## Author: Eric Morenz
## Date: 2020/07/22
## R basics for survival

library(dplyr) ## If you get an error. Please run: install.packages("dplyr")
library(readxl) ## To read .xls files
library(survival) ## Classic and well known

## library(flexsurv) ## mostly an extension of survival package. Not used
## library(foreign) ## only used if we want to load in a Stata dataset (.dta format) 

setwd(here::here())## install.packages("here")

### Reading our data and
### Taking a small subset of the data and renaming some variables
df <- read_excel("PS_Datasets/pbc.xls") %>% 
  select(N, X, D, Z1) %>% 
  rename(ID = N,
         Time = X,
         Death = D,
         Trt = Z1) %>% 
  filter(ID < 200) 

#### Creating a survival object. This is one way to do it.
surv_obj <- df %>% 
  with(Surv(time=Time, event=Death, type="right"))
#### time is the time to event
#### event usually follows 1 = event/death, 0 = censoring

### Basic summary
summary(surv_obj) 

### Creating a Kaplan-Meier curve
survfit_obj <- survfit(surv_obj ~ 1, data=df, 
                          conf.type = "log-log") ## 
## Kaplan-Meier plot with pointwise confidence intervals
plot(survfit_obj, col=c( 'red', 'blue', 'blue'),
     main="Kaplan-Meier Survival Estimate",
     ylab="Survival Probability", xlab="Time")
### There are ways to create KM plots with ggplot2, but they often require additonal packages