Meta-Analysis Project: ART 736-Canet

Overview

This R Markdown document contains a comprehensive meta-analysis workflow for processing and analyzing effect sizes from multiple studies. The project follows a systematic approach through several stages, from data preprocessing to final analysis and visualization.

Videos de Soporte

https://media.upv.es/#/portal/channel/5ab9fa10-22b4-11ef-8f10-ab3091b330f3

pending

• Converger a un modelo junto con art-639
• Documetnos excel de ejemplo

Project Structure

Stage 0: Data Description

• Purpose: Initial exploration and description of the input Excel file
• Input: BRANHAM_V2.xlsx (Excel file with multiple sheets)
• Output: Basic data structure information (column names, data types, sheet structure)

Stage 1: Effect Size Collapsing

• Purpose: Aggregate multiple effect sizes from the same study using Borenstein et al. (2009) formula 24.5
• Key Features:
  • Handles studies with multiple outcomes by calculating weighted averages
  • Assumes correlation of 1.0 between outcomes (conservative approach)
  • Processes data sheet by sheet, identifying study groups separated by blank rows
  • Combines multiple effect sizes into single aggregated values
• Output: Collapsed effect sizes saved as Excel file with timestamp

Stage 2: Individual Meta-Analyses

• Purpose: Perform separate meta-analyses for each analysis group
• Method: Random effects model using Hedges' g
• Calculations Include:
  • Combined effect size and confidence intervals
  • Prediction intervals
  • Credibility intervals
  • Heterogeneity statistics (Q, I², τ²)
  • Z-tests for significance
• Output: Comprehensive results table with all meta-analytic statistics

Stage 3: Visualization

• Forest Plots: Visual representation of individual study effects and overall estimates
• Funnel Plots: Assessment of publication bias using trim-and-fill method
• Features:
  • Automatic detection of missing studies
  • Visual distinction between original and imputed studies
  • Export capabilities to PDF format
  • Text summaries of trim-and-fill results

Stage 4: Between-Group Comparisons

• Purpose: Statistical comparison between different analysis groups
• Method: Mixed-effects meta-regression
• Statistics Calculated:
  • Q-between: Heterogeneity between groups
  • Q-within: Heterogeneity within groups
  • Associated p-values and degrees of freedom
• Output: Comprehensive comparison table for all group pairs

Key Dependencies

library(readxl)      # Excel file reading
library(metafor)     # Meta-analysis calculations
library(writexl)     # Excel file writing
library(dplyr)       # Data manipulation

Input Data Requirements

• Excel file with multiple sheets containing study data
• Required columns:
  • ID: Study identifier
  • ID único de muestra: Unique sample identifier
  • Hedges' g (effect size): Effect size values
  • g variance (std error ^2): Effect size variances
  • Authors: Author information
  • Año: Publication year
  • Outcomes: Outcome measures

Output Files

1. Collapsed Effect Sizes: ART-736canetMeta-EScolapsados_[timestamp].xlsx
2. Meta-Analysis Results: ART-736canetMeta-analisis-[timestamp].xlsx
3. Group Comparisons: ART-736canetMeta-analisis-grupos-[timestamp].xlsx
4. Visualizations: PDF files for forest plots and funnel plots
5. Trim-and-Fill Results: Text files with publication bias analysis

Methodological Notes

Effect Size Aggregation

• Uses Borenstein et al. (2009) formula 24.5 for combining correlated effect sizes
• Conservative approach assumes perfect correlation (r = 1.0) between outcomes
• Increases variance estimates to account for dependency

Publication Bias Assessment

• Trim-and-fill method identifies potentially missing studies
• Visual inspection through funnel plots
• Quantitative assessment of bias impact on overall estimates

Heterogeneity Analysis

• I² statistics for quantifying heterogeneity
• τ² estimates for between-study variance
• Q-statistics for testing homogeneity assumptions

Usage Instructions

1. Ensure input Excel file is in the working directory
2. Update the archivo variable with your file name
3. Run stages sequentially (0 → 1 → 2 → 3 → 4)
4. Check output files for results and visualizations

References

• Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2009). Introduction to meta-analysis. John Wiley & Sons.
• Meta-Essentials FAQ: https://www.erim.eur.nl/research-support/meta-essentials/frequently-asked-questions/

Notes

• The script includes extensive error handling and progress reporting
• Visualization functions support both screen display and PDF export
• All calculations follow established meta-analytic best practices
• Code is modular and can be adapted for different datasets