Machine learning for pattern detection in cochlear implant FDA adverse event reports.

Area of Science:

• Biomedical Engineering
• Data Science
• Medical Device Safety

Background:

• Medical device databases offer insights into patient safety and design improvements.
• Analyzing adverse event reports is crucial for understanding medical device performance.

Purpose of the Study:

• To apply supervised machine learning to identify patterns in cochlear implant (CI) adverse events.
• To predict CI manufacturer and adverse event type using machine learning algorithms.

Main Methods:

• Acquired adverse event reports for top CI manufacturers from a U.S. government database.
• Utilized four supervised machine learning algorithms: random forest, linear SVC, naïve Bayes, and logistic regression.
• Evaluated model performance based on classification prediction accuracy.

Main Results:

• Most adverse events reported were patient injury (16,736), followed by device malfunction (10,760), and death (16).
• Machine learning models achieved high accuracy in predicting CI manufacturer from adverse event narratives: logistic regression (88.6%), naïve Bayes (88.5%), linear SVC (86.0%), and random forest (74.8%).

Conclusions:

• Supervised machine learning models can effectively predict cochlear implant manufacturer and adverse event type from text descriptions.
• These findings highlight the potential of machine learning for enhancing medical device safety and informing design improvements.