From Raw Output to Professional Tables: Transforming Statistical Results with gtsummary

In this article you’ll see the way to communicate to your stakeholders the way to present the results in a clear, accessible format. The gtsummary package transform complex statistical output into publication-ready tables that speak to any audience.

The Problem with Raw Statistical Output

Traditional R output, usually, presents several challenges when communicating with broader audiences:

Issues with Base R Output:

Information overload: Raw output includes MANY technical details (residuals, diagnostic statistics, model fit information) that can overwhelm your stakeholders
Poor formatting: Console output lacks visual hierarchy and professional appearance
Technical jargon: Terms like “Std. Error,” “t value,” and “Pr(>|t|)” require statistical background to interpret
Inconsistent presentation: Different model types produce different output formats, making comparisons difficult

Real-World Scenario:

Imagine presenting regression results from a clinical trial to a medical advisory board. Raw R output might show:

Coefficients:
                    Estimate Std. Error t value Pr(>|t|)    
(Intercept)         45.2341     2.1234   21.29  < 2e-16 ***
treatment_group      8.7654     1.5678    5.59  8.9e-07 ***
age                  0.2345     0.0876    2.68   0.0089 ** 
gender_female       -2.1098     1.2345   -1.71   0.0912 .

While statisticians can quickly interpret this, clinicians need to know: “What does this mean for patient outcomes?”

What is gtsummary?

gtsummary is an R package that creates elegant, publication-ready summary tables from statistical models. It automatically formats results, adds appropriate labels, and presents findings in a way that’s immediately interpretable by diverse audiences.

Core Philosophy:

Clarity over complexity: Present only the most relevant information
Audience-appropriate: Adapt technical results for non-technical stakeholders
Professional appearance: Generate tables ready for publications, presentations, and reports
Consistency: Standardize output across different types of analyses

Key Functions and Applications

1. tbl_regression(): The Workhorse Function

The tbl_regression() function transforms raw regression output into clean, interpretable tables.

Key features: - Automatic formatting of coefficients and confidence intervals - Clear variable labeling - Optional p-values with appropriate formatting - Customizable precision and display options

Example transformation: Instead of raw coefficient output, you get a table showing: - Variable: Treatment Group - Coefficient: 8.77 (95% CI: 5.70, 11.84) - p-value: <0.001

2. tbl_summary(): Descriptive Statistics Made Simple

Creates comprehensive descriptive statistics tables that are immediately publication-ready.

Applications: - Baseline characteristics: Compare treatment groups in clinical trials - Population descriptions: Summarize study participant characteristics - Stratified analyses: Show results by subgroups

3. tbl_survfit(): Survival Analysis Results

Transforms complex survival analysis output into clear, interpretable tables showing: - Median survival times with confidence intervals - Survival probabilities at key time points - Hazard ratios with clear interpretation

Model Versatility: Beyond Linear Regression

Supported Model Types:

Linear Models: - Simple and multiple linear regression - Analysis of variance (ANOVA) - Analysis of covariance (ANCOVA)

Generalized Linear Models: - Logistic regression (odds ratios automatically calculated) - Poisson regression (incidence rate ratios) - Negative binomial regression

Advanced Models: - Cox proportional hazards models - Mixed-effects models (with appropriate packages) - Bayesian models (with brms integration)

Specialized Applications: - Dose-response analyses - Propensity score matching results - Meta-analysis summaries

Example: Logistic Regression Output

Raw R output shows log-odds coefficients that are difficult to interpret:

Coefficients:
                Estimate Std. Error z value Pr(>|z|)
treatment       1.2345     0.3456    3.57   0.0004

gtsummary automatically converts this to odds ratios: - Treatment: OR = 3.44 (95% CI: 1.75, 6.77), p < 0.001

This immediately tells clinicians that treatment increases the odds of success by 244%.

Customization for Different Audiences

For Clinical Audiences:

Emphasize clinical significance: Include effect sizes and confidence intervals
Plain language labels: Replace variable names with descriptive text
Relevant precision: Show appropriate decimal places for clinical context

For Regulatory Submissions:

Complete statistical information: Include all required statistics
Standardized formatting: Follow regulatory guidelines for table presentation
Footnote integration: Add necessary disclaimers and explanations

For Executive Presentations:

Simplified display: Focus on key results only
Visual emphasis: Highlight significant findings
Context provision: Include baseline comparisons

Advanced Features

Statistical Customization:

Confidence interval levels: Adjust from default 95% to other levels
P-value formatting: Control decimal places and significance indicators
Effect size measures: Include standardized coefficients or effect sizes

Visual Enhancement:

Conditional formatting: Highlight significant results
Custom themes: Match organizational branding
Integration capabilities: Export to Word, HTML, or LaTeX

Multi-table Integration:

Model comparison tables: Compare multiple models side-by-side
Stratified analyses: Present results by subgroups
Combined results: Merge different types of analyses

Best Practices for Implementation

1. Know Your Audience

Statistical background: Adjust complexity accordingly
Domain expertise: Use appropriate terminology
Decision-making needs: Highlight actionable results

2. Table Design Principles

Logical organization: Group related variables together
Clear headers: Use descriptive column names
Appropriate precision: Match decimal places to measurement precision
Consistent formatting: Standardize across all tables

3. Interpretation Support

Footnotes: Explain statistical terms when necessary
Reference categories: Clearly identify comparison groups
Clinical context: Include baseline values or normal ranges

Common Implementation Challenges

Challenge 1: Variable Labeling

Problem: R variable names (e.g., trt_grp, age_yrs) aren’t presentation-ready Solution: Use descriptive labels (“Treatment Group”, “Age (years)”)

Challenge 2: Multiple Model Comparisons

Problem: Comparing results across different model types Solution: Standardize presentation format across all models

Challenge 3: Complex Interactions

Problem: Interaction terms are difficult to present clearly Solution: Consider stratified analyses or graphical presentations alongside tables

Integration with Reproducible Research

Benefits for Research Workflow:

Reproducibility: Tables update automatically when data changes
Version control: Track changes in presentation over time
Collaboration: Standardized output facilitates team communication
Quality control: Reduces manual formatting errors

Documentation Advantages:

Audit trail: Clear connection between analysis code and presentation
Transparency: Analysis decisions are explicit in code
Efficiency: Automated formatting saves time and reduces errors

Impact on Statistical Communication

For Researchers:

Increased impact: Clear presentations lead to better understanding
Time savings: Automated formatting reduces manual work
Professional appearance: Publication-ready output enhances credibility

For Decision Makers:

Better understanding: Clear tables facilitate informed decisions
Faster review: Well-organized results speed up evaluation process
Reduced miscommunication: Standardized presentation prevents misinterpretation

For the Field:

Improved standards: Elevates expectations for result presentation
Better science communication: Bridges gap between analysis and application
Enhanced reproducibility: Standardized approaches improve consistency

Future Considerations

New Trends:

Interactive tables: Integration with web-based presentation tools
Automated interpretation: AI-assisted result explanation
Personalized presentation: Audience-specific automatic formatting

Key Takeaway

The transition from raw statistical output to professional presentation isn’t just about aesthetics—it’s about effective scientific communication. Tools like gtsummary transform complex analyses into accessible insights, ensuring that statistical findings can inform decision-making at all levels.

In an era where evidence-based decision making is crucial, the ability to communicate statistical results clearly and professionally has become as important as the analysis itself. By investing in proper presentation tools and techniques, researchers can maximize the impact of their work and ensure that valuable insights reach and influence their intended audiences.

Remember: Great statistics poorly communicated are far less valuable than good statistics clearly presented. The goal is not just to analyze data, but to transform that analysis into actionable knowledge.