A Predictive Model for Forecasting the Severity of Recall of Cardiac Implantable Electronic Devices

This research focuses on developing a predictive model of the severity of recalls related to cardiac implantable electronic devices. Various machine learning models such as Logistic Regression, Naive Bayes, Random Forest, and Deep Neural Networks were used to build predictive models by analyzing textual data from the adverse event reports from the FDA MAUDE database. Data preprocessing techniques such as undersampling and SMOTE were applied to the imbalanced dataset. The Random Forest model showed the best performance in classification around recall severity, thus proving helpful in the post-market surveillance of medical devices, and had the highest accuracy (>0.86). DNN also had strong model performance and accuracy. Feature importance scores also underscored certain words associated with the severity of recall (moderate or severe) or no recall. This study aims to advance post-market surveillance and active recall management

Author

• Mendis, John Udara

Language

• English

Keyword

• Cardiac Implantable Electronic Devices
• Recall Severity Forecasting
• FDA MAUDE Database
• Machine Learning
• Predictive Modeling
• Post-Market Surveillance

Date created

• 2025

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Engineering Management

Degree

• D.Engr.

Advisor

• Blackford, J.P.
• Etemadi, Amir

Committee Member(s)

• Ali, Daniyal

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/kk91fm55c