SAS Day 40: Forest Plot
Tired looking across many overall response tables while comparing the same rate? The forest plot is a perfect solution to summarise data from multiple resources and convert it into a single reader-friendly graph.
Ideal Forest Plot:
When taking all the relevant studies, ask the same question with the same statistical model, identifies a common endpoint in papers and displays them on a single set of axis. The forest plot directly comparing the studies by showing the quantity and quality of the results all in one place.
(🔷 means the most important study, and ⚫ size is according to the study population size)
Done is better than perfect, so we will do a practical forest plot example.
Practical Forest Plot:
Instead of gathering all the relevant studies, we will compare the Overall Survival Rate different categories (Age group, Sex, Hispanic).
The Forest plot did not show one category has significant Overall response rate than another, but Non-Hispanic group did show a little edge in the treatment. The blue dot is the Calculated 95% Confidence Interval based on the data. The red interval is the range of values within which you can be 95% certain the true value lies.
SAS Sample Dataset
proc sort data=crr out=crr1; by descending ord; run;
*graph; options orientation=landscape; ods graphics on/ height=400px width=600px; ods pdf nogfootnote nogtitle ; title1 j=center "Graph 1"; title2 j=center " Forest Plot for Overall Response Rate"; title3 j=center "x Population"; proc sgplot data=orr1; scatter x=oc1 y=lbl /xerrorlower=clow xerrorupper=chigh markerattrs=or (symbol=CirleFilled size=4); refline 40 /axis=x; xaxis values=(0 to 100 by 20) display=(nolabel); yaxis valueattrs=(size=8pt) display=(nolabel); yaxistable c3/ label="ORR(95% CI)" location=inside ; run; goptions reset=all; ods pdf close; ods listing;
Happy Practicing and i decide to use a recycle bag to shopping today, so save some trees to make a forest plot!