MJacobs1985

summary(jointFit1)
str(fitLME)
par(mfrow = c(4,4))
traceplot(jointFit1)
ggtraceplot(jointFit1, "alphas")
ggdensityplot(jointFit1, "alphas")
gelman_diag(jointFit1, "alphas")
densplot(jointFit1, "alphas")
cumuplot(jointFit1, "alphas")

MJacobs1985

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,

MJacobs1985

try2<-total%>%dplyr::filter(!Patientnr%in%8 & !Patientnr%in%20)
try2$meds  <- c(0,0,1,0,1,0,0,0,0,1,0,0,0,1,0,0,0,0,1,1,0,0,1,0,0,0,0,0) 
try2$RT <- recode(try2$Total_dose_spec, "1:2=1; 3=2")
table(try2$RT)
class(try2$RT)
fitLME <- lme(Severity ~ ns(time,3) + 
              Gender + 
              NRS_WorstPain_at_consultation + 
              Age_at_consultation + 
              Opioid_spec + 

MJacobs1985

## Linearity assumption
resMart<-residuals(fitSURV, type="martingale")
plot(pancreas2$NRS_WorstPain_at_consultation, resMart, main="Martingale residuals for Pain", ylab="Residuals", xlab="Ca4"); lines(loess.smooth(pancreas2$NRS_WorstPain_at_consultation, resMart), lwd=2, col="blue"); abline(h=0, col="red", lty=2, lwd=1.5)
plot(pancreas2$Age_at_consultation, resMart, main="Martingale residuals for Age", ylab="Residuals", xlab="Ca4"); lines(loess.smooth(pancreas2$Age_at_consultation, resMart), lwd=2, col="blue"); abline(h=0, col="red", lty=2, lwd=1.5)
## Influential observations
dfbetas<-residuals(fitSURV, type="dfbetas") 
par(mfrow=c(4,3))
plot(dfbetas[,1], type='h',  main="dfBETAS for GenderMale", ylab="DfBETAS", lwd=2);abline(h=c(2/sqrt(dim(dfbetas)[1]),-(2/sqrt(dim(dfbetas)[1]))),lty=2, col="red")
plot(dfbetas[,2], type='h',  main="dfBETAS for NRS Worst Pain", ylab="DfBETAS", lwd=2);abline(h=c(2/sqrt(dim(dfbetas)[1]),-(2/sqrt(dim(dfbetas)[1]))),lty=2, col="red")
plot(dfbetas[,3], type='h',  main="dfBETAS fo

MJacobs1985

# Log-log plot for categorized predictor
KM_Surv             <-with(pancreas2, Surv(SurvTime,Event==1))
KM_by_Tumorsite     <-npsurv(KM_Surv~Tumorsite_in_pancreas, data=pancreas2, type="kaplan-meier", conf.type="log-log")
KM_by_DiseaseStage  <-npsurv(KM_Surv~Disease_stage_at_consultation, data=pancreas2, type="kaplan-meier", conf.type="log-log")
KM_by_Gender        <-npsurv(KM_Surv~Gender, data=pancreas2, type="kaplan-meier", conf.type="log-log")
KM_by_Non_Opioid    <-npsurv(KM_Surv~Non_opioid_spec, data=pancreas2, type="kaplan-meier", conf.type="log-log")
KM_by_Opioid        <-npsurv(KM_Surv~Opioid_spec, data=pancreas2, type="kaplan-meier", conf.type="log-log")
KM_by_Surgery       <-npsurv(KM_Surv~Surgery_prior_to_consultation, data=pancreas2, type="kaplan-meier", conf.type="log-log")
KM_by_TotalDose     <-npsurv(KM_Surv~RT, data=pancreas2, type="kaplan-meier", conf.type="log-log")
par(mfrow=c(4,2))

MJacobs1985

PHassump <- cox.zph(fitSURV)
PHassump
par(mfrow=c(5,2))
plot(PHassump[1], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for Gender"); abline(h=0, lty=2, lwd=1.5, col="red")
plot(PHassump[2], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for NRS"); abline(h=0, lty=2, lwd=1.5, col="red")
plot(PHassump[3], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for Disease Stage"); abline(h=0, lty=2, lwd=1.5, col="red")
plot(PHassump[4], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for Age"); abline(h=0, lty=2, lwd=1.5, col="red")
plot(PHassump[5], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for Tumorsite"); abline(h=0, lty=2, lwd=1.5, col="red")
plot(PHassump[6], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for Non-Opioid"); abline(h=0, lty=2, lwd=1.5, col="red")
plot(PHassump[7], xlab="Time", ylab="Residual", main="Schoenfeld residuals (scaled) for Opioid"); abline(h=0, lty=2, lwd=1.5, col="red")

MJacobs1985

pancreas2<-pancreas%>%dplyr::filter(!Patientnr%in%8 & !Patientnr%in%20) 
pancreas2$meds <- c(0,0,1,0,1,0,0,0,0,1,0,0,0,1,0,0,0,0,1,1,0,0,1,0,0,0,0,0)  
pancreas2$RT <- car::recode(pancreas2$Total_dose_spec, "1:2=1; 3=2")
table(pancreas2$RT)
class(pancreas2$RT)
fitSURV<-coxph(Surv(SurvTimeWeek, Event) ~ 
                 Gender+
                 NRS_WorstPain_at_consultation+
                 Disease_stage_at_consultation+
                 Age_at_consultation+

MJacobs1985

vis_miss(total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38)])
gg_miss_upset(total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38)])
gg_miss_upset(total)
n_var_miss(total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38)])
gg_miss_var(total)
gg_miss_var(total,show_pct = TRUE)
gg_miss_var(total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38,79)],show_pct = TRUE, facet=timef)
gg_miss_case(total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38,79)],show_pct = TRUE, facet=timef)
gg_miss_case(total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38,49)],show_pct = TRUE, facet=Disease_stage_at_consultation)
gg_miss_fct(x=total[,c(3,6,9,12,15,19,20,23,26,29,32,35,38,79)], fct = timef)

MJacobs1985

histogram(~ SurvTimeWeek | is.na(Physical_Functioning), data = total)  
histogram(~ Emotional_Functioning | is.na(Physical_Functioning), data = total)  
histogram(~ SurvTimeWeek | timef*is.na(Interference), data = total,type = "density",
          panel = function(x, ...) {
            panel.histogram(x, ...)
            panel.mathdensity(dmath = dnorm, col = "black",
                              args = list(mean=mean(x),sd=sd(x)))
          } )
histogram(~ Severity | timef*is.na(Relief), data = total) 
histogram(~ Severity | timef*is.na(Pain), data = total) 

MJacobs1985

ggplot(total,aes(x = Relief,y = Severity)) + geom_miss_point() + facet_wrap(~timef) + theme_bw() # not sure this helps, a lof missings deletes already
ggplot(total,aes(x = Age_at_consultation,y = Overall_QOL)) + geom_miss_point() + facet_wrap(~Disease_stage_at_consultation)+theme_bw()
ggplot(total,aes(x = Severity,y = Overall_QOL)) + geom_miss_point() + theme_bw()