JMIR Publications

ABSTRACT

Background:

In China, lung cancer remains a major public health concern and accounts for a substantial proportion of cancer-related deaths nationwide. However, limited research has examined public perceptions of lung cancer in the digital sphere, where health-related information is increasingly disseminated and accessed.

Objective:

This study aims to systematically examine patterns of public attention and perceptions toward lung cancer in China by integrating search engine query data and social media content, thereby enhancing current understanding of online health information dynamics related to lung cancer.

Methods:

Data were collected from Baidu Index (BI) (2011–2025) and Sina Weibo (2010–2025) to represent online search behavior and social media discourse on lung cancer, respectively. Spatiotemporal patterns of BI, per capita Baidu Index (PBI), and Weibo posts were examined to capture temporal trends and spatial variations. Additionally, the spatial autocorrelation of PBI was assessed using global and local Moran’s I statistics. PBI-related explanatory variables were assessed using a spatial panel Durbin model (SPDM). Topic modeling and lexicon-based sentiment analysis were applied to Weibo content to uncover thematic evolution and emotional polarity across year, gender groups and user types.

Results:

Public attention toward lung cancer, as reflected by BI, increased initially, peaked in 2019, and subsequently declined, whereas Weibo discussions demonstrated a fluctuating but generally upward trend before stabilizing after 2022. Similar temporal patterns were observed across most provinces. Significant spatial heterogeneity was identified, with higher BI levels concentrated in eastern coastal regions and persistently lower levels in western and southwestern provinces. Spatial autocorrelation analysis revealed stable positive clustering over time, with low-low clusters particularly concentrated in southwestern regions such as Guangxi and no significant high-high clusters detected. Panel spatial regression analyses indicated that the provincial PBI was positively associated with gross domestic product (GDP) per capita and average years of education per capita, but negatively associated with the urbanization rate. Moreover, significant spatial spillover effects were observed, suggesting that socioeconomic factors were associated with not only local public attention but also that of neighboring regions. Topic modeling revealed a clear thematic evolution over time. Although personal experiences initially dominated online discourse, discussions progressively shifted toward healthcare service–related issues, which became the most prominent theme by 2025. Sentiment analysis indicated an overall positive emotional tone throughout the study period, with “Good” and “Disgust” representing the predominant positive and negative emotions, respectively. Emotional expression varied across demographic groups and user types, with noticeable differences in both intensity and temporal trends.

Conclusions:

This study offers a comprehensive overview of public attention and discourse on lung cancer in China’s digital landscape, providing valuable evidence to inform targeted health communication and policy interventions.