Enhanced Cross Tables with danchaltiel/crosstable - Professional Clinical Research • ClinicoPath

Introduction to Enhanced Cross Tables

The enhancedcrosstable function in ClinicoPath integrates the powerful danchaltiel/crosstable package to provide advanced cross-tabulation capabilities beyond standard contingency tables. This module offers sophisticated data analysis features specifically designed for clinical research, including automated statistical test selection, comprehensive effect size calculations, and publication-ready formatting.

What Makes Enhanced Cross Tables Special?

Advanced Functionality

Tidyselect Syntax: Modern R syntax for variable selection and manipulation
Formula Interface: Transform variables on-the-fly during analysis
Automated Testing: Intelligent statistical test selection based on data characteristics
Effect Sizes: Comprehensive effect size calculations with confidence intervals
Professional Output: Publication-ready formatting for medical journals

Clinical Research Focus

Medical Terminology: Clinically appropriate variable labels and descriptions
Regulatory Compliance: Output formats suitable for regulatory submissions
Flexible Reporting: Multiple export options including Word, HTML, and CSV
Statistical Rigor: Appropriate tests for medical research standards

When to Use Enhanced Cross Tables

Perfect for: - Clinical trial baseline characteristic tables (Table 1) - Case-control studies examining risk factor associations - Diagnostic test evaluation and validation studies - Epidemiological research with categorical exposures and outcomes - Multi-center studies requiring standardized reporting - Regulatory submissions requiring detailed statistical analysis

Key Advantages: - Superior functionality compared to basic cross-tabulation - Automated statistical test selection reduces analysis errors - Professional formatting saves time in manuscript preparation - Comprehensive documentation supports reproducible research

Getting Started

Installation Requirements

The enhanced cross-table functionality requires the crosstable package:

# Install from CRAN
install.packages("crosstable")

# For enhanced export features
install.packages("officer")
install.packages("flextable")

# Supporting packages (automatically handled)
install.packages(c("janitor", "labelled", "stringr", "dplyr"))

Basic Syntax Comparison

Standard crosstable vs Enhanced crosstable

# Standard ClinicoPath crosstable
crosstable(
  data = histopathology,
  row_var = "Grade_Level",
  col_var = "Group",
  add_margins = TRUE
)

# Enhanced crosstable with danchaltiel integration
enhancedcrosstable(
  data = histopathology,
  vars = c("Grade_Level", "LVI", "PNI"),
  by_var = "Group",
  test_auto = TRUE,
  effect_size = TRUE,
  export_format = "html"
)

Key Differences

Multiple Variables: Enhanced version analyzes multiple variables simultaneously
Automated Testing: Statistical tests selected automatically based on data
Effect Sizes: Comprehensive effect size calculations included
Advanced Formatting: Professional output with customizable patterns

Clinical Research Applications

Example 1: Clinical Trial Baseline Characteristics

This example demonstrates creating a comprehensive Table 1 for a clinical trial using the histopathology dataset:

library(ClinicoPath)

# Load the histopathology dataset
data("histopathology")

# Basic enhanced cross-table for baseline characteristics
enhancedcrosstable(
  data = histopathology,
  vars = c("Age", "Sex", "Race", "Grade_Level", "LVI", "PNI"),
  by_var = "Group",
  percent_pattern = "col_percent",
  test_auto = TRUE,
  effect_size = FALSE,
  funs_arg = "mean_sd",
  show_total = TRUE,
  show_total_row = TRUE,
  use_labels = TRUE,
  export_format = "html"
)

Clinical Interpretation: - Automatically selects appropriate tests (t-test for Age, chi-square for categorical variables) - Column percentages show distribution within each treatment group - Professional formatting suitable for publication - Variable labels enhance readability for clinical audiences

Example 2: Risk Factor Analysis with Effect Sizes

Examining associations between pathological features and outcomes:

# Risk factor analysis with comprehensive effect sizes
enhancedcrosstable(
  data = histopathology,
  vars = c("LVI", "PNI", "LymphNodeMetastasis", "Grade_Level"),
  by_var = "Mortality5yr",
  percent_pattern = "row_percent",
  test_auto = TRUE,
  effect_size = TRUE,
  margin = "row",
  show_interpretation = TRUE,
  export_format = "flextable"
)

Advanced Features: - Row percentages show risk within each exposure category - Effect sizes (Cramer’s V, odds ratios) quantify association strength - Statistical interpretation provides clinical context - Flextable output ready for Word documents

Example 3: Diagnostic Test Evaluation

Analyzing diagnostic test performance across different populations:

# Diagnostic test evaluation
enhancedcrosstable(
  data = histopathology,
  vars = c("New Test", "Golden Standart"),
  by_var = "Disease Status",
  percent_pattern = "col_percent",
  test_auto = TRUE,
  effect_size = TRUE,
  cor_method = "spearman",
  showNA = "ifany",
  compact = FALSE,
  summary_statistics = TRUE
)

Key Features: - Evaluates test performance across disease status - Spearman correlation for ordinal diagnostic scores - Missing data handling with transparency - Summary statistics provide additional context

Advanced Features and Customization

Percentage Pattern Options

The enhanced cross-table offers sophisticated percentage display patterns:

Pattern Types

col_percent: {n} ({p_col}%) - Percentages within columns
row_percent: {n} ({p_row}%) - Percentages within rows
total_percent: {n} ({p_tot}%) - Percentages of total
count_only: {n} only - Raw counts without percentages
percent_only: {p_col} only - Percentages without counts

Clinical Applications

# For prevalence studies (column percentages)
enhancedcrosstable(
  data = histopathology,
  vars = "LVI",
  by_var = "Sex",
  percent_pattern = "col_percent"  # Shows prevalence within each sex
)

# For risk analysis (row percentages)
enhancedcrosstable(
  data = histopathology,
  vars = "Grade_Level",
  by_var = "Mortality5yr", 
  percent_pattern = "row_percent"  # Shows mortality risk within each grade
)

Statistical Test Automation

The enhanced cross-table automatically selects appropriate statistical tests:

Test Selection Logic

Chi-square test: For expected frequencies ≥ 5
Fisher’s exact test: For small expected frequencies (< 5)
t-test: For continuous variables with 2 groups
ANOVA: For continuous variables with >2 groups
Correlation tests: For continuous-continuous associations

Custom Test Selection

# Force specific statistical approaches
enhancedcrosstable(
  data = histopathology,
  vars = c("Age", "MeasurementA"),
  by_var = "Group",
  cor_method = "spearman",  # Non-parametric correlations
  funs_arg = "median_q1q3"  # Robust summary statistics
)

Effect Size Calculations

Comprehensive effect size analysis provides clinical significance context:

Available Effect Sizes

Cramer’s V: For categorical associations (small: 0.1, medium: 0.3, large: 0.5)
Odds Ratios: For 2×2 tables with confidence intervals
Phi Coefficient: For binary associations
Cohen’s d: For continuous variable comparisons

Clinical Interpretation

# Comprehensive effect size analysis
enhancedcrosstable(
  data = histopathology,
  vars = c("LVI", "PNI", "Grade_Level"),
  by_var = "Outcome",
  effect_size = TRUE,
  show_interpretation = TRUE
)

Interpretation Guidelines: - Effect sizes quantify practical significance beyond p-values - Confidence intervals provide precision estimates - Clinical guidelines help interpret magnitude of effects - Essential for evidence-based medicine applications

Export and Reporting Options

Publication-Ready Outputs

HTML Tables

# Professional HTML output for web reports
enhancedcrosstable(
  data = histopathology,
  vars = c("Sex", "Age", "Grade_Level"),
  by_var = "Group",
  export_format = "html",
  compact = FALSE,
  use_labels = TRUE
)

Word Document Integration

# Flextable output for Word manuscripts
enhancedcrosstable(
  data = histopathology,
  vars = c("LVI", "PNI", "LymphNodeMetastasis"),
  by_var = "Mortality5yr",
  export_format = "flextable",
  digits = 2,
  show_n_col = TRUE
)

CSV Data Export

# CSV export for further analysis
enhancedcrosstable(
  data = histopathology,
  vars = c("Grade_Level", "TStage"),
  by_var = "Group",
  export_format = "csv",
  exclude_missing = TRUE
)

Customization Options

Display Formatting

# Customized display options
enhancedcrosstable(
  data = histopathology,
  vars = c("Age", "OverallTime"),
  by_var = "Mortality5yr",
  digits = 3,                    # Decimal precision
  compact = TRUE,                # Space-efficient layout
  show_total = TRUE,             # Include totals
  show_total_row = TRUE,         # Include row totals
  show_n_col = TRUE              # Display column sample sizes
)

Missing Data Handling

# Comprehensive missing data options
enhancedcrosstable(
  data = histopathology,
  vars = c("Race", "Smoker"),
  by_var = "Group",
  exclude_missing = FALSE,       # Include missing in analysis
  showNA = "ifany",              # Show missing if present
  use_labels = TRUE              # Use descriptive labels
)

Best Practices for Clinical Research

Design Recommendations

Variable Selection

Primary Variables: Focus on clinically meaningful associations
Confounding Variables: Include known confounders for stratified analysis
Sample Size: Ensure adequate power for subgroup analyses

Statistical Considerations

Multiple Testing: Consider Bonferroni or FDR correction for multiple comparisons
Effect Sizes: Always report alongside p-values for clinical significance
Confidence Intervals: Provide precision estimates for key associations

Reporting Standards

CONSORT Guidelines: Follow trial reporting standards
STROBE Statement: Adhere to observational study guidelines
Journal Requirements: Check specific journal formatting requirements

Quality Assurance

Data Validation

# Pre-analysis data quality check
summary(histopathology[c("Group", "Mortality5yr", "LVI")])

# Enhanced cross-table with validation
enhancedcrosstable(
  data = histopathology,
  vars = c("LVI", "PNI"),
  by_var = "Group",
  exclude_missing = TRUE,        # Remove incomplete cases
  show_interpretation = TRUE,    # Include statistical guidance
  summary_statistics = TRUE      # Display sample size information
)

Reproducibility Standards

# Document analysis parameters for reproducibility
analysis_params <- list(
  dataset = "histopathology v1.0",
  variables = c("LVI", "PNI", "Grade_Level"),
  grouping = "Group",
  test_method = "automatic",
  missing_data = "excluded",
  effect_sizes = TRUE,
  alpha_level = 0.05
)

# Include in analysis output
enhancedcrosstable(
  data = histopathology,
  vars = analysis_params$variables,
  by_var = analysis_params$grouping,
  exclude_missing = TRUE,
  test_auto = TRUE,
  effect_size = TRUE
)

Comparison with Standard Cross Tables

Functionality Comparison

Feature	Standard crosstable	Enhanced crosstable
Variable Selection	Single pair	Multiple variables
Statistical Tests	Manual selection	Automated selection
Effect Sizes	Manual calculation	Automated calculation
Export Options	Basic HTML	HTML, Word, CSV
Missing Data	Basic handling	Comprehensive options
Tidyselect Syntax	Not supported	Full support
Formula Interface	Limited	Advanced transformations

When to Use Each Approach

Standard crosstable

Quick exploratory analysis
Simple 2×2 or small contingency tables
Basic statistical requirements
Limited formatting needs

Enhanced crosstable

Comprehensive clinical research
Multiple variable analysis
Publication-ready output
Advanced statistical features
Regulatory submission requirements

Troubleshooting and Support

Common Issues

Package Dependencies

# Check required packages
required_packages <- c("crosstable", "officer", "janitor", "labelled")

for (pkg in required_packages) {
  if (!requireNamespace(pkg, quietly = TRUE)) {
    install.packages(pkg)
  }
}

Data Formatting Issues

# Ensure proper factor levels
histopathology$Group <- factor(histopathology$Group)
histopathology$Mortality5yr <- factor(histopathology$Mortality5yr)

# Check for adequate sample sizes
table(histopathology$Group, histopathology$Mortality5yr)

Export Problems

# Fallback to HTML if flextable fails
enhancedcrosstable(
  data = histopathology,
  vars = "LVI",
  by_var = "Group",
  export_format = "html"  # Reliable fallback option
)

Getting Help

Documentation Resources

Function Help: ?enhancedcrosstable for parameter details
Package Documentation: help(package = "crosstable") for advanced features
Clinical Examples: Additional vignettes for specific research contexts

Community Support

GitHub Issues: Report bugs and request features
Clinical Research Forums: Discuss analysis approaches with peers
Statistical Consulting: Seek professional guidance for complex studies

Advanced Topics

Integration with Clinical Workflows

EMR Data Integration

# Example with electronic medical record data
emr_data <- data.frame(
  patient_id = 1:200,
  age_group = factor(sample(c("Young", "Middle", "Old"), 200, replace = TRUE)),
  comorbidity = factor(sample(c("Yes", "No"), 200, replace = TRUE)),
  treatment = factor(sample(c("Standard", "Experimental"), 200, replace = TRUE)),
  outcome = factor(sample(c("Success", "Failure"), 200, replace = TRUE))
)

# Clinical research analysis
enhancedcrosstable(
  data = emr_data,
  vars = c("age_group", "comorbidity"),
  by_var = "outcome",
  test_auto = TRUE,
  effect_size = TRUE,
  show_interpretation = TRUE
)

Multi-Center Studies

# Stratified analysis by site
multi_center_analysis <- function(data, site_var) {
  sites <- unique(data[[site_var]])
  
  for (site in sites) {
    cat("\n=== Analysis for", site, "===\n")
    site_data <- data[data[[site_var]] == site, ]
    
    result <- enhancedcrosstable(
      data = site_data,
      vars = c("treatment", "severity"),
      by_var = "outcome",
      test_auto = TRUE,
      compact = TRUE
    )
    
    print(result)
  }
}

Future Enhancements

Planned Features

Tidyverse Integration: Full dplyr and tidyselect syntax support
Advanced Visualizations: Integrated plots for categorical associations
Regulatory Templates: Pre-configured formats for FDA, EMA submissions
Real-time Analysis: Streaming data support for clinical trials

Research Applications

Precision Medicine: Biomarker-treatment interaction analysis
Digital Health: Mobile health data cross-tabulation
Population Health: Large-scale epidemiological studies
Health Economics: Cost-effectiveness analysis support

Conclusion

The enhanced cross-table functionality represents a significant advancement in clinical research analysis capabilities. By integrating the sophisticated danchaltiel/crosstable package with ClinicoPath’s clinical focus, researchers gain access to:

Professional Quality: Publication-ready outputs that meet journal standards
Statistical Rigor: Automated test selection and comprehensive effect sizes
Clinical Relevance: Features designed specifically for medical research
Efficiency: Streamlined workflow from analysis to manuscript
Reproducibility: Comprehensive documentation and parameter tracking

Key Benefits

Enhanced Productivity: Automated statistical analysis reduces manual work
Improved Quality: Professional formatting and statistical rigor
Clinical Focus: Designed for medical research applications
Flexibility: Multiple output formats and customization options
Integration: Seamless workflow with ClinicoPath ecosystem

Recommendations

Start Simple: Begin with basic cross-tabulations before exploring advanced features
Validate Results: Compare with standard methods during initial implementation
Document Parameters: Maintain detailed analysis logs for reproducibility
Seek Training: Consider statistical consulting for complex research designs
Stay Updated: Monitor package updates for new features and improvements

For optimal results in clinical research, combine the enhanced cross-table functionality with appropriate study design, adequate sample sizes, and careful interpretation of statistical results in clinical context.

Enhanced Cross Tables with danchaltiel/crosstable - Professional Clinical Research

Advanced cross-tabulation analysis with tidyselect syntax and automated reporting

ClinicoPath

2025-07-13