Meta Analysis in R and Ai

Module 1: Introduction to Meta-Analysis

Day 1 – Foundations of Meta-Analysis

• Concept and purpose of meta-analysis

• Effect sizes and study heterogeneity

• Fixed vs. random effects models

• Key steps in conducting meta-analysis

• Overview of R packages (meta, metafor, dmetar)

Day 2 – Data Preparation

• Structure of meta-analysis data (effect size, SE, CI, sample size)

• Importing and cleaning data in R (readxl, dplyr)

• Handling missing data and transformation of variables

• Practice: create a structured dataset from published studies

Module 2: Effect Size Computation

Day 3 – Continuous Outcomes

• Mean difference (MD), Standardized mean difference (SMD)

• Calculating effect sizes using escalc()

• Interpreting Cohen’s d, Hedges g

Day 4 – Binary Outcomes

• Odds ratio, risk ratio, risk difference

• Computing log OR and SE

• Example: meta-analysis of treatment success vs. control

Module 3: Model Fitting and Interpretation

Day 5 – Fixed-Effect Model

• Theory and formula

• Performing fixed-effect meta-analysis in R (metagen())

• Forest plot interpretation

• Hands-on practice

Day 6 – Random-Effects Model

• Understanding heterogeneity (τ², I², Q-test)

• Fitting random-effects model (rma(), metagen())

• Comparing fixed vs random models

Day 7 – Forest Plots & Visualizations

• Customizing forest plots (meta, metafor, ggplot2)

• Adding subgroup, study labels, effect sizes, and weights

• Exporting publication-quality plots (PNG/PDF 600 dpi)

Module 4: Heterogeneity & Bias Assessment

Day 8 – Heterogeneity Diagnostics

• Cochran’s Q, I², τ² statistics

• Subgroup analyses to explore heterogeneity

• Forest plot by subgroup

Day 9 – Publication Bias

• Funnel plot interpretation

• Egger’s regression test, Begg’s test

• Duval & Tweedie’s trim-and-fill method

• Asymmetry visualization in R

Module 5: Advanced Methods

Day 10 – Meta-Regression

• Concept and interpretation

• Model building using moderators (rma() with moderators)

• Example: Effect size ~ sample size + region

Day 11 – Subgroup Analysis & Sensitivity

• Grouping studies by categorical variables (e.g., gender, region)

• Leave-one-out analysis

• Influence diagnostics and Baujat plot

Day 12 – Network and Multivariate Meta-Analysis

• Concept of network meta-analysis

• Introduction to multivariate models in metafor

• Overview of netmeta and mvmeta packages

Module 6: Reporting & Automation

Day 13 – Reporting Results

• PRISMA workflow

• Summary tables and model output interpretation

• Writing a results section (APA/Journal style)

• Export to Word/Excel using officer or openxlsx