JMIR Publications

ABSTRACT

Background:

Nowadays, new media has played an important role in providing information about cancer prevention and treatment. A growing body of work has been devoted to examining the access and communication effects of cancer information on social media. However, there has been limited understanding of the overall presentation of cancer prevention and treatment on social media. Further, research on comparing the differences between medical social media and common social media remained limited.

Objective:

Based on big data analytics, this study aimed to comprehensively map the characteristics of cancer treatment and prevention information on medical social media and common social media, which was promisingly helpful in cancer coverage and patients’ treatment decision.

Methods:

We collected all posts (N=60,843) from 4 medical WeChat official accounts (classified as medical social media in this paper), and 5 health and lifestyle WeChat official accounts (classified as common social media in this paper). By applying latent Dirichlet allocation topic model, we extracted cancer-related posts (N=8,427) and obtained 6 cancer themes in common social media and medical social media separately. After manually labeling posts according to our codebook, we adopted a neural-based method to label different articles automatically. To be more specific, we defined our task as a multi-label task and chose different pre-trained models, say, Bert and Glove, to learn document level semantic representations for labelling.

Results:

Themes in common social media were more related to lifestyle, while medical social media were more related to medical attributions. Early screening and testing, healthy diet, and physical exercise were the most frequently mentioned preventive measures. Compared with common social media, medical social media mentioned vaccinations to prevent cancer more frequently. Both types of media provided limited coverage of radiation prevention (including sun protection) and breastfeeding. Surgery, chemotherapy, and radiotherapy were the most mentioned treatment measures. Medical social media discussed treatment information more than common social media.

Conclusions:

Cancer prevention and treatment information on social media revealed a lack of balance. The focus on cancer prevention and treatment information was mainly limited to a few aspects. The cancer coverage on preventive measures and treatments in social media required further improvement. Additionally, the study's findings underscored the potential of applying machine learning to content analysis as a promising research paradigm for mapping the key dimensions of cancer information on social media. The findings provided methodological and practical significance in future study and health promotion.