add pointwise confidence interval to prediction of survival curve plot

Question

I have a data set of patients on a waiting list for an organ transplant. Event is death time. The set containts the following variables: patientID, time, status , age, gender, bmi, disease. I made a model of the survival time of a patient on the waiting list as function of the covariates age, gender, bmi and disease by using accelerated failure time models, in which the baseline distribution was given by a Lognormal distribution.

fitLog <- survreg(Surv(time, status)~age+gender+bmi+disease, dist = "lognormal", data = patients)

Now, I need to give a prediction of the survival curve for a male Fibrosis (disease = 2) patient of age 60 and bmi of 23 and I also need to add a pointwise confidence interval to this curve.

new_data <- with(patients, data.frame(gender = c(1),
                                  age = c(60),
                                  bmi = c(23),
                                  disease = c(2)))

plot(predict(fitLog, newdata=new_data, type="quantile", p=seq(.01,.99,by=.01)), seq(.99,.01,by=-.01), col="red", type="l", xlab = "Time (days)", ylab = "Survival probability (%)", main="Survival time")

So when I run this code, I get my prediction model, but now I have to add a pointwise confidence interval to this plot. How do I have to do this? I already tried conf.type but that didn't do anything.

Edit: I used the following code to calculate and plot the confidence intervals:

pred <- predict(fitLog, newdata=new_data, 
type="quantile",p=seq(.01,.99,by=.01), se=TRUE)

plot(pred$fit - 1.96*pred$se.fit, seq(.99,.01,by=-.01),col="red", type="l", 
lty=2, xlab = "Time (days)", ylab = "Survival probability (%)", 
main="Survival time")

lines(pred$fit + 1.96*pred$se.fit,seq(.99,.01,by=-.01),col="red", type="l", 
lty=2)

lines(predict(fitLog, newdata=new_data, 
type="quantile",p=seq(.01,.99,by=.01)),seq(.99,.01,by=-.01),col="red", 
type="l")

So now I can plot the confidence interval, but the question is now how to make the dotted line continue until the end of the follow up?

image of the plot

Does this help? https://stackoverflow.com/questions/9151591/how-to-plot-the-survival-curve-generated-by-survreg-package-survival-of-r or this http://blogs2.datall-analyse.nl/2016/02/19/rcode_confidence_intervals_parametric_survival_models/ or maybe try the `broom::tidy` function https://www.rdocumentation.org/packages/broom/versions/0.5.2/topics/tidy.survreg — MrFlick, May 25 '19 at 18:48
The survival probability can not be < 0 by definition. Thus lower CI can also not go below 0. — adibender, May 27 '19 at 11:27
Voting to close as unclear. Makes no sense to ask for survival curve to cross the zero line. — IRTFM, May 29 '19 at 16:17
unless the question is how to make the dotted line continue until the end of the follow up? But not clear from the question — adibender, May 31 '19 at 16:13

add pointwise confidence interval to prediction of survival curve plot

0 Answers0