Excess Mortality and Health System Collapse in Protracted Conflict: A Population-Based Cohort Study from Yemen (2015–2025)

This repository contains all supplementary materials, analytical code, and de-identified data for the manuscript entitled "Excess Mortality and Health System Collapse in Protracted Conflict: Evidence from a Population-Based Cohort in Yemen", under consideration in Conflict and Health. Study Overview: We conducted a population-based retrospective cohort study of 6,000 individuals from 1,152 households in Al-Dhalea and Lahj governorates, southern Yemen, over a decade of conflict (2015–2025). A multidimensional cumulative war exposure index was constructed to capture direct violence, displacement, livelihood loss, healthcare access interruption, and frontline residence. The primary outcome was all-cause mortality, analysed using multivariable Cox proportional hazards models, competing risk regression, and population attributable fractions (PAF). Key Findings: - Clear exposure–response gradient: mortality 2.8% (low), 6.1% (medium), 8.8% (high exposure) - High exposure associated with 2.64-fold increased mortality risk (aHR 2.64, 95% CI 2.01–3.46) - 43.7% (95% CI 35.2–51.1) of deaths attributable to conflict-related exposures under a causal interpretation - 65.9% of deaths occurred outside health facilities; 46.6% received no medical care before death; 35.0% died from treatable conditions - Health system collapse, not direct violence, was the dominant pathway of excess mortality Repository Contents: File S1: Core Study Questionnaire Modules File S2: Informed Consent Form (English translation) File S3: Methodological and Reporting Checklists (STROBE, Ethics, FAIR, Peer Review) File S4: Data Dictionary and Variable Specifications File S5: De-identified dataset (controlled access) File S6: Supplementary Methods Appendix (including DAG and sensitivity analysis plans) File S7: Statistical Code Reproducibility Appendix (complete R scripts) Spatial_Visualization_Code.R: Code for geographic visualization (Figure S2) Supplementary_Tables.docx: Tables S1–S16 (including S7b and S7c for sensitivity analyses) Supplementary_Figures.docx: Figures S1–S4 Data Availability: The de-identified individual-level dataset is available under controlled access to protect participant confidentiality in an active conflict setting. Access requests can be submitted via this OSF repository and are subject to a data use agreement. All analytical code is fully reproducible and openly available. Ethical Approval: Radfan College University Institutional Review Board, University of Lahej, Yemen (YEM-IRB-2024-016; March 15, 2024), with official endorsement from the Office of the Ministry of Public Health and Population – Al-Dhalea Governorate. Funding: This research received no specific grant from any funding agency. Correspondence: Naif Taleb Ali (n.taleb@ust.edu) | ORCID: 0000-0002-2266-1569 Khalil A. Saleh (dr.khalil.aden@gmail.com) | ORCID: 0000-0002-6965-9968 Repository DOI: https://doi.org/10.17605/OSF.IO/WJHYU