1.9777   9.5  1  48  3.3617  1  1.5682   5  10

7.00  1  1.0414   5.1  0  61  3.7324  1  2.0000   1  10

7.00  1  1.1761  11.4  1  53  3.7243  1  1.5185   1  13

9.00  1  1.7243   8.2  1  55  3.7993  1  1.7404   0  12

11.00  1  1.1139  14.0  1  61  3.8808  1  1.2788   0  10

11.00  1  1.2304  12.0  1  43  3.7709  1  1.1761   1   9

11.00  1  1.3010  13.2  1  65  3.7993  1  1.8195   1  10

11.00  1  1.5682   7.5  1  70  3.8865  0  1.6721   0  12

11.00  1  1.0792   9.6  1  51  3.5051  1  1.9031   0   9

13.00  1  0.7782   5.5  0  60  3.5798  1  1.3979   2  10

14.00  1  1.3979  14.6  1  66  3.7243  1  1.2553   2  10

15.00  1  1.6021  10.6  1  70  3.6902  1  1.4314   0  11

16.00  1  1.3424   9.0  1  48  3.9345  1  2.0000   0  10

16.00  1  1.3222   8.8  1  62  3.6990  1  0.6990  17  10

17.00  1  1.2304  10.0  1  53  3.8808  1  1.4472   4   9

17.00  1  1.5911  11.2  1  68  3.4314  0  1.6128   1  10

18.00  1  1.4472   7.5  1  65  3.5682  0  0.9031   7   8

19.00  1  1.0792  14.4  1  51  3.9191  1  2.0000   6  15

19.00  1  1.2553   7.5  0  60  3.7924  1  1.9294   5   9

24.00  1  1.3010  14.6  1  56  4.0899  1  0.4771   0   9

25.00  1  1.0000  12.4  1  67  3.8195  1  1.6435   0  10

26.00  1  1.2304  11.2  1  49  3.6021  1  2.0000  27  11

32.00  1  1.3222  10.6  1  46  3.6990  1  1.6335   1   9

35.00  1  1.1139   7.0  0  48  3.6532  1  1.1761   4  10

37.00  1  1.6021  11.0  1  63  3.9542  0  1.2041   7   9

41.00  1  1.0000  10.2  1  69  3.4771  1  1.4771   6  10

41.00  1  1.1461   5.0  1  70  3.5185  1  1.3424   0   9

51.00  1  1.5682   7.7  0  74  3.4150  1  1.0414   4  13

52.00  1  1.0000  10.1  1  60  3.8573  1  1.6532   4  10

54.00  1  1.2553   9.0  1  49  3.7243  1  1.6990   2  10

58.00  1  1.2041  12.1  1  42  3.6990  1  1.5798  22  10

66.00  1  1.4472   6.6  1  59  3.7853  1  1.8195   0   9

67.00  1  1.3222  12.8  1  52  3.6435  1  1.0414   1  10

88.00  1  1.1761  10.6  1  47  3.5563  0  1.7559  21   9

89.00  1  1.3222  14.0  1  63  3.6532  1  1.6232   1   9

92.00  1  1.4314  11.0  1  58  4.0755  1  1.4150   4  11

4.00  0  1.9542  10.2  1  59  4.0453  0  0.7782  12  10

4.00  0  1.9243  10.0  1  49  3.9590  0  1.6232   0  13

7.00  0  1.1139  12.4  1  48  3.7993  1  1.8573   0  10

7.00  0  1.5315  10.2  1  81  3.5911  0  1.8808   0  11

8.00  0  1.0792   9.9  1  57  3.8325  1  1.6532   0   8

12.00  0  1.1461  11.6  1  46  3.6435  0  1.1461   0   7

11.00  0  1.6128  14.0  1  60  3.7324  1  1.8451   3   9

12.00  0  1.3979   8.8  1  66  3.8388  1  1.3617   0   9

13.00  0  1.6628   4.9  0  71  3.6435  0  1.7924   0   9

16.00  0  1.1461  13.0  1  55  3.8573  0  0.9031   0   9

19.00  0  1.3222  13.0  1  59  3.7709  1  2.0000   1  10

19.00  0  1.3222  10.8  1  69  3.8808  1  1.5185   0  10

28.00  0  1.2304   7.3  1  82  3.7482  1  1.6721   0   9

41.00  0  1.7559  12.8  1  72  3.7243  1  1.4472   1   9

53.00  0  1.1139  12.0  1  66  3.6128  1  2.0000   1  11

57.00  0  1.2553  12.5  1  66  3.9685  0  1.9542   0  11

77.00  0  1.0792  14.0  1  60  3.6812  0  0.9542   0  12

;

data myeloma;

    format ID 3.;

    set myeloma;

    id=_n_;

run;

proc phreg data=Myeloma;

   model Time*VStatus(0)=LogBUN HGB Platelet Age LogWBC

                         Frac LogPBM Protein SCalc;

output out=surv_data survival=surv_estimate;

run;

proc print data=surv_data;

run;

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 5 replies
  • 1510 views
  • 0 likes
  • 2 in conversation