Year: 2022
Type: dissertation
Abstract: Recent technological advancements (i.e., internet of things [IoT]), radio-frequency identification technology, electronic medical records/electronic health records) and new achievements in machine learning have provided hospitals the platform necessary to invest in data-driven modern technologies to enhance care delivery and therefore to reduce cost and increase revenue generation. Therefore, this Ph.D. research identifies the impacts and demonstrates the applications of these advanced technologies to shape the current hospital system by accomplishing three projects: (i) Evaluate the impacts of IoT on the quality of care domains and staff requirements of a hospital unit, (ii) Design a validation framework and apply it to evaluate the impact of a smart privacy auditing tool on operational excellence, and (iii) Predict the number of beds to clean within the next few hours as well as foresee the staff requirements in the emergency department (ED) applying machine learning algorithms and discrete-event simulation (DES) modeling to facilitate decision-making and reduce ED crowding. To identify the impacts of IoT on quality care domains (e.g., patient safety, efficiency, timeliness), a pre-/post-intervention analysis of patient falls and length of stay as well as a post-intervention analysis on hand hygiene compliance rate and call response time has been conducted using interrupted time series and other statistical analyses. Additionally, semi-structured interviews to find staff experiences have been performed. A favorable impact of IoT on these outcome measures is reported. The black box method of user acceptance testing has been adopted to design the evaluation framework for the automated privacy auditing tools, considering several scenarios on appropriate and inappropriate behavior. Validation of the framework with one hospital’s data shows that the effectiveness of one such tool is 98.09%. Poisson regression performs the best for predicting the number of beds to clean at the ED. Prediction model results are used as the input for a DES model to identify staff requirements and to assist in decision-making. This data-driven Ph.D. research unfolds the evidence of the consequences of smart hospital applications and solves the current complex problems related to decision making. Thus, this research informs hospital management and policymakers about the effectiveness of smart hospital applications.
Source: TSpace (University of Toronto)
Author: Yesmin, Tahera
Institution: University of Toronto
Language: English
Field: Decision Sciences
Domain: Social Sciences
Keywords: smart hospitals, Internet of Things, hospital quality of care, patient falls, length of hospital stay, hand hygiene compliance, user acceptance testing, emergency department overcrowding, staffing requirements, discrete event simulation, electronic health records, radio frequency identification, hospital management decision-making
Open Access status: green
OpenAlex ID: W7133033064
