An Analytical Framework for Mental Healthcare Operations Management

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 Abstract x 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 Abstract xi 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.xii AbrgLa gestion des oprations de sant est ddie l'amlioration des processus au sein des systmes de sant pour optimiser la qualit des soins et amliorer les rsultats pour les patients.Les dfis de ce domaine prennent une complexit plus grande dans le soin de sant mentale.En rponse, j'ai conu un cadre analytique pour examiner les effets des soins dlivrs sur les rsultats des patients psychiatriques.Cette thse comprend trois tudes distinctes avec pour thmes principaux la gestion des soins de sant mentale.Dans la premire tude, j'investigue les caractristiques oprationnelles des hpitaux contribuant la radmission des patients en psychiatrie, peu aprs leur sortie.Je propose que la dure du sjour dans le service d'hospitalisation mdie les effets des caractristiques de l'hpital sur le risque de radmission.J'utilise un ensemble de donnes d'environ 15 000 patients en psychiatrie admis dans 25 hpitaux au Qubec, Canada, en utilisant un modle probit ajust pour l'endognit travers des variables instrumentales pour conduire une analyse causale.J'ai illustr que le nombre de patients admis annuellement dans un hpital, c'est--dire, le volume de patients, augmente le risque de radmission, tandis que ce risque xvi an immeasurable level of moral support, patience, and understanding throughout my Ph.D. studies.Among them, I want to express my particular appreciation to three persons.First, my brother, Ali Hejazian, who was the pillar of my life.Next, my dearest friends, Dr.Ata Jalili Marand and Dr. Milad Jamali, whose presences were the sources of trust and

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.