Analyzing Patient Stories on Social Media Using Text Analytics

Area of Science:

• Health Informatics
• Natural Language Processing
• Patient Experience Research

Background:

• Social media platforms like Care Opinion collect numerous patient healthcare experiences.
• Analyzing this large volume of qualitative data is challenging for healthcare providers.

Purpose of the Study:

• To apply text mining techniques to analyze patient stories from the Care Opinion platform.
• To explore healthcare experiences and identify areas for quality improvement.

Main Methods:

• Collected 367,573 patient stories (September 2005-2019).
• Utilized topic modeling (Latent Dirichlet Allocation) and sentiment analysis.
• Identified 16 topics across five healthcare experience aspects.

Main Results:

• Identified key themes: communication, clinical/non-clinical service quality, human aspects, and satisfaction.
• 99% of stories had discernible sentiment.
• Over 55% of stories regarding information requests, treatment, or appointments showed negative sentiment, indicating dissatisfaction.

Conclusions:

• Text mining and sentiment analysis effectively analyze large patient story datasets.
• Patient stories offer valuable, specific insights for healthcare quality improvement.
• These narratives are distinct from general social media posts and crucial for understanding patient care.