JMIR Publications

ABSTRACT

Background:

In the era of internet-based governance, online public appeals—particularly in the healthcare domain—have become a vital channel for expressing citizens’ demands and a focal point for researchers and policymakers seeking to improve service delivery and responsiveness.

Objective:

This study aims to uncover the thematic structure, emotional tone, and governmental response logic associated with public medical appeals in China, thereby offering empirical insights to enhance the responsiveness and quality of health-related public services.

Methods:

A dual-track analytical framework was developed, incorporating both textual and contextual (non-textual) factors. (1) Key themes of public appeals were identified using the Term Frequency–Inverse Document Frequency (TF-IDF) model to extract feature words, followed by hierarchical cluster analysis. (2) Sentiment classification was conducted via supervised machine learning, with additional validation using sentiment scores derived from a lexicon-based approach. (3) A binary logistic regression model was employed to assess the effects of textual factors (appeal theme, content, sentiment polarity, and title clarity) and non-textual factors (issue resolution difficulty, benefit attribution, internet penetration, education level, and economic development) on the likelihood of receiving a government response. A Probit model was used to validate the robustness of the results.

Results:

Public medical appeals mainly focused on pandemic control, fertility policies, healthcare institutions, and insurance systems. Negative sentiment was predominant, accounting for 85.8% (3328/3877) of complaint/help-seeking posts, 70.0% (1666/2381) of consultation posts, and 65.6% (1710/2606) of suggestion posts. Logistic regression results indicated that textual characteristics, issue complexity, and benefit-related content were not significant predictors of government responsiveness (P > .05). In contrast, internet penetration (P < .01), education level (P < .001), and economic development (P < .001) were positively and significantly associated with the probability of obtaining a government response.

Conclusions:

Public medical appeals exhibit five key characteristics: urgency driven by pandemic conditions, linkage to fertility policies, tensions between medical service efficiency and costs, cross-regional insurance coverage challenges, and a predominance of negative sentiment. The findings suggest that textual and issue-specific content have limited influence on government responsiveness—possibly due to the politically sensitive and complex nature of healthcare topics. Instead, macro-level environmental factors play a decisive role. These findings contribute to optimizing response mechanisms on digital health platforms and provide new theoretical and empirical perspectives for advancing health information dissemination and digital governance in the healthcare sector.