RapTLR

Rapid Top-Line Report - An R package for streamlined creation of clinical study reports with dynamic text generation and embedded appendices.

Overview

RapTLR is designed to enhance the generation of time-sensitive formal reports, particularly clinical study reports. It provides tools to:

• Manipulate Microsoft Word (.docx) documents programmatically in R
• Replace template keywords with dynamically generated text and statistics
• Generate formatted narrative text for clinical reports
• Create appendices with linked Tables, Listings, and Figures (TLFs)
• Build natural language summaries from CDISC ADaM datasets

Installation

# Install from GitHub
# install.packages("devtools")
devtools::install_github("johnsonandjohnson/RapTLR")

Quick Start

1. Import the Demo Script

Have a look at the complete demo and process from appendice creation to TLR generation. Run import_demo() and a script called demo_TLR.R will be created in your current working directory.

library(RapTLR)

# Copy the demo script to your working directory
import_demo()

2. Basic Keyword Replacement

Replace placeholder keywords in Word documents with dynamic content:

library(RapTLR)

# Load a template document
docx_path <- system.file("extdata", "Keywords_replace.docx", package = "RapTLR")

# Replace a keyword with new text
textReplace(
  doc_object = docx_path,
  keyword = "TTsubjectTT",
  replacement = "150 participants were enrolled"
)

Chain multiple replacements using the pipe operator:

library(officer)

docx_path <- system.file("extdata", "Keywords_replace.docx", package = "RapTLR")

read_docx(docx_path) |>
  textReplace(keyword = "TTsubjectTT", replacement = "150 participants") |>
  textReplace(keyword = "TTsiteTT", replacement = "across 25 sites") |>
  textReplace(keyword = "TTconclusionTT",
              replacement = "The primary endpoint was met",
              return_to_file = "final_report.docx")

3. Generate Statistical Summaries

Create formatted narrative text from clinical data:

data(tlr_adsl)

# Generate baseline statistics text
fct_bsl_stats(
  arg_dataset = tlr_adsl,
  arg_grp = ARM,
  arg_var_num = AGE,
  arg_var1 = "mean",
  arg_var2 = "sd",
  arg_unit = "years",
  arg_label = "age"
)
# Output: "For age the overall mean was 75.1 years (SD 8.2), for Placebo..."

4. Create Appendices with TLFs

Automatically build appendices with linked tables, listings, and figures:

path_TLFs <- system.file("extdata/TLF_outputs", package = "RapTLR")
path_docx <- system.file("extdata/Keywords_replace.docx", package = "RapTLR")
TLF_list <- system.file("extdata/TLF_list.xlsx", package = "RapTLR")

run_apdx(
  path_TLFs = path_TLFs,
  docx_object = path_docx,
  sections_structure = TLF_list,
  return_to_file = "report_with_appendix.docx"
)

String Concatenation Operators

RapTLR provides custom infix operators for building narrative text:

Operator	Separator	Example	Result
%c%	space	"Hello" %c% "World"	"Hello World"
%c,%	comma	"red" %c,% "blue"	"red, blue"
%c&%	and	"efficacy" %c&% "safety"	"efficacy and safety"
%c.%	period	"First" %c.% "Second"	"First. Second"

Build complex sentences naturally:

"The study enrolled" %c% "120 participants" %c.%
"Most were" %c% "White (85%)" %c&% "female (62%)"
# "The study enrolled 120 participants. Most were White (85%) and female (62%)"

Function Reference

Document Manipulation

Function	Description
textReplace()	Replace keywords in Word documents with dynamic text
run_apdx()	Create appendices with linked TLF files
import_demo()	Copy overall demo script to working directory

Statistical Summary Functions

Function	Description
fct_bsl_stats()	Generate baseline statistics text (mean/median with SD/range)
fct_grp_tbl()	Count unique subjects by grouping variable
fct_mst_aes()	Summarize most common adverse events by treatment group
fct_smpl_rest()	Count records with custom text formatting
fct_var_text()	Create formatted text describing variable distribution
var_text_group()	Generate grouped variable summary text

String Operators

Operator	Description
%c%	Concatenate with space separator
%c,%	Concatenate with comma separator
%c&%	Concatenate with "and" separator
%c.%	Concatenate with period separator

Included Datasets

RapTLR includes two example clinical trial datasets in CDISC ADaM format:

• tlr_adsl: Subject-level analysis dataset with demographics and baseline characteristics
• tlr_adae: Adverse events analysis dataset with treatment-emergent adverse events

data(tlr_adsl)
data(tlr_adae)

Example Templates

The package includes template files accessible via system.file():

# Word template with keywords
system.file("extdata", "Keywords_replace.docx", package = "RapTLR")

# TLR shell document
system.file("extdata", "TLR_Shell.docx", package = "RapTLR")

# Appendix structure files
system.file("extdata", "TLF_list.xlsx", package = "RapTLR")
system.file("extdata", "TLF_list.csv", package = "RapTLR")

# Example TLF outputs
system.file("extdata/TLF_outputs", package = "RapTLR")

Complete Workflow Example

library(RapTLR)
library(officer)
library(dplyr)

# Load clinical data
data(tlr_adsl)
data(tlr_adae)

# Step 1: Create appendix with TLFs
path_TLFs <- system.file("extdata/TLF_outputs", package = "RapTLR")
path_docx <- system.file("extdata/TLR_Shell.docx", package = "RapTLR")
TLF_list <- system.file("extdata/TLF_list.xlsx", package = "RapTLR")

run_apdx(
  path_TLFs = path_TLFs,
  docx_object = path_docx,
  sections_structure = TLF_list,
  return_to_file = "TLR_with_appendix.docx"
)

# Step 2: Generate dynamic text for report sections
TLR <- read_docx("TLR_with_appendix.docx")

# Demographics section
demo_text <- fct_smpl_rest(
  arg_dataset = tlr_adsl,
  arg_text_desc = "This study randomized ",
  arg_text_end = " participants"
) %c%
fct_var_text(
  arg_dataset = tlr_adsl,
  arg_var_rest = TRT01A,
  arg_lbl = "nbr-pct"
) %c.%
fct_bsl_stats(
  arg_dataset = tlr_adsl,
  arg_grp = ARM,
  arg_var_num = AGE,
  arg_var1 = "median",
  arg_var2 = "range",
  arg_unit = "years",
  arg_label = "age"
)

# Step 3: Replace keywords with generated text
textReplace(TLR, "TT_geninfo_TT", demo_text)

# Safety section
safety_text <- fct_mst_aes(
  arg_data = tlr_adae,
  arg_data_adsl = tlr_adsl,
  arg_grp = ARM,
  arg_var = AEBODSYS,
  arg_trt_flt = `Xanomeline High Dose`,
  arg_frq = 10
)

textReplace(TLR, "TT_mosttaes_TT", safety_text,
            return_to_file = "final_TLR.docx")

Dependencies

RapTLR relies on the following packages:

• officer - Word document manipulation
• dplyr - Data transformation
• stringr - String manipulation
• purrr - Functional programming
• rlang - Non-standard evaluation
• xml2 - XML parsing
• crosstable - Text formatting
• readxl - Excel file reading
• magrittr - Pipe operator
• uuid - Unique identifiers

License

MIT License

Authors

• Xiang Li
• Lauren Crow
• Caesar Li
• Yannick Vandendijck
• Antoine Stos