Peer-reviewed article
Mapping Indonesian Public Discourse on Oral Health: A Content Analysis of Social Media Using Latent Dirichlet Allocation Topic Modelling.
Amalia R, Dewanto I, Wirasto A, Sandy LPA ·
Summary
Written by AI, not yet AICS's view · claude-haiku-4-5-20251001This peer-reviewed study analyzed social media posts about oral health from Indonesian platforms Quora and Threads over 18 months to identify public discourse patterns. Using automated data extraction and topic modeling techniques, researchers identified 10 dominant themes including oral hygiene, children's dental visits, and mouth ulcers. The study demonstrates that social media can complement traditional epidemiological monitoring to guide public health communications about oral health.
Key points
- 858 posts analyzed from Quora (541) and Threads (317) over 18 months
- Most frequently discussed issues were mouth ulcers and toothache
- 10 dominant themes identified through Latent Dirichlet Allocation modeling
- Quora discussions were symptom-focused while Threads contained emotionally expressive narratives
- Significant association found between platform type and topics discussed (P < .001)
Limitations
This study focuses on discourse analysis and does not evaluate health outcomes or the effectiveness of any intervention. It provides insights into what people discuss online but not whether children's oral health actually improved or what would be needed to improve it.
Abstract
Background The potential of social media that offers public health insights in middle-income countries remains underexplored. This study examines patterns of oral health discourse on Indonesian social media and highlighting such insights into promotive and preventive oral health strategies. Methods A discourse and content analysis was conducted on user-generated posts from Quora and Threads over an 18-month period (July 1, 2023-January 31, 2025). Posts were retrieved through automated data extraction and computational text retrieval based on keyword searches and coded into 19 predefined categories. Latent Dirichlet allocation (LDA) was applied to identify underlying thematic structures. Associations between platforms and discussion topics were tested using chi-square and Monte Carlo simulations. Comparative analyses examined platform-specific trends, thematic frequencies and inquiry types, with findings cross-referenced against national survey data. Results A total of 858 posts (541 from Quora, 317 from Threads) were analysed. The most frequently discussed issues were mouth ulcers and toothache. LDA revealed 10 dominant themes, including oral hygiene practices, children's dental visits, gum swelling and postsurgical recovery. Quora discussions were dominated by symptom-focused inquiries, whereas Threads contained more personal and emotionally expressive narratives. Statistical testing confirmed a significant association between platform type and topics discussed (P < .001). Conclusion This study demonstrated the utility of automated data extraction and LDA-based thematic modelling in uncovering community-level oral health patterns. Social media provides an immediate context-relevant perspective for monitoring oral health needs, supporting targeted communication strategies and responsive public health interventions. Clinical relevance This study introduces a complementary approach to conventional epidemiological monitoring. The findings can guide policymakers and practitioners in developing targeted preventive initiatives and public engagement strategies for improved oral health outcomes.