-
Notifications
You must be signed in to change notification settings - Fork 0
/
Check_rcs.R
19 lines (14 loc) · 1.32 KB
/
Check_rcs.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
#######################################################################################################################
####################################### Script to check if restricted cubic splines are needed ######################3
##################################### If yes, also make the proper transformations for the model & run it ########
#######################################################################################################################
SMSCdata_glmer$nr.Gd.enhanced.lesions<-as.factor(SMSCdata_glmer$nr.Gd.enhanced.lesions)
SMSCdata_glmer$nr.relapses.2y.prior.study<-as.factor(SMSCdata_glmer$nr.relapses.2y.prior.study)
SMSCdata_glmer$treatment.naive.prior.visit<-as.factor(SMSCdata_glmer$treatment.naive.prior.visit)
SMSCdata_glmer$gender<-as.factor(SMSCdata_glmer$gender)
SMSCdata_glmer$treatment.during.cycle<-as.factor(SMSCdata_glmer$treatment.during.cycle)
SMSCdata_glmer$outcome<-as.factor(SMSCdata_glmer$outcome)
glm_rcs<-lrm((outcome)~ rcs(age,4) + rcs(disease.duration,4) + rcs(edss,4) + nr.Gd.enhanced.lesions + nr.relapses.2y.prior.study +
rcs(months.since.last.relapse,4) + treatment.naive.prior.visit + gender + treatment.during.cycle,
x=TRUE,y=TRUE,data=SMSCdata_glmer,maxit=1000)
anova(glm_rcs) #none of the variables has non-linear relationship with the outcome