topic Re: cumulative percentage plot of event over time (multiple arms) in SAS Procedures

cumulative percentage plot of event over time (multiple arms)

fengyuwuzu — Thu, 18 Jan 2018 16:17:25 GMT

I want to plot the cumulative percentage of event over time, and have multiple arms, so I want to plot the curves in the single graph.

Time is continuous. event is 0 or 1.

Can anyone recommend an example how to do this?

Re: cumulative percentage plot of event over time (multiple arms) [how to improve your question]

Community_Guide — Thu, 18 Jan 2018 16:25:10 GMT

Hello @fengyuwuzu,

Your question requires more details before experts can help. Can you revise your question to include more information?

Review this checklist:

Specify a meaningful subject line for your topic. Avoid generic subjects like "need help," "SAS query," or "urgent."
When appropriate, provide sample data in text or DATA step format. See this article for one method you can use.
If you're encountering an error in SAS, include the SAS log or a screenshot of the error condition. Use the Photos button to include the image in your message.
It also helps to include an example (table or picture) of the result that you're trying to achieve.

To edit your original message, select the "blue gear" icon at the top of the message and select Edit Message. From there you can adjust the title and add more details to the body of the message. Or, simply reply to this message with any additional information you can supply.

SAS experts are eager to help -- help them by providing as much detail as you can.

This prewritten response was triggered for you by fellow SAS Support Communities member @ballardw

Re: cumulative percentage plot of event over time (multiple arms)

ballardw — Thu, 18 Jan 2018 16:29:29 GMT

Example data

Which percentages you want calculated and plotted

The basic plot would be created once the data is summarized assuming you have variable to indicate which "arm" (what ever that may mean) the represents with:

proc sgplot data=want;

series x=time y=percentagevalue/group=arm;

run;

Without seeing the data exactly how to calculate your percentages is problematical though Proc Freq is the first thing that comes to mind to do cumulative percentages of counts.

Re: cumulative percentage plot of event over time (multiple arms)

fengyuwuzu — Thu, 18 Jan 2018 20:20:50 GMT

data have1;
 input ID $ arm $ event days ;
 datalines;
 001 A 1 10.5
 002 B 1 49.7
 003 A 1 14.5
 004 B 1 28.5
 005 A 1 37.9
 006 B 1 41.7
 007 A 1 18.3
 008 A 1 26.5
 009 A 1 39.3
 010 B 1 51.9
 011 B 1 31.7
 012 A 1 17.4
 013 B 1 25.3
 014 A 1 47.4
 015 A 1 15.7
 016 B 1 34.3
 017 B 1 47.5
 018 A 1 56.1
 019 B 1 27.8
 020 A 1 17.4
 021 B 1 43.3
 022 A 1 52.7
 023 A 1 64.8
 024 B 1 43.3
 025 A 1 52.7
 026 A 1 64.8
 027 B 1 43.3
 028 B 1 52.7
 029 B 1 64.8
 030 A 1 44.9
;
 run;


 data have2;
 input ID $ arm $ event days ;
 datalines;
 001 A 1 10.5
 002 B 1 49.7
 003 A 0 14.5
 004 B 0 28.5
 005 A 1 37.9
 006 B 1 41.7
 007 A 0 18.3
 008 A 1 26.5
 009 A 1 39.3
 010 B 0 51.9
 011 B 1 31.7
 012 A 1 17.4
 013 B 0 25.3
 014 A 0 47.4
 015 A 1 15.7
 016 B 1 34.3
 017 B 1 47.5
 018 A 1 56.1
 019 B 0 27.8
 020 A 0 17.4
 021 B 1 43.3
 022 A 0 52.7
 023 A 1 64.8
 024 B 1 43.3
 025 A 1 52.7
 026 A 0 64.8
 027 B 0 43.3
 028 B 1 52.7
 029 B 0 64.8
 030 A 0 44.9
;
 run;

In data have1, every one has event (event=1);

in data have2, only some have event (event=1)

want to get a plot of cumulative percentage of event in both situation

Re: cumulative percentage plot of event over time (multiple arms)

ballardw — Thu, 18 Jan 2018 20:53:20 GMT

Does this do something like what you want for Have1?

proc sort data=have1;
   by arm days;
run;
proc freq data=have1;
   by arm;
   tables days/out=h1freq outcum;
run;

proc sgplot data=h1freq;
   series x=days y=cum_pct/group=arm;
   label cum_pct='Cumulative  Percentage';
run;

If that is close then the main difference for doing have2 would be to include a WHERE event=1; in the Proc freq.

Re: cumulative percentage plot of event over time (multiple arms)

fengyuwuzu — Fri, 19 Jan 2018 04:05:35 GMT

Thank you very much ballardw.

Yes, for the first one, it is what I want.

I did not know proc freq can be used for continuous variable (in my case the time is continuous).

for the second one, is it possible to include the event=0 in the denominator, so the overall % will not reach 100% in the end?