i am using the following code proc sgpanel data=dataset noautolegend; format ....; panelby cohort / onepanel spacing=2 headerattrs=(size=14) novarname; vline timevar / response=var stat=mean limits=both markers group=grps; rowaxis label="my variable" labelattrs=(size=16) valueattrs=(size=14) ; colaxis label="time" labelattrs=(size=16) valueattrs=(size=14) ; run; quit; but i want to plot box plots over time, not the mean and CI which this is producing. I can't see any way to do this using sgpanel. I could do something like "vbox var / category=timevar" but then i lose grps (each panel should have 2 series of box plots over time, one for each group, and time is discrete/categorical) cheers!
... View more
i have two subgroups in my dataset, and i estimated kappa and obtained confidence intervals for each as per this code: proc freq data=.... order=data ; tables var1*var2 / agree noprint; weight frequency / zeros; run; is it possible to test whether the agreement (as measured by kappa) is the same within these subgroups ie Ho: kappa1 = kappa2?
... View more
i would think so, ie write the model out as it should be in the data step (with any nonlinear terms) and then see if the plot makes sense. Eg these authors found "a significant nonlinear relationship with regard to the composite of death/HF hospitalization (P = 0.0039), with a plateau of effect beyond 180 ms for the composite outcome" [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3855551/] and this is seen in the plot: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3855551/figure/EHT290F3/?report=objectonly it is a shame that authors don't routinely make programming code available on publication
... View more
but once you have the model estimates you can just calculate the HR for any range of values of the continuous variable ie do it in a data step rather than exporting from phreg edit, something like: ods output parameterestimates=out (keep=estimate parameter where=(parameter="xxxxx")); proc phreg data=xxxxx; class aaa bbb ccc; model surv*cens(1)= ...../ rl alpha=0.05; run; data .....; set out; do x=.... to .... by ....; val=exp(estimate*x); output; end; run; ??
... View more
are you at a university or company? they ought to provide the instructions. If you go into your SAS/utilities folder in the start menu you should see "renew sas software"
... View more
sort by subject and cmtrt, then maybe "proc sort data=.... nodupkey dupout=XXX" see the dataset XXX for the duplicates. Or once sorted you can identify duplicates in a data step
... View more
i deleted my comment when i realised you were restricting the random number (and thus you wouldn't get negative survival as i suggested). Your new approach sounds ideal ie informed as much as possible. I think in the industry it's not uncommon to replace missing days with 15, but that's when there's a decade between diagnosis and death. With shorter survival, as you say, the impact of how missings are handled is more apparent
... View more
seems fine. if you're unsure you could always create the indicator variables for S in a data step and then fit these in model and confirm the results correspond etc
... View more
it's a text variable? if so, maybe: format newdate date9.; day=input(substr(olddate,1,2),2.); monthtxt=substr(olddate,4,3); if monthtxt ne '' then month=month(input("01"||substr(monthtxt,1,3)||"2000",date9.)); year=input(substr(olddate,8,4),4.); if day ne . and month ne . and year ne . then newdate=mdy(month,day,year);
... View more
