关键词: 公共图书馆, 健康信息服务空间, 信息共享空间, 用户体验障碍, 信息行为, 文本挖掘, Louvain聚类

Abstract:

[Purpose/Significance] Public libraries are increasingly expected to support health knowledge dissemination, health literacy improvement, and equitable access to trustworthy information. In China, the policy context shaped by the Healthy China strategy, national health literacy initiatives, and the public service mission of libraries, has created a practical demand for health information services that extend beyond simply providing collections. Existing studies have mainly examined service items, activity models, or institutional pathways, while relatively limited attention has been paid to whether the supply of physical and digital service elements is aligned with users' actual experiences. To address this gap, this study develops an analytical framework of "supply coverage–experience barriers–supply-demand adaptation" and applies it to health information service spaces in public libraries. The contribution of this study lies in combining verifiable supply-side evidence with user-generated review texts, thereby shifting the discussion of service space from normative description to evidence-based diagnosis and differentiated governance. This study also clarifies the conceptual boundary of health information service space. The core service layer consists of health-themed resources, consultations, activities, online access points, and interaction mechanisms. Lighting, seating, hygiene, accessibility, and other basic environmental conditions are supporting conditions that affect accessibility, continuity of use, and perceived service quality. These conditions are not considered core health information services themselves. [Method/Process] This study adopts an extended secondary data analysis design. Based on a publicly released dataset and supplementary verification materials, supply-side evidence from 20 public libraries was reorganized into 12 binary indicators, including six physical-space indicators and six digital-space indicators. Each indicator was coded as 1 when explicit and verifiable public evidence was available and as 0 when such evidence was absent or insufficient. Physical-space coverage, digital-space coverage, and overall coverage were then calculated, and the median values of physical and digital coverage were used to construct a four-quadrant supply typology. On the demand side, 658 valid health-related user reviews from a mainstream consumer review platform were cleaned and analyzed. Text similarity networks, TF-IDF character n-grams, Louvain community detection, manual verification, and barrier coding were used to identify thematic clusters and experience barriers. For the supply-demand matching analysis, 12 libraries with complete supply-side evidence and at least 10 valid health-related reviews were retained. Spearman correlation analysis was used to examine exploratory associations between coverage indicators, average health-related ratings, and the proportions of different barrier types. Relative indicators, rather than absolute review counts, were used to reduce the influence of differences in city activity, library size, and platform popularity. [Results/Conclusions] The results show that the sampled libraries present a pattern of medium overall coverage and structural differentiation. The high physical–high digital coordination type accounts for the largest share, but low physical-low digital and unevenly developed types remain visible, indicating that physical and digital health information service elements have not yet formed a balanced system in all libraries. User reviews are mainly concentrated in five themes: environmental experience and basic services, health education and outreach activities, accessibility and health resource acquisition, mental health and emotional support, and sports and health management. High-intensity barriers are concentrated in supporting environmental conditions and key service touchpoints, including hygiene problems, insufficient seating, crowding, service response, appointment restrictions, platform usability, and access-point visibility. Correlation results indicate that physical-space coverage is significantly and positively associated with average health-related ratings, while the associations between digital-space coverage and platform barriers, and between physical-space coverage and supporting environmental barriers, are negative but not statistically significant. These findings should therefore be interpreted as exploratory rather than causal. Accordingly, public libraries should adopt differentiated governance strategies based on their supply typology. Libraries with extensive physical and digital resources should transition from ensuring basic availability to focusing on quality control and sustainable operations. Physical-dominant libraries should improve unified online access points, appointment processes, and feedback channels. Digital-dominant libraries should reinforce on-site capacity, accessibility, and order management. Libraries with limited resources on both sides should first establish basic accessible and usable service touchpoints before introducing more complex intelligent tools. The limitations of this study include the small matched sample, the limited visibility of public evidence, the use of secondary data, and possible biases among users of a single review platform. Future research may use a combination of larger regional samples, internal library service data, questionnaires, interviews, and longitudinal tracking to examine the effects and improvement mechanisms of health information service spaces in public libraries more thoroughly.

Key words: public libraries, health information service spaces, information commons, user experience barriers, information behavior, text mining, Louvain clustering