How do I zoom in the survival graph with many group for analysis

[RAVI2000]

Hello Everyone,

I have created a survival graph with multiple groups as below.

As you look here it is a bit clustered at the top of the graph. 

Is there any procedure or way to zoom the graph? so can analyze properly.

ODS PDF file='survival_Ethnicity white race by SES.pdf';
ods graphics on / imagefmt=png imagename= "Surival for Ethnicity by White Race and SES";
proc lifetest data = adsl3 method=KM plots=s(test) 
		timelist = 0 1 2 5 10 15 20 25 30 CONFTYPE= linear;
	time Years*death(0);
	strata race_ses_white;
title1 "Overall Survival Ethnicity by White Race by SES";
run ;

[whymath]

height= option at ods graphics can modify the height of graph.
And locate the legend at right bottom inside the axis area may helps, too.

[ballardw]

May also want to set the graph to make a wider display using Ods graphics /width = <something>.

Or create a data set and use that data for SGPLOT where you can apply more controls.

[RAVI2000]

options nodate nonumber; 
ods escapechar='^';
ods listing close;
ods graphics on;
ods output survivalplot = sj.survplot;
proc lifetest data = sj.adsl3 method = km plots=s(test) 
		timelist = 0 1 2 5 10 15 20 25 30 conftype = linear;
	time years * death(0);
	strata ses;
run;


ods tagsets.rtf file = "/home/Projects/programs/exhibit_01.rtf";
ods graphics on / height = 8in width = 9in;
title1 j=c h=12pt font = arial "Overall Survival by SES";

proc sgplot data = sj.survplot;
		step x=time y=survival /group=stratum name='s' lineattrs=(pattern=solid);
	xaxis label="Years" values=(0 to 5 by 1);
	yaxis label="Survival Probability" values=(0 to 1 by 0.2);
	scatter x=time y=censored / markerattrs=(symbol=plus) name='c';
	scatter x=time y=censored / markerattrs=(symbol=plus) group=stratum;
	keylegend 'c'/location=inside position=topright;
	keylegend 's'/location=inside position=bottomleft;
run; 

ods tagsets.rtf close;
ods listing;

I have given height and width option in ODS but still not getting desired output zoomed.

I still feel the graph is clustered. Any other options @ballardw @Ksharp @whymath ? Thankyou!

[djrisks · Accepted solution]

Hi @RAVI2000,

Perhaps you can use the MIN option in the YAXIS statement to make the graph go from 0.4 to 1.

Please see the example below:

options nodate nonumber; 
ods escapechar='^';
ods listing close;
ods graphics on;
ods output survivalplot = sj.survplot;
proc lifetest data = sj.adsl3 method = km plots=s(test) 
		timelist = 0 1 2 5 10 15 20 25 30 conftype = linear;
	time years * death(0);
	strata ses;
run;


ods tagsets.rtf file = "/home/Projects/programs/exhibit_01.rtf";
ods graphics on / height = 8in width = 9in;
title1 j=c h=12pt font = arial "Overall Survival by SES";

proc sgplot data = sj.survplot;
		step x=time y=survival /group=stratum name='s' lineattrs=(pattern=solid);
	xaxis label="Years" values=(0 to 5 by 1);
	yaxis label="Survival Probability" values=(0.4 to 1 by 0.2);
	scatter x=time y=censored / markerattrs=(symbol=plus) name='c';
	scatter x=time y=censored / markerattrs=(symbol=plus) group=stratum;
	keylegend 'c'/location=inside position=topright;
	keylegend 's'/location=inside position=bottomleft;
run; 

ods tagsets.rtf close;
ods listing;

[JeffMeyers]

To add to this you might want to turn off the censor markers (not plot the SCATTER statements above). Your graphs have a lot of censor indicators making the lines into what Dr. Therneau would call "Wooly Bear" lines. If you remove them it will be easier to see the lines for comparison and they aren't adding much information to the plot currently.

[RAVI2000]

I wonder I cannot remove the censor as this is a Spatial data analysis. I have a lot of data. Please correct me.

[JeffMeyers]

I don't think they add much information to the plot and make the lines hard to read. Another option you can do is to add a Number of cumulative censors to the bottom of the graph (like a patient's at risk type of table) if you want to keep track of where and how many censors have occurred.

[RAVI2000]

Hello @djrisks,

Can you convert this into a macro to generate multiple outputs? meaning how can I give multiple destination files?

Thanks!

[Ksharp]

Maybe could try this :

yaxis label="Survival Probability" values=(0.5 to 1 by 0.1);

[Ksharp]

Check this macro and option "height=7in" :

https://communities.sas.com/t5/SAS-Communities-Library/Kaplan-Meier-Survival-Plotting-Macro-NEWSURV/...

[JeffMeyers]

Nice to see a shout out to my macro. @RAVI2000 if you want to try using it feel free to ask me any questions you may have.

[Ksharp]

Your macro is awesome . The problem I meet is can't display Chinese . and I have to use option FONTFAMILY= 宋体 to display Chinese characters .

And I need change "+ Censored" in legend into Chinese "+ 删失" .and I have to modify your macro original code to get that .