BookmarkSubscribeRSS Feed
Quyhuynh
Calcite | Level 5
Hello all,
Based on time-to-event ADaM dataset, I have to output CI for 10th, 25th and 50th percentiles. In PROC LIFETEST, we can easily get 25th & 50th, but not 10th.
Does anyone know how to calculate CI for 10th percentile manually?
9 REPLIES 9
Reeza
Super User

You can get that information from the OUTSURV data set. I believe you take the value just above the percentile you're looking for, since it's a step type curve.

Ksharp
Super User
proc univariate data=sashelp.heart  cipctldf cipctlnormal ;
var weight;
run;
PGStats
Opal | Level 21

Note that if your data contains censored values you should be using a proc that will take that fact into account.

 

Proc quantlife is a good choice:

 

data heart;
set sashelp.heart;
ageAtDeathCensored = ageAtDeath;
if missing(ageAtDeathCensored) then do;
    ageAtDeathCensored = AgeAtStart;
    censored = 1;
    end;
else censored = 0;
run;

proc quantlife data=heart method=na ;
model ageAtDeathCensored*censored(1) = / quantile=(0.1);
run;
                 Summary of the Number of Event and Censored Values

                                                         Percent
                       Total       Event    Censored    Censored

                        5209        1991        3218       61.78


                                 Parameter Estimates

                                     Standard    95% Confidence
Quantile  Parameter    DF  Estimate     Error        Limits        t Value  Pr > |t|

  0.1000  Intercept     1   57.7854    0.4854   56.8341   58.7367   119.06    <.0001

Note that proc quantlife can also handle regressors.

PG
Quyhuynh
Calcite | Level 5

I am sorry. I have written wrong request. Actually I need an estimated value and its lower and upper limit for 10th.

PGStats
Opal | Level 21

Above example shows how to get the 10th survival percentile estimate and confidence limits. You want to estimate something else?

PG
Quyhuynh
Calcite | Level 5

I mean the limit value at 10th percentile.

Reeza
Super User

@Quyhuynh wrote:

I mean the limit value at 10th percentile.


Did you try the QUANTLIFE approach? What values did you get and what are you expecting?

Quyhuynh
Calcite | Level 5

I figured it out. Thank you all for your kind help.

Reeza
Super User
Please post the correct answer for future users and people who were trying to answer your question.

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

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
  • 9 replies
  • 1712 views
  • 7 likes
  • 4 in conversation