Understanding Communication in an Online Cancer Forum: Content Analysis Study

Area of Science:

• Computational linguistics
• Psycholinguistics
• Social sciences

Background:

• Online cancer forums serve as crucial platforms for individuals affected by cancer to share experiences and seek support.
• The support needs and concerns expressed may vary significantly between cancer patients and their family members or friends.
• Language used in online posts can reflect these distinct user perspectives and needs.

Purpose of the Study:

• To investigate the differences in support needs and concerns between cancer patients and their support network (family/friends) in online forums.
• To utilize natural language processing (NLP) to analyze language patterns in forum posts from these distinct user groups.

Main Methods:

• Latent Dirichlet Allocation (LDA), a topic modeling technique, was employed to identify key themes in user posts.
• Linguistic Inquiry and Word Count (LIWC), a psycholinguistic dictionary, was used to analyze word usage and emotional content.
• Posts from self-declared cancer patients were compared with those from self-declared family members or friends.

Main Results:

• Cancer patients' posts frequently discussed hospital visits, health-related terms, and anxiety.
• Family members and friends' posts were more associated with topics of losing a loved one, past-oriented language, and death.
• Statistically significant differences (Cohen's d values provided) were observed in topic and word category associations between the two groups.

Conclusions:

• Distinct linguistic patterns and concerns emerge from online cancer forum posts by patients versus their caregivers.
• Understanding these differences is essential for providing appropriate and effective support in online cancer communities.
• Responses to online forum posts should be tailored to acknowledge the specific needs and perspectives of patients and their support networks.