Abstract
OBJECTIVES: This study aims to construct and validate an interdisciplinary framework based on Library and Information Science (LIS) to improve the timeliness and accuracy of health education resource development during public health emergencies, and to provide a practical technical approach and theoretical framework through a complete "analysis-generation-evaluation" cycle for resolving the conflict between "information overload" and "precise targeting" in crisis communication. METHODS: A total of 1,026 epidemic bulletins from various levels of government in China (2020-2024) were collected as the primary data source. In-depth knowledge development was achieved through core Library and Information Science (LIS) methods such as knowledge graph construction, thematic analysis, natural language processing (NLP), and association rule mining. Building upon these analytical results, an automated resource generation system was developed based on the Technology Acceptance Model (TAM). The system was subsequently evaluated using questionnaires administered to 305 users. RESULTS: A topic modeling analysis was conducted on 1,026 epidemic announcements, revealing five themes, with preventive measures being the most prominent (32.7%). Association rule mining indicated significant co-occurrence patterns among key protective factors (support >0.6, confidence >0.8). An automated resource generation system based on the Technology Acceptance Model (TAM) was evaluated using 305 valid questionnaires, showing a high level of user acceptance. Specifically, Path analysis confirmed that perceived usefulness (β = 0.42, p < 0.001) was the strongest predictor of behavioral intention, followed by perceived ease of use (β = 0.31, p < 0.01). Logistic regression further showed that trust in official information sources (OR = 2.05) and eHealth literacy levels (OR = 1.87) were important factors influencing perceived resource effectiveness. CONCLUSION: The core of this study established a pathway that rapidly and automatically converts authoritative epidemic announcements into personalized health education resources. The framework utilizes LIS technologies such as knowledge graphs and association rule mining to analyze the content of the announcements and achieve automatic resource generation. Empirical research shows that user acceptance of these resources depends primarily on their perceived usefulness and ease of use, with eHealth literacy playing an important moderating role in this process. The study's "analyze-generate-evaluate" closed-loop model can be extended to other crisis situations.