Turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- Home
- /
- Analytics
- /
- Stat Procs
- /
- How can MST exists in Proc Lifetest when more than 50% of observations...

Options

- RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page

🔒 This topic is **solved** and **locked**.
Need further help from the community? Please
sign in and ask a **new** question.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Posted 08-17-2018 09:15 AM
(2658 views)

Hello,

I'm using the following code to perform a survival analysis, just looking for the general survival information for my cases:

```
proc lifetest data = combine2;
time days_survived*censor(1);
run;
```

Here, I have 42 cases, 6 with missing censor information (and thus were deleted), and of the remaining 36, 63.89% were censored. This seems to be done appropriately given the output box I received from proc lifetest below:

It is my understanding that when over 50% of your cases are censored the median survival time does not exist. Using the proc lifetest code above, in which over 50% of my cases are censored, I received the following output:

Here is the other survival output in case it helps:

How is it that Proc Lifetest can compute a median survival time when more than 50% of the cases are censored? I'm using SAS 9.4.

I'd appreciate your help!

1 ACCEPTED SOLUTION

Accepted Solutions

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Hello @whitknee48 and welcome to the SAS Support Communities!

More than 50% of observations may be censored, but whether median survival time can be estimated or not also depends on *when* they were censored. The formula for the median survival time (see section "Breslow, Fleming-Harrington, and Kaplan-Meier Methods" in the documentation) uses *Kaplan-Meier estimates* (you used the default, KM, of option METHOD= of the PROC LIFETEST statement). Their values can very well drop below 0.5 if many subjects were censored *early* (as is the case in your data). Then the median survival time *can* be calculated. (Censored subjects are not expected to survive indefinitely.)

PROC LIFETEST usually plots a survival curve by default. So, you should have got this graph (showing that the survival curve clearly drops below 0.5):

If you added, say, 1000 to the days_survived values of sufficiently many censored observations, you'd see the difference in the survival curve and get a missing MST.

PS: Next time please post test data in the form of a data step, not as pictures.

2 REPLIES 2

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Hello @whitknee48 and welcome to the SAS Support Communities!

More than 50% of observations may be censored, but whether median survival time can be estimated or not also depends on *when* they were censored. The formula for the median survival time (see section "Breslow, Fleming-Harrington, and Kaplan-Meier Methods" in the documentation) uses *Kaplan-Meier estimates* (you used the default, KM, of option METHOD= of the PROC LIFETEST statement). Their values can very well drop below 0.5 if many subjects were censored *early* (as is the case in your data). Then the median survival time *can* be calculated. (Censored subjects are not expected to survive indefinitely.)

PROC LIFETEST usually plots a survival curve by default. So, you should have got this graph (showing that the survival curve clearly drops below 0.5):

If you added, say, 1000 to the days_survived values of sufficiently many censored observations, you'd see the difference in the survival curve and get a missing MST.

PS: Next time please post test data in the form of a data step, not as pictures.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Thanks @FreelanceReinh. This was very helpful!

Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. **Registration is now open through August 30th**. Visit the SAS Hackathon homepage.

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.