JMIR Publications

ABSTRACT

Background:

Depression, a highly prevalent global mental disorder, has prompted significant research concerning its association with social media usage and its impact during Hong Kong’s social unrest and COVID-19. However, other mainstream media, specifically online news, has been largely overlooked. Despite extensive research conducted in countries such as the United States, Australia, and Canada to investigate the latent subthemes, sentiments, and coping strategies portrayed in depression-related news, the landscape in Hong Kong remains unexplored.

Objective:

The purpose of this study is to uncover the latent subthemes that were presented in the online news coverage of depression in Hong Kong, examine the sentiment conveyed in the news, and assess whether coping strategies have been provided throughout the news for individuals experiencing depression.

Methods:

This study utilized natural language processing (NLP) techniques, namely the Latent Dirichlet Allocation Topic Modelling and Vader sentiment analysis, to fulfill the first and second objectives. Coping strategies were rigorously assessed and manually labeled with designated categories for content analysis. The online news was collected from February 2019 to May 2024 from Hong Kong mainstream news websites to examine the latest portrayal of depression, especially during and after the social unrest and COVID-19.

Results:

2,435 news articles were retained for analysis after the news screening process. A total of 7 subthemes were identified based on the topic modeling results. ‘Societal system,’ ‘Law enforcement,’ ‘Global recession,’ ‘Lifestyle,’ ‘Leisure,’ ‘Health issues,’ and ‘US politics’ were the latent subthemes. Moreover, the overall news exhibited a slightly positive sentiment. The correlations between the sentiment scores and the latent subthemes indicated that the ‘Societal system,’ ‘Law enforcement,’ ‘Health issues, ’ and ‘US politics’ revealed positive tendencies, while the remainders leaned toward a negative sentiment. The coping strategies for depression were substantially lacking; however, the category emphasizing the dissemination of information on skills and resources and individual adjustment to cope with depression emerged as the priority focus.

Conclusions:

This pioneering study utilized a mixed method approach where NLP is employed to investigate latent subthemes and underlying sentiment in online news. Content analysis was also performed to examine available coping strategies. The findings of this research enhance our understanding of how depression is portrayed through online news in Hong Kong and the preferable coping strategies being employed to mitigate depression. The potential impact on readers was discussed. Future research is recommended to address the mentioned limitations and apply the NLP techniques to new mental issue cases.