Three essays on data-driven models in health care operations management
Notice bibliographique
Résumé
Though 20th century has seen life expectancy largely lengthened worldwide, aging population, chronic diseases, worsening food supply with deficit nutrition and environmental problems add to the burden of healthcare systems all around the world. Data analytics, which has been seen as a significant power in other industries, is expected to contribute to the improvement of efficiency and effectiveness in healthcare. This thesis aims to identify and promote more effective and efficient strategic, operations and clinical policies in healthcare systems through descriptive, predictive and prescriptive analytics. To this end, this thesis focuses on three essays, i.e. three data-driven problems based on medium to large size of real life datasets, on: i) design of financial incentive systems for maternity care; ii) design of specialist response policies and modified triage coding to reduce waiting times in emergency departments (EDs), and iii) design of observation units for hearth failure patients. The first essay focuses on strategic level and aims to design a two-level financial incentive mechanisms to reimburse physicians, in order to reduce unnecessary C-sections while retain it for those who need it, resulting in enhanced birth quality with alleviated economic burden for overall health care system. Contributing to clinical decision-making, we first cluster the patients according to their pregnancy complexities, and characterize a threshold between spontaneous birth and medically necessary planned C-section by analyzing 12.7 million annual birth records from National Bureau of Economics Research through statistical learning methods. Then we compare payment systems analytically vis-\'{a}-vis a variety of performance measures within two-level hierarchy, (i) mainstream payment models and (ii) compensation on the top of mainstream payment, and provide insights about the effectiveness of alternative payment models in the context of maternity care. Finally, we propose optimal payment for physicians to maximize the value for patients under the principal and agent framework, from the strategic perspective.The second paper focuses on operational level and targets to reduce the length of stay in EDs by designing a systematic response policy for various specialists depending on ED clinical demands. This work is motivated by and verified with 40,000 ED visits to a local community hospital in Montreal. We first identify a class of patients who are more likely to require specialist consultation based on their clinical information available at the triage stage through statistical analysis. Then we analyze several alternative policies for specialists' response to consultation requests using queuing models with non-homogeneous Poisson arrival rates. Moreover, we examine an integrated ED decision-making by incorporating specialist consultation requests in the triage system. Finally, our proposed optimal specialist response policy and associated modified triage coding are verified through a comprehensive simulation model. We provide a feasible guideline of integrated patient streamlining to shorten length of stay and alleviate overcrowding in ED. The third paper focuses on clinical level and propose a framework to design a dedicated observation unit for acute decomposition heart failure patients, in order to provide proper treatment and reduce unnecessary hospitalization and chance of post-discharge events. To this end, we, first, use multiple analytical models to figure out the proper number of bed for this observation unit based on historical patient arrival data from a local community hospital. Based on the confined range of analytical capacity, we use simulation models to analyze different discharge and admission policies. We propose an optimal discharge-admission criteria for this dedicated observation unit to realize cost-saving and quality enhancement of treating acute decomposition heart failure patients.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,012 | 0,034 |
| Méta-épidémiologie (sens strict) | 0,002 | 0,001 |
| Méta-épidémiologie (sens large) | 0,001 | 0,002 |
| Bibliométrie | 0,002 | 0,003 |
| Études des sciences et des technologies | 0,001 | 0,006 |
| Communication savante | 0,006 | 0,008 |
| Science ouverte | 0,002 | 0,003 |
| Intégrité de la recherche | 0,004 | 0,007 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,006 | 0,001 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».