Boundary Conditions of Ω in Healthcare Systems: A Comparative Evaluation Against Continuous Baselines
Fuente:
Zenodo
Enregistré dans:
| Auteur principal: | |
|---|---|
| Format: | Recurso digital |
| Publié: |
Zenodo
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866901349320359936 |
|---|---|
| author | Aizawa, Hiroaki |
| author_facet | Aizawa, Hiroaki |
| contents | <p>This repository contains a structured evaluation of the directional metric Ω in the context of healthcare system collapse detection.</p> <p>Using publicly available Scottish healthcare operational data (A&E weekly activity, Delayed Discharge monthly statistics, and Scottish Ambulance Service weekly operational data), we compare a modified Ω-based directional persistence model against a continuous baseline logistic regression model (M1) under identical preprocessing, training, and evaluation conditions.</p> <p>The study investigates whether Ω is indispensable for collapse detection.</p> <p>Three collapse definitions are evaluated:</p> <p>Break-A (Surface Congestion): A&E 12-hour exceedance rate above rolling 24-month 90th percentile.</p> <p>Break-B (Structural Exit Blockage): Delayed Bed Days above 90th percentile AND Delayed Census above 75th percentile.</p> <p>Break-C (Composite Structural Shock): At least two of Delayed Beds, Delayed Census, or SAS Turnaround exceed 90th percentile.</p> <p>Results show that Ω does not demonstrate universal dominance. Under surface congestion (Break-A), the continuous baseline model outperforms Ω in AUC and operational loss. However, under structural persistence-based collapse definitions (Break-B and Break-C), Ω shows comparative advantage, particularly in low-false-negative operating regions and cost-weighted evaluation.</p> <p>The findings suggest that Ω is not a general congestion detector but may provide domain-conditional advantage in detecting sustained structural deterioration characterized by directional persistence.</p> <p>This repository includes:</p> <p>A technical evaluation report</p> <p>A complete version with figures and fixed numerical results</p> <p>An appendix detailing model equations, variable definitions, cost function specification, and evaluation protocol</p> <p>The objective of this release is to document the boundary conditions of Ω rather than to claim universal superiority.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_18814018 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Boundary Conditions of Ω in Healthcare Systems: A Comparative Evaluation Against Continuous Baselines Aizawa, Hiroaki Omega metric directional persistence healthcare systems collapse detection logistic regression Pareto frontier structural deterioration operational risk <p>This repository contains a structured evaluation of the directional metric Ω in the context of healthcare system collapse detection.</p> <p>Using publicly available Scottish healthcare operational data (A&E weekly activity, Delayed Discharge monthly statistics, and Scottish Ambulance Service weekly operational data), we compare a modified Ω-based directional persistence model against a continuous baseline logistic regression model (M1) under identical preprocessing, training, and evaluation conditions.</p> <p>The study investigates whether Ω is indispensable for collapse detection.</p> <p>Three collapse definitions are evaluated:</p> <p>Break-A (Surface Congestion): A&E 12-hour exceedance rate above rolling 24-month 90th percentile.</p> <p>Break-B (Structural Exit Blockage): Delayed Bed Days above 90th percentile AND Delayed Census above 75th percentile.</p> <p>Break-C (Composite Structural Shock): At least two of Delayed Beds, Delayed Census, or SAS Turnaround exceed 90th percentile.</p> <p>Results show that Ω does not demonstrate universal dominance. Under surface congestion (Break-A), the continuous baseline model outperforms Ω in AUC and operational loss. However, under structural persistence-based collapse definitions (Break-B and Break-C), Ω shows comparative advantage, particularly in low-false-negative operating regions and cost-weighted evaluation.</p> <p>The findings suggest that Ω is not a general congestion detector but may provide domain-conditional advantage in detecting sustained structural deterioration characterized by directional persistence.</p> <p>This repository includes:</p> <p>A technical evaluation report</p> <p>A complete version with figures and fixed numerical results</p> <p>An appendix detailing model equations, variable definitions, cost function specification, and evaluation protocol</p> <p>The objective of this release is to document the boundary conditions of Ω rather than to claim universal superiority.</p> |
| title | Boundary Conditions of Ω in Healthcare Systems: A Comparative Evaluation Against Continuous Baselines |
| topic | Omega metric directional persistence healthcare systems collapse detection logistic regression Pareto frontier structural deterioration operational risk |
| url | https://doi.org/10.5281/zenodo.18814018 |