Closed amirabadiza921 closed 1 year ago
It is not clear what exactly the issue is. If the model has converged, the magnitude of the coefficients needs to be interpreted, accounting for the scale of the longitudinal outcome. For example, is the random slopes are very small in magnitude per subject, it is logical to get a large coefficient for the slope.
Hi Dr. Rizopoulos, When I run the current values+ current slopes, I get point estimates of assoct = 4.27 (SE = 1.03) and assoct.s = 28.10 (SE=4.86).
model.cox<-coxph(Surv(surv_time,incidenceCKD)~sex15_1+age+FHD+ waist+fbs+bs2HR+BMI,data=data.cox,x=T,model=TRUE) fit.lm1<-lme(FT4~time,random= ~time | code,data=dataLong) ffm<-list("FT4" = ~ value(FT4) + slope(FT4)) fit.JM<-jm(model.cox,fit.lm1,functional_forms = ffm,time_var="time") summary(fit.JM) Thanks