Congestion-Responsive Management of Reserve Treatment Capacity in Emergency Care with Evolving Patient Acuity

Description

This Version 2 release is synchronized with manuscript version V12.7 of “Congestion-Responsive Management of Reserve Treatment Capacity in Emergency Care with Evolving Patient Acuity.” It contains the sparse continuous-time Markov chain implementation, an independently implemented discrete-event simulation, synthetic configurations, threshold-policy grids, source data for Tables 6–12, arrival-load and switching-action-cost reoptimization results, baseline and non-diagonal MMAP validation, simulation seeds, environment specifications, automated tests, verification records, and the seven high-resolution manuscript figures. Version 2 corrects the narrative cost-component decomposition and the reference mapping in Supplementary Table A.1. The mathematical model, computational inputs, numerical output files, policy selections, and figures remain unchanged. The legacy K = 2 policy is represented as the mathematically equivalent adjacent threshold pair (1, 2) under a common action-counting convention. All inputs are synthetic. No patient-level, hospital-identifying, or personally identifiable data are included. Code is provided under the MIT License; data, documentation, configurations, and generated outputs are provided under the CC BY 4.0 License.

Steps to reproduce

Step 1 — Download and extract HCMS_V12_7_Reproducibility_Supplementary_V2.zip. Step 2 — Open a terminal and move to HCMS_Reproducibility_Data_and_Code_V2/repository. Step 3 — Use Python 3.12; the certified environment used Python 3.12.13. Install the required packages with “python -m pip install -r environment/requirements.txt”. Step 4 — Set the source directory on the Python path. On Linux, macOS, WSL, or Git Bash, run “export PYTHONPATH=src”. On Windows PowerShell, run “$env:PYTHONPATH = 'src'”. Step 5 — Run “python -m pytest -q” and then “python scripts/verify_release.py” as two separate commands. The certified V12.7 release passes all 11 automated tests and all 32 release-verification checks. Step 6 — Reproduce the complete computational workflow with “bash scripts/run_all.sh”. Windows users may run this command through WSL or Git Bash. Step 7 — To rerun only the arrival-load and switching-action-cost policy screen and regenerate the corresponding policy-map figure, run “python scripts/run_load_switching_cost_reoptimization.py” and then “python scripts/generate_final_policy_map_figure.py” as separate commands. The exact manuscript figures are preserved in the publication_figures directory. Newly rendered figures should be numerically equivalent, although byte-level identity may vary across operating systems and graphics-library versions. Archive integrity can be checked against MANIFEST_SHA256.txt. All inputs are synthetic, and no patient-level or hospital-identifying data are required.

Institutions

• Jeonbuk National University

  Jeollabuk-do, Jeonju
• Sangji University

  Gangwon-do, Wŏnju

Categories

Licence