Analyzing Patient Experience on Weibo: Machine Learning Approach to Topic Modeling and Sentiment Analysis

Area of Science:

• Health Services Research
• Social Media Analytics
• Patient Experience Research

Background:

• Social media platforms are crucial for understanding patient experiences with healthcare services.
• These platforms facilitate open communication and information sharing about healthcare interactions.

Purpose of the Study:

• To identify common patient experience topics discussed on social media.
• To pinpoint patient concerns that healthcare providers should address.

Main Methods:

• Spatiotemporal analysis of Weibo posts for volume, sentiment, and topics.
• Supervised machine learning (logistic regression, Naive Bayes, random forest) for topic modeling and sentiment analysis.
• Human annotation and multiclassifier voting for robust analysis.

Main Results:

• Over 4000 posts analyzed, with 82.81% expressing negative sentiment.
• Twenty patient experience themes identified, with interpersonal aspects (73.53%) being most discussed.
• Top topics: healthcare professional attitude, access to care, communication, technical competence, and treatment efficacy.

Conclusions:

• Healthcare providers should monitor social media for patient feedback.
• Machine learning enhances the analysis of social media data for patient experience insights.
• Interpersonal aspects of care are critical "moments of truth" in patient evaluations of hospital services.