Dissertation/Thesis Abstract

An Automated Model to Estimate the Probability of a Use Error Related Adverse Event for In Vitro Diagnostic Medical Devices
by Jengelley, Toni-Ann N., D.Engr., The George Washington University, 2019, 155; 22617835
Abstract (Summary)

The FDA, the MHRA, and other regulatory authorities recommend that during the development process of a device, manufacturers should aim to understand the use errors of comparable devices to the ones of interest. Knowing the probability and severity of use errors for similar products, they can be eliminated or reduced by implementing HFE/UE principles related to them. In this study, the MAUDE database was used as the data source to create an automated model that is able to estimate the probability of use related errors associated with IVD devices. Several characteristics related to the device, operator, error type and location were found to be important in identifying the probabilities of a use error related adverse event that are readily available to a user of the proposed model and do not require a burdensome number of characteristics to generate accurate probability results. The final model provides an objective and time saving approach using the Bootstrap Forest algorithm with these characteristics. It is shown to accurately characterize use error related adverse events with a generalized R-squared value of 0.8587 and provides a highly accurate method with a low misclassification rate of 6.95% and is an effective model for distinguishing if an event is an adverse event with a high AUC of 97.5%. In addition, a knowledge model for use errors is utilized that provides an understanding from a human factor and usability perspective and allows the design team to address the design based on the cognitive areas that are impacted for the new device rather than a specific design issue. The long term goal is to facilitate device design improvements to ensure safety and prevent patient injury and death caused by use errors adverse events associated with IVD medical devices.

Indexing (document details)
Advisor: Grenn, Michael
Commitee: Etemadi, Amir, Malalla, Ebrahim
School: The George Washington University
Department: Engineering Management
School Location: United States -- District of Columbia
Source: DAI-B 81/2(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Industrial engineering, Biomedical engineering, Engineering
Keywords: Adverse event, Human factors engineering, IVD device, Machine learning, MAUDE database, Use error
Publication Number: 22617835
ISBN: 9781085673136
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest