run_pretty_km_output: Wrapper for KM Model Output, with Log-Rank p value

In z2thet/MoffittFunctions: Moffitt Functions

Description

This function takes a dataset, along with variables names for time and event status for KM fit, and possibly strata

Usage

run_pretty_km_output(
  strata_in = NA,
  model_data,
  time_in,
  event_in,
  event_level = NULL,
  time_est = NULL,
  group_name = NULL,
  title_name = NULL,
  conf_level = 0.95,
  surv_est_prefix = "Time",
  surv_est_digits = 2,
  median_est_digits = 1,
  p_digits = 4,
  output_type = NULL,
  sig_alpha = 0.05,
  background = "yellow",
  ...
)

Arguments

strata_in	name of strata variable, or NA (default) if no strata desired
model_data	dataset that contains strata_in, time_in, and event_in variables
time_in	name of time variable component of outcome measure
event_in	name of event status variable. If event_level = NULL then this must be the name of a F/T or 0/1 variable, where F or 0 are considered the censored level, respectively
event_level	event level for event status variable.
time_est	numerical vector of time estimates. If NULL (default) no time estimates are calculated
group_name	strata variable name. If NULL and strata exists then using variable
title_name	title to use
conf_level	the confidence level required (default is 0.95).
surv_est_prefix	prefix to use in survival estimate names. Default is Time (i.e. Time:5, Time:10,...)
surv_est_digits	number of digits to round p values for survival estimates for specified times
median_est_digits	number of digits to round p values for Median Survival Estimates
p_digits	number of digits to round p values for Log-Rank p value
output_type	output type, either NULL (default), "latex", or "html" (making special charaters latex friendly)
sig_alpha	the defined significance level. Default = 0.05
background	background color of significant values, or no highlighting if NULL. Default is "yellow"
...	other params to pass to pretty_pvalues (i.e. bold or italic)

Value

A tibble with: Name (if provided), Group (if strata variable in fit), Level (if strata variable in fit), Time:X (Survival estimates for each time provided), Median Estimate. In no strata variable tibble is one row, otherwise nrows = number of strata levels.

Examples

# Basic survival model examples
set.seed(542542522)
ybin <- sample(0:1, 100, replace = TRUE)
ybin2 <- sample(0:1, 100, replace = TRUE)
ybin3 <- sample(c('Dead','Alive'), 100, replace = TRUE)
y <- rexp(100,.1)
x1 <- factor(sample(LETTERS[1:2],100,replace = TRUE))
x2 <- factor(sample(letters[1:4],100,replace = TRUE))
my_data <- data.frame(y, ybin, ybin2, ybin3, x1, x2)
Hmisc::label(my_data$x1) <- "X1 Variable"

 # Single runs 
run_pretty_km_output(strata_in = 'x1', model_data = my_data, 
     time_in = 'y', event_in = 'ybin', time_est = NULL)
run_pretty_km_output(strata_in = 'x1', model_data = my_data, 
     time_in = 'y', event_in = 'ybin', time_est = c(5,10))
run_pretty_km_output(strata_in = 'x2', model_data = my_data, 
     time_in = 'y', event_in = 'ybin3', event_level = 'Dead', time_est = c(5,10))

# Multiple runs for different variables
library(dplyr) 
vars_to_run = c(NA, 'x1', 'x2')
purrr::map_dfr(vars_to_run, run_pretty_km_output, model_data = my_data,
     time_in = 'y', event_in = 'ybin', event_level = '0', time_est = NULL) %>% 
   select(Group, Level, everything())
   
km_info <- purrr::map_dfr(vars_to_run, run_pretty_km_output, model_data = my_data, time_in = 'y', 
     event_in = 'ybin3', event_level = 'Dead', time_est = c(5,10), surv_est_prefix = 'Year', 
     title_name = 'Overall Survival') %>% 
   select(Group, Level, everything())
   
km_info2 <- purrr::map_dfr(vars_to_run, run_pretty_km_output, model_data = my_data, time_in = 'y', 
     event_in = 'ybin2', time_est = c(5,10), surv_est_prefix = 'Year', 
     title_name = 'Cancer Specific Survival') %>% 
   select(Group, Level, everything())

options(knitr.kable.NA = '')
kableExtra::kable(bind_rows(km_info, km_info2), escape = FALSE
, longtable = FALSE, booktabs = TRUE, linesep = '', 
     caption = 'Survival Percentage Estimates at 5 and 10 Years') %>%
  kableExtra::collapse_rows(c(1:2), row_group_label_position = 'stack'
  , headers_to_remove = 1:2) 


  # Real World Example
  data(Bladder_Cancer)
  
  vars_to_run = c(NA, 'Gender', 'Clinical_Stage_Grouped', 'PT0N0', 'Any_Downstaging')
  
  purrr::map_dfr(vars_to_run, run_pretty_km_output, model_data = Bladder_Cancer, 
       time_in = 'Survival_Months', event_in = 'Vital_Status', event_level = 'Dead', 
       time_est = c(24,60), surv_est_prefix = 'Month', p_digits=5) %>% 
   select(Group, Level, everything())
  

z2thet/MoffittFunctions documentation built on July 17, 2021, 9:51 a.m.