yamathir/bayesian-fishermen-respiratory

bayesian-statistics occupational-health epidemiology respiratory-disease copd chronic-bronchitis stan r mcmc public-health morocco diagnostic-bias # Bayesian Methods for Correcting Diagnostic Bias in Mediterranean Fishermen Respiratory Disease Studies ## Overview This repository contains code and documentation for a Bayesian analysis correcting diagnostic bias in occupational respiratory disease prevalence among Mediterranean fishermen. The study addresses the challenge of comparing respiratory disease prevalence across 25 years (1995-2020) of evolving diagnostic criteria, from pre-GOLD clinical evaluation to strict post-GOLD spirometry. **Key Finding:** Morocco's observed 9.7% prevalence (2018) corrects to 20.3-23.9% when accounting for diagnostic evolution and selection biases, converging with Spain's 1995 baseline and revealing substantial undetected disease burden. ## Repository Contents ``` bayesian-fishermen-respiratory/ │ ├── README.md # This file ├── LICENSE # MIT License ├── CITATION.cff # Machine-readable citation │ ├── code/ │ ├── full_analysis.qmd # Complete R/Stan analysis (Quarto) │ └── full_analysis.html # Rendered HTML output │ ├── data/ │ └── README_data.md # Data sources and citations │ ├── docs/ │ ├── Technical_Supplement_EN.pdf # Prior justifications │ └── Validation_Document_EN.pdf # 52 validation tests │ └── manuscript/ ├── IMH-Article-BAYESIAN-FISHERMEN.qmd # Article source (Quarto) ├── IMH-Article-BAYESIAN-FISHERMEN.pdf # Article PDF └── IMH-Article-BAYESIAN-FISHERMEN.docx # Article Word format ``` ## Key Methods Four independent Bayesian methods converged on corrected prevalence: 1. **Method 1 (Monte Carlo):** 21.5% [17.8-26.2%] 2. **Method 2 (Dual Bias Separation):** 23.9% [21.2-26.8%] 3. **Method 3 (Trajectory Predictions):** 22.7% [16.5-24.8%] 4. **Method 4 (Three-Level Hierarchical):** 20.3% [12.3-32.0%] **Average:** 22.1% (range: 20.3-23.9%) ## Data Sources This analysis uses data from four Mediterranean countries: - **Spain 1995 …