東華大學圖書館 |

An Analytical Framework for Mental Healthcare Operations Management /

Record Type:	Electronic resources : Monograph/item
Title/Author:	An Analytical Framework for Mental Healthcare Operations Management // Hossein Hejazian.
Author:	Hejazian, Hossein,
Description:	1 electronic resource (240 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 86-04, Section: B.
Contained By:	Dissertations Abstracts International86-04B.
Subject:	Patients. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31527265
ISBN:	9798342116695

An Analytical Framework for Mental Healthcare Operations Management /
Hejazian, Hossein,

An Analytical Framework for Mental Healthcare Operations Management /Hossein Hejazian. - 1 electronic resource (240 pages)

Source: Dissertations Abstracts International, Volume: 86-04, Section: B.

Healthcare operations management is dedicated to enhancing the processes within health systems to optimize the quality of care and improve patient outcomes. This field's challenges take on greater complexity in mental healthcare. In response, I have designed an analytical framework for scrutinizing the effects of delivered care on psychiatric patient outcomes. This thesis comprises three distinct studies with the overarching themes of mental healthcare management.In the first study, I investigate the operational characteristics of hospitals contributing to the readmission of psychiatry patients, shortly after being discharged. I propose that the length of stay in the inpatient ward mediates the effects of hospital characteristics on the risk of readmission. I utilize a data set of about 15,000 psychiatry patients admitted to 25 hospitals in Quebec, Canada, using a probit model adjusted for endogeneity through instrumental variables to conduct a causal analysis. I illustrated that the number of patients admitted to a hospital annually, i.e., patient volume, increases the risk of readmission, whereas this risk reduces with the hospital specializing in certain diagnosis classes, i.e., hospital focus. These relationships are moderated by patients' intensity of resource usage at the emergency department. Moreover, a nonlinear relationship between LOS and the risk of readmission is highlighted.In the second project, I apply a Difference-in-Differences (DID) framework to assess the causal effects of the COVID-19 pandemic on psychiatric patient outcomes. This study specifically examines changes in the probability of admission and the timing of return visits to the emergency department (ED), informed by a dataset tracking individual visits daily across a three-year span. The ED return time is the primary outcome of interest which is the time between two successive ED visits of a patient. This quantity can indicate the overall quality of care the patient receives in the ED visit. The analysis is segmented by various stages of the pandemic, with the Oxford COVID-19 Government Response Tracker (OxCGRT) providing insights into the impact of policy interventions. The results reveal that the pandemic reduced the ED return time of psychiatric patients by lowering their hospitalization likelihood. The study underscores the complex interplay between a reduction in ED visits, likely due to concerns of contracting the virus, and an increased demand for emergency psychiatric care as mental health crises intensified amidst pandemic conditions.In the third study, I explore the utilization of advanced interpretability techniques applied to deep learning models, particularly Bidirectional Encoder Representations from Transformers (BERT), to analyze psychiatric patient clinical notes for predicting 30-day readmissions in a single hospital. I illustrate how machine learning (ML) models can unravel insightful details within the unstructured data of clinical notes, aiding healthcare providers in understanding patient conditions and risks. More specifically, I illustrate how the ML algorithms predict the 30-day readmission likelihood of a patient by detecting the most influential words. Moreover, I highlight how this chapter can pivot towards the integration of text analytics approaches into hospital decision support systems and explore the potential of large language models (LLMs) to further refine the analysis of clinical narratives for more accurate patient outcome predictions.

English

ISBN: 9798342116695Subjects--Topical Terms:

1961957
Patients.

An Analytical Framework for Mental Healthcare Operations Management /
LDR:08940nmm a22004333i 4500 001 2438243
005 20251210113952.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798342116695
035 $a (MiAaPQD)AAI31527265
035 $a (MiAaPQD)McGill8c97kx020
035 $a AAI31527265
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Hejazian, Hossein, $e author. $3 3829487
245 1 3 $a An Analytical Framework for Mental Healthcare Operations Management / $c Hossein Hejazian.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (240 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-04, Section: B.
500 $a Advisors: Nasiry, Javad; Verter, Nafiz Vedat.
502 $b Ph.D. $c McGill University (Canada) $d 2024.
520 $a Healthcare operations management is dedicated to enhancing the processes within health systems to optimize the quality of care and improve patient outcomes. This field's challenges take on greater complexity in mental healthcare. In response, I have designed an analytical framework for scrutinizing the effects of delivered care on psychiatric patient outcomes. This thesis comprises three distinct studies with the overarching themes of mental healthcare management.In the first study, I investigate the operational characteristics of hospitals contributing to the readmission of psychiatry patients, shortly after being discharged. I propose that the length of stay in the inpatient ward mediates the effects of hospital characteristics on the risk of readmission. I utilize a data set of about 15,000 psychiatry patients admitted to 25 hospitals in Quebec, Canada, using a probit model adjusted for endogeneity through instrumental variables to conduct a causal analysis. I illustrated that the number of patients admitted to a hospital annually, i.e., patient volume, increases the risk of readmission, whereas this risk reduces with the hospital specializing in certain diagnosis classes, i.e., hospital focus. These relationships are moderated by patients' intensity of resource usage at the emergency department. Moreover, a nonlinear relationship between LOS and the risk of readmission is highlighted.In the second project, I apply a Difference-in-Differences (DID) framework to assess the causal effects of the COVID-19 pandemic on psychiatric patient outcomes. This study specifically examines changes in the probability of admission and the timing of return visits to the emergency department (ED), informed by a dataset tracking individual visits daily across a three-year span. The ED return time is the primary outcome of interest which is the time between two successive ED visits of a patient. This quantity can indicate the overall quality of care the patient receives in the ED visit. The analysis is segmented by various stages of the pandemic, with the Oxford COVID-19 Government Response Tracker (OxCGRT) providing insights into the impact of policy interventions. The results reveal that the pandemic reduced the ED return time of psychiatric patients by lowering their hospitalization likelihood. The study underscores the complex interplay between a reduction in ED visits, likely due to concerns of contracting the virus, and an increased demand for emergency psychiatric care as mental health crises intensified amidst pandemic conditions.In the third study, I explore the utilization of advanced interpretability techniques applied to deep learning models, particularly Bidirectional Encoder Representations from Transformers (BERT), to analyze psychiatric patient clinical notes for predicting 30-day readmissions in a single hospital. I illustrate how machine learning (ML) models can unravel insightful details within the unstructured data of clinical notes, aiding healthcare providers in understanding patient conditions and risks. More specifically, I illustrate how the ML algorithms predict the 30-day readmission likelihood of a patient by detecting the most influential words. Moreover, I highlight how this chapter can pivot towards the integration of text analytics approaches into hospital decision support systems and explore the potential of large language models (LLMs) to further refine the analysis of clinical narratives for more accurate patient outcome predictions.
520 $a La gestion des operations de sante est dediee a l'amelioration des processus au sein des systemes de sante pour optimiser la qualite des soins et ameliorer les resultats pour les patients. Les defis de ce domaine prennent une complexite plus grande dans le soin de sante mentale. En reponse, j'ai concu un cadre analytique pour examiner les effets des soins delivres sur les resultats des patients psychiatriques. Cette these comprend trois etudes distinctes avec pour themes principaux la gestion des soins de sante mentale.Dans la premiere etude, j'investigue les caracteristiques operationnelles des hopitaux contribuant a la readmission des patients en psychiatrie, peu apres leur sortie. Je propose que la duree du sejour dans le service d'hospitalisation medie les effets des caracteristiques de l'hopital sur le risque de readmission. J'utilise un ensemble de donnees d'environ 15 000 patients en psychiatrie admis dans 25 hopitaux au Quebec, Canada, en utilisant un modele probit ajuste pour l'endogeneite a travers des variables instrumentales pour conduire une analyse causale. J'ai illustre que le nombre de patients admis annuellement dans un hopital, c'est-a-dire, le volume de patients, augmente le risque de readmission, tandis que ce risque diminue avec l'hopital se specialisant dans certaines classes de diagnostic, c'est-a-dire, le focus de l'hopital. Ces relations sont moderees par l'intensite de l'utilisation des ressources par les patients au service des urgences. De plus, une relation non lineaire entre la duree de sejour et le risque de readmission est mise en evidence.Dans le deuxieme projet, j'applique un cadre de Difference dans les Differences (DID) pour evaluer les effets causaux de la pandemie de COVID-19 sur les resultats des patients psychiatriques. Cette etude examine specifiquement les changements dans la probabilite d'admission et le timing des retours aux urgences (ED), informee par un ensemble de donnees suivant les visites individuelles quotidiennement sur une periode de trois ans. Le temps de retour a l'ED est le principal resultat d'interet qui est le temps entre deux visites successives a l'ED d'un patient. Cette quantite peut indiquer la qualite globale des soins que le patient recoit lors de la visite a l'ED. L'analyse est segmentee par diverses etapes de la pandemie, avec le Tracker de Reponse Gouvernementale COVID-19 d'Oxford (OxCGRT) fournissant des insights sur l'impact des interventions politiques. Les resultats revelent que la pandemie a reduit le temps de retour a l'ED des patients psychiatriques en diminuant leur probabilite d'hospitalisation. L'etude souligne l'interaction complexe entre une reduction des visites a l'ED, probablement due aux preoccupations de contracter le virus, et une demande accrue pour les soins psychiatriques d'urgence alors que les crises de sante mentale s'intensifiaient dans les conditions de pandemie.Dans la troisieme etude, j'explore l'utilisation de techniques d'interpretabilite avancees appliquees aux modeles d'apprentissage profond, en particulier les Representations de l'Encodeur Bidirectionnel a partir de Transformateurs (BERT), pour analyser les notes cliniques des patients psychiatriques afin de predire les readmissions a 30 jours dans un seul hopital. J'illustre comment les modeles d'apprentissage machine (ML) peuvent devoiler des details perspicaces au sein des donnees non structurees des notes cliniques, aidant les prestataires de soins de sante a comprendre les conditions et les risques des patients. Plus specifiquement, j'illustre comment les algorithmes ML predisent la probabilite de readmission a 30 jours d'un patient en detectant les mots les plus influents. De plus, je souligne comment ce chapitre peut pivoter vers l'integration des approches d'analyse de texte dans les systemes de support decisionnel des hopitaux et explorer le potentiel des modeles de langage de grande taille (LLMs) pour affiner davantage l'analyse des recits cliniques pour des predictions de resultats des patients plus precises.
546 $a English
590 $a School code: 0781
650 4 $a Patients. $3 1961957
650 4 $a Emergency medical care. $3 3563184
650 4 $a Psychiatry. $3 517037
650 4 $a Deep learning. $3 3554982
650 4 $a Length of stay. $3 3560751
650 4 $a Pandemics. $3 3464080
650 4 $a Mental health. $3 534751
650 4 $a Large language models. $3 3775019
650 4 $a Hospitalization. $3 825068
650 4 $a Clinical outcomes. $3 3682127
650 4 $a COVID-19. $3 3554449
650 4 $a Epidemiology. $3 568544
690 $a 0347
690 $a 0800
690 $a 0766
690 $a 0769
690 $a 0796
710 2 $a McGill University (Canada). $e degree granting institution. $3 3759190
720 1 $a Nasiry, Javad $e degree supervisor.
720 1 $a Verter, Nafiz Vedat $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-04B.
790 $a 0781
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31527265