eligmodel

The eligmodel package facilitates the analysis of real-world data (RWD) by identifying and augmenting patient populations who satisfy key clinical trial eligibility criteria. It estimates the probability that RWD patients meet these criteria (i.e., weights) and supports weighted analyses for time-to-event outcomes. The package features an interactive Shiny application for implementing and visualizing the analyses. Methodological details are available at Jemielita et al. (2021).

US geographic information used in the package is licensed under public domain and can be found
here.

Installation

You can install the development version of eligmodel like so:

# via CRAN
install.packages("eligmodel")

# via GitHub
# install.packages("remotes")
remotes::install_github("MSDLLCpapers/eligmodel")

Contributions

To contribute, please open a GitHub issue with reproducible examples and sanitized logs.

Example (Shiny Application)

Step 1: Generate Example Data

This step will generate example data to be used for the Shiny application

library(eligmodel)
library(dplyr)
library(purrr)

#Generate example data
data_list = app_data_gen(n=1000, n_trials=3, n_rwd=2, miss_pct=0.1)

elig_dat_list <- data_list$elig_dat_list # A list of data.frames with example eligibility criteria for each trial/RWD combination
char_dat_list <- data_list$char_dat_list # A list of data.frames with example patient characteristics for each trial/RWD combination 
surv_dat_list <- data_list$surv_dat_list # A list of data.frames with example time-to-event outcomes for each trial/RWD combination
data_dict_list <- data_list$data_dict_list # A list of data.frames with example data dictionary for each trial/RWD combination

Step 2: Launch the Shiny application

This step will launch the interactive Shiny application to conduct relevant analyses and visualize results.

# Launch the Shiny application with time-to-event outcomes
EligModelShinyApp(elig_dat_list=elig_dat_list, char_dat_list=char_dat_list, data_dict_list=data_dict_list, surv_dat_list=surv_dat_list, color_elig="#6b7ad1", color_inelig="#99c418", color_unkn="#69B8F7", color_aug="#0C2340", surv_outcome="Overall Survival")

# Launch the Shiny application without time-to-event outcomes
EligModelShinyApp(elig_dat_list=elig_dat_list, char_dat_list=char_dat_list, data_dict_list=data_dict_list, color_elig="#6b7ad1", color_inelig="#99c418", color_unkn="#69B8F7", color_aug="#0C2340")

Step 3: Add additional data in the analysis

If additional trial or RWD is of interest, this step will add additional data to the Shiny application

# Generate example data for another trial and real-world data source
data_list_new <- app_data_gen(n=1000, n_trials=1, n_rwd=1, miss_pct=0.1) 

elig_dat_new <- data_list_new$elig_dat_list %>%
  map(~ mutate(.x, trial = "Trial new"))
char_dat_new <- data_list_new$char_dat_list %>%
  map(~ mutate(.x, trial = "Trial new"))
surv_dat_new <- data_list_new$surv_dat_list %>%
  map(~ mutate(.x, trial = "Trial new"))
data_dict_new <- data_list_new$data_dict_list %>%
  map(~ mutate(.x, trial = "Trial new"))

# Update the input data lists with the new data
elig_dat_list_new <- append(elig_dat_list, elig_dat_new)
char_dat_list_new <- append(char_dat_list, char_dat_new)
surv_dat_list_new <- append(surv_dat_list, surv_dat_new)
data_dict_list_new <- append(data_dict_list, data_dict_new)

# Relaunch the Shiny application with the new input data list
EligModelShinyApp(elig_dat_list=elig_dat_list_new, char_dat_list=char_dat_list_new, data_dict_list=data_dict_list_new, surv_dat_list=surv_dat_list_new, color_elig="#6b7ad1", color_inelig="#99c418", color_unkn="#69B8F7", color_aug="#0C2340", surv_outcome="Overall Survival")