Identifying Influences in Patient Decision-making Processes in Online Health Communities: Data Science Approach

Area of Science:

• Health Informatics
• Computational Social Science
• Artificial Intelligence in Healthcare

Background:

• Online health communities (OHCs) are increasingly used by patients for information and support.
• Understanding patient decision-making processes and influences within OHCs is crucial.

Purpose of the Study:

• To identify posts within discussion threads that influence users' health care decision-making.
• To analyze the impact of online health community interactions on user behavior.

Main Methods:

• Proposed a definition for influence relationships in OHC discussion threads.
• Developed a deep learning framework combining text relevance features and post characteristics (question/action probability).
• Utilized state-of-the-art text relevance measurement and deep learning for influence identification.

Main Results:

• Evaluated the approach on a cancer survivor OHC discussion dataset.
• Demonstrated the effectiveness of the proposed deep learning model in identifying influential posts.

Conclusions:

• Identifying influence relationships in OHCs is feasible and valuable.
• Influential discussions significantly affect user decision-making and engagement.
• Findings can enhance information quality, user engagement, and user experience in OHCs.