Abstract
Purpose: Digital communication has transformed the landscape of psychosocial oncology information for survivors and professionals, potentially influencing the unmet informational needs and supports of cancer survivors and healthcare professionals. This study aims to explore the psychosocial oncology information on social media and generative artificial intelligence (GenAI) using a topic modeling approach.
Methods: A qualitative design incorporating structural topic modeling (STM) was utilized to analyze the psychosocial oncology information from social media (WeChat and Rednote) and GenAI (DeepSeek and ChatGPT). Social media data (n = 1,177) were collected using the main keywords "cancer" and "psychosocial", while GenAI data were collected through nine standardized prompts. STM identified topic prevalence, keywords, topic relations, and temporal trends specific to social media posts.
Results: Five topics were identified from social media, highlighting caregivers' reflection, psychological intervention and assessment, family dynamics, and pediatric cancer care. Sharing caregiving reflection as an emerging trend for social support. STM identified seven topics in DeepSeek and five in ChatGPT, commonly focusing on psychological distress, coping strategies and support systems, healthcare professional practices, and holistic care. DeepSeek demonstrated more critical reflection on survivorship challenges and psychosocial care barriers, but identified misinformation. Social media presents in-depth topics, whereas GenAI offers synthesized overviews. Help-seeking resources were limited across both sources.
Conclusions: Findings emphasized the significance of information strategies in psychosocial oncology care. GenAI should be fine-tuning/training with relevant data for improving practices. Social media and content creators are effective tools for promoting evidence-based information. A hybrid model that integrates professional oversight, GenAI, and cloud platforms is essential to ensure ethical compliance with psychosocial care.
Methods: A qualitative design incorporating structural topic modeling (STM) was utilized to analyze the psychosocial oncology information from social media (WeChat and Rednote) and GenAI (DeepSeek and ChatGPT). Social media data (n = 1,177) were collected using the main keywords "cancer" and "psychosocial", while GenAI data were collected through nine standardized prompts. STM identified topic prevalence, keywords, topic relations, and temporal trends specific to social media posts.
Results: Five topics were identified from social media, highlighting caregivers' reflection, psychological intervention and assessment, family dynamics, and pediatric cancer care. Sharing caregiving reflection as an emerging trend for social support. STM identified seven topics in DeepSeek and five in ChatGPT, commonly focusing on psychological distress, coping strategies and support systems, healthcare professional practices, and holistic care. DeepSeek demonstrated more critical reflection on survivorship challenges and psychosocial care barriers, but identified misinformation. Social media presents in-depth topics, whereas GenAI offers synthesized overviews. Help-seeking resources were limited across both sources.
Conclusions: Findings emphasized the significance of information strategies in psychosocial oncology care. GenAI should be fine-tuning/training with relevant data for improving practices. Social media and content creators are effective tools for promoting evidence-based information. A hybrid model that integrates professional oversight, GenAI, and cloud platforms is essential to ensure ethical compliance with psychosocial care.
| Original language | English |
|---|---|
| Article number | 103018 |
| Number of pages | 14 |
| Journal | European Journal of Oncology Nursing |
| Volume | 79 |
| Early online date | 20 Oct 2025 |
| DOIs | |
| Publication status | Published - Dec 2025 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 3 Good Health and Well-being
User-Defined Keywords
- Psychosocial oncology
- Oncology nursing
- Generative artificial intelligence
- Social media
- Machine learning
- Caregivers
- Cancer survivors
- Health communication
Fingerprint
Dive into the research topics of 'How do GenAI and social media provide psychosocial information support for cancer survivors and professionals? An analysis using structural topic modeling for exploration and comparison'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver