gistfile1.txt

fitLME <- lme(Severity ~ ns(time,3),           
              random = ~ time | Patientnr,            
              data=try2,
              control=lmeControl(maxIter = 10000),
              na.action=na.exclude)
fitSURV<-coxph(Surv(SurvTimeWeek, Event) ~ 1,
               data = pancreas2, x = TRUE, 
               na.action=na.exclude, 
               cluster = Patientnr)
jointFit1 <- jm(fitSURV,
                fitLME, 
                time_var = "time",
                id_var="Patientnr",
                data_Surv = pancreas2, 
                n_iter = 50000L, 
                n_burnin = 10000L, 
                n_thin = 5L, 
                n_chains=4)
try_pred<-try2%>%dplyr::select(Patientnr, time, Severity)%>%arrange(Patientnr, time)
pancreas_pred<-pancreas2%>%dplyr::select(Patientnr, SurvTimeWeek, Event)
nd<-list(newdataL = try_pred, 
         newdataE = pancreas_pred)
predLong1<-predict(jointFit1, 
                   newdata=nd,
                   return_newdata = TRUE, 
                   type = "mean_subject")
predLong1%>%
  dplyr::arrange(time)%>%
  dplyr::select(time, pred_Severity , low_Severity, upp_Severity)%>%
  dplyr::group_by(time)%>%
  dplyr::summarise(pred_m = mean(pred_Severity, na.rm=TRUE),
                   ci_l_m = mean(low_Severity, na.rm=TRUE),
                   ci_h_m = mean(upp_Severity, na.rm=TRUE))%>%
  ungroup()%>%
  ggplot(., aes(x=time, group=1))+
  geom_point(aes(x=time, y=pred_m), color = "red")+
  geom_line(aes(y=pred_m), color = "red", size=0.8)+
  geom_ribbon(aes(ymin=ci_l_m, ymax=ci_h_m), fill="red", alpha=0.2)+
  ggtitle(paste("A")) +
  theme_survminer() + 
  scale_x_continuous("Time", 
                     breaks = c(0,1,2,4,5,7,11,19,31))+
  scale_y_continuous("BPI - Severity 0-10 scale", 
                     breaks = seq(0, 10, by = 2))