Abstract

Objectives: The study aims to analyze social media posts to understand patient perceptions and emotions regarding BII and to assess the relationship between the sentiment of these posts and breast implant removal rates. Introduction: Breast Implant Illness (BII) is identified by patients as a range of symptoms related to breast implants, with increasing public interest despite no conclusive research on causality. Social media has become a key platform for sharing patient experiences and perceptions of BII. The RoBERTa NLP model, trained on extensive social media data, offers insights into these discussions. Materials / method: Posts mentioning BII from January 2014 to December 2023 on platform X were collected and analyzed using two NLP models: a binary sentiment model and an emotion-based model. The models categorized posts as positive or negative and identified dominant emotions. Correlations between post sentiments, emotions, and implant removal data were examined using the Pearson correlation coefficient. Results: Out of 6,099 posts over ten years, the majority were negative (75.4%), with neutrality (35.9%) and fear (35.6%) being the most common emotions. Excluding neutral posts, fear dominated 55.5% of the discourse. A strong positive correlation (r > 0.80) was found between negative, neutral, and fear posts and the rates of implant removal. Conclusion: The analysis reveals that discussions on BII are mainly negative, with a significant focus on fear, which correlates with the increase in breast implant removals. This highlights the role of NLP in understanding patient sentiments and aiding healthcare providers in addressing patient concerns.