Advanced Clinical Reporting with officer and flextable

Create sophisticated Word reports programmatically with R

Clinical Reporting

Word

Intermediate

Author

David Gohel (Ardata)

Overview

Intermediate Clinical Reporting Word Documents

Master the creation of sophisticated clinical reports in Word format using R with a reproducible approach. This workshop covers the complete journey from clinical data to pharmaceutical-grade report generation, focusing on advanced document structure management, complex table creation, and integration of ggplot2 visualizations.

What You’ll Learn

📄 Document structure management in Word
📊 Complex tables following pharmaceutical standards
📈 ggplot2 integration for publication-ready figures
🎨 Styling and formatting for professional outputs
🔗 Cross-references and automated numbering
✅ Complete end-to-end CSR generation

Prerequisites

Required Knowledge:

Intermediate R programming
Basic understanding of clinical trial data
Familiarity with data manipulation (dplyr)

Recommended:

Experience with ggplot2
Knowledge of clinical study report structure

Key Packages & Tools

{officer}

{flextable}

{ggplot2}

{officedown}

Microsoft Word

Workshop Materials

Resources

Workshop Website: https://ardata.fr/r-in-pharma-2025/

GitHub Code: https://github.com/ardata-fr/r-in-pharma-2025-codes

Full Tutorial: https://github.com/ardata-fr/r-in-pharma-reporting-with-officer-flextable

Part 1: Fundamentals

Introduction to {officer}

{officer} allows you to create and manipulate Word documents programmatically:

library(officer)

# Create a new document
doc <- read_docx()

# Add content
doc <- doc %>%
  body_add_par("Clinical Study Report", style = "heading 1") %>%
  body_add_par("Protocol ABC-123", style = "heading 2") %>%
  body_add_par("This report summarizes the results of...", 
               style = "Normal")

# Save
print(doc, target = "clinical_report.docx")

Introduction to {flextable}

{flextable} creates publication-ready tables with fine-grained control:

library(flextable)
library(dplyr)

# Create a demographic table
demographics %>%
  flextable() %>%
  set_header_labels(
    age = "Age (years)",
    sex = "Sex, n (%)",
    race = "Race, n (%)"
  ) %>%
  theme_vanilla() %>%
  autofit()

Part 2: Advanced Clinical Reporting

Document Structure Management

Creating a Structured CSR:

library(officer)

# Start with a template
doc <- read_docx("csr_template.docx")

# Add title page
doc <- doc %>%
  body_add_par("CLINICAL STUDY REPORT", style = "Title") %>%
  body_add_par(sprintf("Protocol: %s", protocol_id), 
               style = "Subtitle") %>%
  body_add_break()

# Table of Contents (auto-updated)
doc <- doc %>%
  body_add_toc(level = 3)

# Section structure
doc <- doc %>%
  body_add_par("1. Synopsis", style = "heading 1") %>%
  body_add_par("1.1 Study Objectives", style = "heading 2")

Complex Table Creation

Pharmaceutical-Standard Tables:

library(flextable)

# Adverse events table with nested headers
create_ae_table <- function(data) {
  data %>%
    flextable() %>%
    # Multi-level headers
    add_header_row(
      values = c("", "Treatment A", "Treatment B"),
      colwidths = c(1, 2, 2)
    ) %>%
    add_header_row(
      values = c("System Organ Class", "n", "%", "n", "%"),
      top = FALSE
    ) %>%
    # Merge cells for SOC
    merge_v(j = 1) %>%
    # Styling
    theme_booktabs() %>%
    bold(part = "header") %>%
    align(align = "center", part = "header") %>%
    align(j = 1, align = "left", part = "body") %>%
    # Borders
    hline(i = ~ SOC_change, border = fp_border()) %>%
    # Conditional formatting
    bg(i = ~ rate > 10, bg = "#FFCCCC") %>%
    # Footnotes
    add_footer_lines("SOC = System Organ Class; n = number of subjects") %>%
    autofit()
}

Headers and Footers

Customized Headers/Footers:

# Define header
header_text <- fpar(
  ftext("Protocol ABC-123", prop = fp_text(font.size = 8)),
  ftext(" | ", prop = fp_text(font.size = 8)),
  ftext("Page ", prop = fp_text(font.size = 8)),
  run_pagenum(),
  ftext(" of ", prop = fp_text(font.size = 8)),
  run_totalpages()
)

# Add to document
doc <- doc %>%
  add_header(value = as_paragraph(header_text)) %>%
  add_footer(value = as_paragraph(
    ftext("Confidential", prop = fp_text(italic = TRUE))
  ))

Cross-References

Automated Table/Figure Numbering:

# Add bookmarked table
doc <- doc %>%
  body_add_par("Demographic Characteristics", 
               style = "Table Caption") %>%
  body_bookmark("table_demographics") %>%
  body_add_flextable(demographics_table)

# Reference it elsewhere
doc <- doc %>%
  body_add_par(
    fpar(
      ftext("As shown in Table "),
      run_reference("table_demographics"),
      ftext(", the treatment groups were well balanced...")
    )
  )

Integrating ggplot2 Visualizations

Publication-Ready Figures:

library(ggplot2)

# Create plot
survival_plot <- ggplot(survival_data, 
                        aes(x = time, y = surv, color = arm)) +
  geom_step(linewidth = 1) +
  geom_ribbon(aes(ymin = lower, ymax = upper, fill = arm), 
              alpha = 0.2) +
  scale_color_manual(values = c("#0072B2", "#D55E00")) +
  scale_fill_manual(values = c("#0072B2", "#D55E00")) +
  labs(
    title = "Kaplan-Meier Survival Curves",
    x = "Time (months)",
    y = "Survival Probability",
    color = "Treatment Arm",
    fill = "Treatment Arm"
  ) +
  theme_minimal() +
  theme(
    plot.title = element_text(hjust = 0.5, face = "bold"),
    legend.position = "bottom"
  )

# Add to document
doc <- doc %>%
  body_add_par("Figure 1: Overall Survival", 
               style = "Figure Caption") %>%
  body_add_gg(value = survival_plot, 
              width = 6, height = 4) %>%
  body_add_par("Shaded areas represent 95% confidence intervals.")

Workshop Structure

Session 1: Basics (90 minutes)

{officer} fundamentals
{flextable} basics
Simple document creation
Basic styling

Session 2: Advanced Techniques (90 minutes)

Complex table structures
Section management
Headers/footers customization
Cross-reference handling

Session 3: Complete CSR Example (120 minutes)

End-to-end workflow
Template-based approach
Automated report generation
QC considerations

Practical Examples

Example 1: Demographics Table

create_demographics <- function(data) {
  summary_stats <- data %>%
    group_by(ARM) %>%
    summarise(
      N = n(),
      Age_mean = mean(AGE, na.rm = TRUE),
      Age_sd = sd(AGE, na.rm = TRUE),
      Sex_M = sum(SEX == "M"),
      Sex_F = sum(SEX == "F")
    ) %>%
    mutate(
      Age = sprintf("%.1f (%.1f)", Age_mean, Age_sd),
      `Male, n (%)` = sprintf("%d (%.1f%%)", 
                              Sex_M, 100 * Sex_M / N),
      `Female, n (%)` = sprintf("%d (%.1f%%)", 
                                Sex_F, 100 * Sex_F / N)
    ) %>%
    select(ARM, N, Age, `Male, n (%)`, `Female, n (%)`)
  
  summary_stats %>%
    flextable() %>%
    set_header_labels(ARM = "Treatment Arm") %>%
    align(j = 2:5, align = "center", part = "all") %>%
    bold(part = "header") %>%
    theme_booktabs() %>%
    autofit()
}

Example 2: Adverse Events Summary

create_ae_summary <- function(ae_data) {
  ae_summary <- ae_data %>%
    group_by(AESOC, AEDECOD, TRTA) %>%
    summarise(n = n_distinct(USUBJID)) %>%
    ungroup() %>%
    pivot_wider(names_from = TRTA, values_from = n, 
                values_fill = 0)
  
  ae_summary %>%
    flextable() %>%
    set_header_labels(
      AESOC = "System Organ Class",
      AEDECOD = "Preferred Term"
    ) %>%
    merge_v(j = "AESOC") %>%
    valign(valign = "top") %>%
    bold(j = "AESOC") %>%
    theme_booktabs() %>%
    autofit() %>%
    add_footer_lines(
      "Table includes treatment-emergent adverse events"
    )
}

Example 3: Complete Report Template

generate_csr <- function(data_list, output_path) {
  doc <- read_docx("templates/csr_template.docx")
  
  # Title page
  doc <- doc %>%
    body_add_par("CLINICAL STUDY REPORT", style = "Title") %>%
    body_add_par(data_list$protocol_id, style = "Subtitle")
  
  # ToC
  doc <- doc %>%
    body_add_break() %>%
    body_add_toc()
  
  # Demographics
  doc <- doc %>%
    body_add_break() %>%
    body_add_par("Demographics", style = "heading 1") %>%
    body_add_flextable(create_demographics(data_list$adsl))
  
  # Efficacy
  doc <- doc %>%
    body_add_break() %>%
    body_add_par("Efficacy Results", style = "heading 1") %>%
    body_add_gg(create_efficacy_plot(data_list$adtte))
  
  # Safety
  doc <- doc %>%
    body_add_break() %>%
    body_add_par("Safety Results", style = "heading 1") %>%
    body_add_flextable(create_ae_summary(data_list$adae))
  
  # Save
  print(doc, target = output_path)
}

Best Practices

✅ Do’s

Use templates for consistency
Separate data processing from formatting
Document your styling choices
Test cross-references thoroughly
Version control your templates

❌ Don’ts

Hard-code numbers in text
Mix styling approaches
Forget page breaks
Ignore Word template styles
Skip QC on final document

Advanced Features

Conditional Formatting

table %>%
  flextable() %>%
  bg(i = ~ p_value < 0.05, j = "p_value", bg = "yellow") %>%
  bold(i = ~ significant == TRUE) %>%
  color(i = ~ trend == "increasing", j = "value", color = "red")

Custom Themes

my_pharma_theme <- function(ft) {
  ft %>%
    border_remove() %>%
    hline_top(border = fp_border(width = 2)) %>%
    hline_bottom(border = fp_border(width = 2)) %>%
    font(fontname = "Arial", part = "all") %>%
    fontsize(size = 10, part = "body") %>%
    fontsize(size = 11, part = "header")
}

Learning Outcomes

By the end of this workshop, you will be able to:

✅ Create professional Word documents programmatically
✅ Build complex, pharmaceutical-standard tables
✅ Integrate ggplot2 visualizations seamlessly
✅ Manage document structure with sections and breaks
✅ Implement automated cross-referencing
✅ Generate complete CSRs from R scripts
✅ Apply best practices for reproducible reporting

Troubleshooting

Common Issues:

Tables not fitting page width → Use autofit() or manual width() settings
Plots appear pixelated → Increase DPI with body_add_gg(width, height, res = 300)
Cross-references not updating → Right-click → “Update Field” in Word
Styles not applying → Ensure template has required style definitions

Integration with Other Tools

{officedown}: R Markdown → Word via officer
{gtsummary}: Statistical tables → flextable
{officer} + {mschart}: Native Office charts
{pharmaRTF}: Alternative for RTF output

Next Steps

After this workshop:

Explore {officedown} for R Markdown integration
Try {mschart} for editable Office charts
Investigate {pharmaRTF} for RTF submissions
Join ardata community for support

Similar Workshops

Cardinal: Clinical Reporting - Standardized TLG templates
Branded Documents with Quarto - Alternative approach

Next Steps

For tables: See Cardinal workshop
For multi-format: Try Quarto workshop
Industry adoption: Automation trends

Last updated: November 2025 | R/Pharma 2025 Conference

--- title: "Advanced Clinical Reporting with officer and flextable" subtitle: "Create sophisticated Word reports programmatically with R" author: "David Gohel (Ardata)" categories: [Clinical Reporting, Word, Intermediate] --- ## Overview [Intermediate]{.badge .badge-intermediate} [Clinical Reporting]{.badge .badge-category} [Word Documents]{.badge .badge-category} Master the creation of sophisticated clinical reports in Word format using R with a reproducible approach. This workshop covers the complete journey from clinical data to pharmaceutical-grade report generation, focusing on advanced document structure management, complex table creation, and integration of ggplot2 visualizations. ### What You'll Learn - 📄 **Document structure** management in Word - 📊 **Complex tables** following pharmaceutical standards - 📈 **ggplot2 integration** for publication-ready figures - 🎨 **Styling and formatting** for professional outputs - 🔗 **Cross-references** and automated numbering - ✅ **Complete end-to-end** CSR generation ## Prerequisites ::: requirements **Required Knowledge:** - Intermediate R programming - Basic understanding of clinical trial data - Familiarity with data manipulation (dplyr) **Recommended:** - Experience with ggplot2 - Knowledge of clinical study report structure ::: ## Key Packages & Tools ::: tool-tag {officer} ::: ::: tool-tag {flextable} ::: ::: tool-tag {ggplot2} ::: ::: tool-tag {officedown} ::: ::: tool-tag Microsoft Word ::: ## Workshop Materials ::: callout-note ## Resources **Workshop Website:** <https://ardata.fr/r-in-pharma-2025/> **GitHub Code:** <https://github.com/ardata-fr/r-in-pharma-2025-codes> **Full Tutorial:** <https://github.com/ardata-fr/r-in-pharma-reporting-with-officer-flextable> ::: ## Part 1: Fundamentals ### Introduction to {officer} **{officer}** allows you to create and manipulate Word documents programmatically: ``` r library(officer) # Create a new document doc <- read_docx() # Add content doc <- doc %>% body_add_par("Clinical Study Report", style = "heading 1") %>% body_add_par("Protocol ABC-123", style = "heading 2") %>% body_add_par("This report summarizes the results of...", style = "Normal") # Save print(doc, target = "clinical_report.docx") ``` ### Introduction to {flextable} **{flextable}** creates publication-ready tables with fine-grained control: ``` r library(flextable) library(dplyr) # Create a demographic table demographics %>% flextable() %>% set_header_labels( age = "Age (years)", sex = "Sex, n (%)", race = "Race, n (%)" ) %>% theme_vanilla() %>% autofit() ``` ## Part 2: Advanced Clinical Reporting ### Document Structure Management **Creating a Structured CSR:** ``` r library(officer) # Start with a template doc <- read_docx("csr_template.docx") # Add title page doc <- doc %>% body_add_par("CLINICAL STUDY REPORT", style = "Title") %>% body_add_par(sprintf("Protocol: %s", protocol_id), style = "Subtitle") %>% body_add_break() # Table of Contents (auto-updated) doc <- doc %>% body_add_toc(level = 3) # Section structure doc <- doc %>% body_add_par("1. Synopsis", style = "heading 1") %>% body_add_par("1.1 Study Objectives", style = "heading 2") ``` ### Complex Table Creation **Pharmaceutical-Standard Tables:** ``` r library(flextable) # Adverse events table with nested headers create_ae_table <- function(data) { data %>% flextable() %>% # Multi-level headers add_header_row( values = c("", "Treatment A", "Treatment B"), colwidths = c(1, 2, 2) ) %>% add_header_row( values = c("System Organ Class", "n", "%", "n", "%"), top = FALSE ) %>% # Merge cells for SOC merge_v(j = 1) %>% # Styling theme_booktabs() %>% bold(part = "header") %>% align(align = "center", part = "header") %>% align(j = 1, align = "left", part = "body") %>% # Borders hline(i = ~ SOC_change, border = fp_border()) %>% # Conditional formatting bg(i = ~ rate > 10, bg = "#FFCCCC") %>% # Footnotes add_footer_lines("SOC = System Organ Class; n = number of subjects") %>% autofit() } ``` ### Headers and Footers **Customized Headers/Footers:** ``` r # Define header header_text <- fpar( ftext("Protocol ABC-123", prop = fp_text(font.size = 8)), ftext(" | ", prop = fp_text(font.size = 8)), ftext("Page ", prop = fp_text(font.size = 8)), run_pagenum(), ftext(" of ", prop = fp_text(font.size = 8)), run_totalpages() ) # Add to document doc <- doc %>% add_header(value = as_paragraph(header_text)) %>% add_footer(value = as_paragraph( ftext("Confidential", prop = fp_text(italic = TRUE)) )) ``` ### Cross-References **Automated Table/Figure Numbering:** ``` r # Add bookmarked table doc <- doc %>% body_add_par("Demographic Characteristics", style = "Table Caption") %>% body_bookmark("table_demographics") %>% body_add_flextable(demographics_table) # Reference it elsewhere doc <- doc %>% body_add_par( fpar( ftext("As shown in Table "), run_reference("table_demographics"), ftext(", the treatment groups were well balanced...") ) ) ``` ### Integrating ggplot2 Visualizations **Publication-Ready Figures:** ``` r library(ggplot2) # Create plot survival_plot <- ggplot(survival_data, aes(x = time, y = surv, color = arm)) + geom_step(linewidth = 1) + geom_ribbon(aes(ymin = lower, ymax = upper, fill = arm), alpha = 0.2) + scale_color_manual(values = c("#0072B2", "#D55E00")) + scale_fill_manual(values = c("#0072B2", "#D55E00")) + labs( title = "Kaplan-Meier Survival Curves", x = "Time (months)", y = "Survival Probability", color = "Treatment Arm", fill = "Treatment Arm" ) + theme_minimal() + theme( plot.title = element_text(hjust = 0.5, face = "bold"), legend.position = "bottom" ) # Add to document doc <- doc %>% body_add_par("Figure 1: Overall Survival", style = "Figure Caption") %>% body_add_gg(value = survival_plot, width = 6, height = 4) %>% body_add_par("Shaded areas represent 95% confidence intervals.") ``` ## Workshop Structure ### Session 1: Basics (90 minutes) - {officer} fundamentals - {flextable} basics - Simple document creation - Basic styling ### Session 2: Advanced Techniques (90 minutes) - Complex table structures - Section management - Headers/footers customization - Cross-reference handling ### Session 3: Complete CSR Example (120 minutes) - End-to-end workflow - Template-based approach - Automated report generation - QC considerations ## Practical Examples ### Example 1: Demographics Table ``` r create_demographics <- function(data) { summary_stats <- data %>% group_by(ARM) %>% summarise( N = n(), Age_mean = mean(AGE, na.rm = TRUE), Age_sd = sd(AGE, na.rm = TRUE), Sex_M = sum(SEX == "M"), Sex_F = sum(SEX == "F") ) %>% mutate( Age = sprintf("%.1f (%.1f)", Age_mean, Age_sd), `Male, n (%)` = sprintf("%d (%.1f%%)", Sex_M, 100 * Sex_M / N), `Female, n (%)` = sprintf("%d (%.1f%%)", Sex_F, 100 * Sex_F / N) ) %>% select(ARM, N, Age, `Male, n (%)`, `Female, n (%)`) summary_stats %>% flextable() %>% set_header_labels(ARM = "Treatment Arm") %>% align(j = 2:5, align = "center", part = "all") %>% bold(part = "header") %>% theme_booktabs() %>% autofit() } ``` ### Example 2: Adverse Events Summary ``` r create_ae_summary <- function(ae_data) { ae_summary <- ae_data %>% group_by(AESOC, AEDECOD, TRTA) %>% summarise(n = n_distinct(USUBJID)) %>% ungroup() %>% pivot_wider(names_from = TRTA, values_from = n, values_fill = 0) ae_summary %>% flextable() %>% set_header_labels( AESOC = "System Organ Class", AEDECOD = "Preferred Term" ) %>% merge_v(j = "AESOC") %>% valign(valign = "top") %>% bold(j = "AESOC") %>% theme_booktabs() %>% autofit() %>% add_footer_lines( "Table includes treatment-emergent adverse events" ) } ``` ### Example 3: Complete Report Template ``` r generate_csr <- function(data_list, output_path) { doc <- read_docx("templates/csr_template.docx") # Title page doc <- doc %>% body_add_par("CLINICAL STUDY REPORT", style = "Title") %>% body_add_par(data_list$protocol_id, style = "Subtitle") # ToC doc <- doc %>% body_add_break() %>% body_add_toc() # Demographics doc <- doc %>% body_add_break() %>% body_add_par("Demographics", style = "heading 1") %>% body_add_flextable(create_demographics(data_list$adsl)) # Efficacy doc <- doc %>% body_add_break() %>% body_add_par("Efficacy Results", style = "heading 1") %>% body_add_gg(create_efficacy_plot(data_list$adtte)) # Safety doc <- doc %>% body_add_break() %>% body_add_par("Safety Results", style = "heading 1") %>% body_add_flextable(create_ae_summary(data_list$adae)) # Save print(doc, target = output_path) } ``` ## Best Practices ### ✅ Do's - Use templates for consistency - Separate data processing from formatting - Document your styling choices - Test cross-references thoroughly - Version control your templates ### ❌ Don'ts - Hard-code numbers in text - Mix styling approaches - Forget page breaks - Ignore Word template styles - Skip QC on final document ## Advanced Features ### Conditional Formatting ``` r table %>% flextable() %>% bg(i = ~ p_value < 0.05, j = "p_value", bg = "yellow") %>% bold(i = ~ significant == TRUE) %>% color(i = ~ trend == "increasing", j = "value", color = "red") ``` ### Custom Themes ``` r my_pharma_theme <- function(ft) { ft %>% border_remove() %>% hline_top(border = fp_border(width = 2)) %>% hline_bottom(border = fp_border(width = 2)) %>% font(fontname = "Arial", part = "all") %>% fontsize(size = 10, part = "body") %>% fontsize(size = 11, part = "header") } ``` ## Learning Outcomes By the end of this workshop, you will be able to: ✅ Create professional Word documents programmatically\ ✅ Build complex, pharmaceutical-standard tables\ ✅ Integrate ggplot2 visualizations seamlessly\ ✅ Manage document structure with sections and breaks\ ✅ Implement automated cross-referencing\ ✅ Generate complete CSRs from R scripts\ ✅ Apply best practices for reproducible reporting ## Troubleshooting **Common Issues:** 1. **Tables not fitting page width** → Use `autofit()` or manual `width()` settings 2. **Plots appear pixelated** → Increase DPI with `body_add_gg(width, height, res = 300)` 3. **Cross-references not updating** → Right-click → "Update Field" in Word 4. **Styles not applying** → Ensure template has required style definitions ## Integration with Other Tools - **{officedown}:** R Markdown → Word via officer - **{gtsummary}:** Statistical tables → flextable - **{officer}** + **{mschart}:** Native Office charts - **{pharmaRTF}:** Alternative for RTF output ## Next Steps After this workshop: - Explore {officedown} for R Markdown integration - Try {mschart} for editable Office charts - Investigate {pharmaRTF} for RTF submissions - Join ardata community for support ------------------------------------------------------------------------ ### Similar Workshops - [Cardinal: Clinical Reporting](cardinal-tlgs.qmd) - Standardized TLG templates - [Branded Documents with Quarto](branded-quarto.qmd) - Alternative approach ### Related Presentations - [autoslideR](../presentations/europe-us-sessions.qmd#autoslider-streamlining-slide-deck-generation) - Automated slide decks ### Next Steps - **For tables:** See [Cardinal workshop](cardinal-tlgs.qmd) - **For multi-format:** Try [Quarto workshop](branded-quarto.qmd) - **Industry adoption:** [Automation trends](../summary/trends-insights.qmd#4-automation--efficiency-) ------------------------------------------------------------------------ *Last updated: November 2025 \| R/Pharma 2025 Conference*