How to output LogFit from sliced restricted cubic spline (created through proc logistic)?

[njgrubic]

Hi,

I was able to output a a restricted cubic spline (through effectplot statement in proc logistic) between my independent variable and binary outcome, sliced by a grouping variable. I would now like to change the axis and am trying to figure out how to save a "Logfit" output that also includes my additional grouping variable. I am then planning to recreate this graph in proc sgplot. I have noted the two main questions I have in red below. 

Thanks for your help!

proc logistic data=splines descending;
class ruca(ref="1_urbancore") / param=ref;
effect spl = spline(responsetime/ details naturalcubic 
basis=tpf(noint)knotmethod=percentiles(5));
model neurooutcome = ruca spl ruca*spl / expb cl outpred=Pred;
effectplot slicefit (x=responsetime sliceby=ruca) / noobs clm;
ods output fitplot=Logfit; (I was not able to output the LogFit with a group variable?)
run;

proc sgplot data=Logfit noautolegend; (how to I add a statement to indicate the separate series for a group variable?)
        band upper=_uclm lower=_lclm x=_xcont1 / transparency=.3;
        series y=_predicted x=_xcont1;
        xaxis values=(0 to 1200 by 100) grid offsetmin=.05 offsetmax=.05;
        yaxis values=(0 to 0.2 by 0.05) grid offsetmin=.05 offsetmax=.05 
              label="Predicted Probability of Survival with CPC 1 or 2";
        title "Effect of RT";
        run;

[Rick_SAS · Accepted solution]

I suspect the problem is that you are using the wrong name for the object that you are trying to save. It is probably SliceFitPanel rather than FitPlot.  You can use ODS TRACE ON to find out:

ods trace on;
proc logistic data= ...;
   ...
   ods output SliceFitPanel=LogFit;
run;
ods trace off;

if you are unfamiliar with ODS TRACE, see this article about how to use the ODS OUTPUT statement.