Machine learning-based analysis of adverse events in mesh implant surgery reports

Abstract

AbstractMesh implant surgery, commonly used for various medical conditions, has been linked to a range of negative effects, significantly impacting patient experiences and outcomes. Additionally, the growing volume of medical data, especially text-based reports, presents challenges in deriving meaningful insights for informed healthcare decisions. To address these challenges, our study extensively analyzed the Manufacturer and User Facility Device Experience (MAUDE) dataset from 2000 to 2021. We have meticulously identified a range of adverse events associated with mesh implant surgeries, including Dyspareunia, Urinary problems, Chronic Inflammation, Prolapse Recurrence, Mesh Erosion, Urinary Tract Infections, discomfort, and sleep disturbances. Using topic modeling, we explored patient experiences and the interrelationships among these adverse events. This approach uncovered key topics linked to mesh surgery, such as Stress Urinary Incontinence, Incisional Hernia, Inguinal Hernia, and Umbilical Hernia, along with their side effects. While the analysis focused on common symptoms such as pain, infection, and bleeding, it also brought to light specific symptoms like sleeping issues, mental stress, and discomfort. We also examined the interconnectedness of these adverse events with identified topics and their temporal trends, revealing shifts in patient experiences over time. Notably, there was an increase in reports of Stress Urinary Incontinence around 2011–2012 and a surge in Inguinal Hernia concerns in 2017–2018. This study provides a comprehensive understanding of adverse events and associated topics in mesh implant surgeries, contributing valuable insights into patient experiences and aiding in informed healthcare decision-making.