Privacy-Preserving ICU Coordination: Simulation Code and Synthetic Data

This repository contains the simulation code and synthetic data used in the study “Privacy-Preserving ICU Coordination under Homomorphic Encryption.” The package implements a regional ICU allocation framework in which hospitals compute homomorphically encrypted deterioration risk logits locally, while a regional coordinator performs encrypted, order-preserving allocation without access to any patient-level plaintext information. Contents include: • Fully reproducible Python simulation code for ICU allocation policies • Synthetic and aggregated data derived from MIMIC-IV v2.2 • Configuration files for surge, capacity, and noise sensitivity experiments • Scripts for generating all figures and tables reported in the manuscript No identifiable patient data are included. All experiments rely exclusively on synthetic or aggregated representations that comply with HIPAA and GDPR standards. This repository supports transparency, reproducibility, and independent verification of privacy-preserving healthcare operations research.