Abstract
Objective
Social media has emerged as a transformative platform for data-driven health communication about breast cancer. To reveal social cognition and enhance public awareness of breast cancer, this study analyzes emotional tendencies and core topics in Weibo discourse.
Methods
A web crawler was employed to collect 185,486 Weibo posts containing the keyword “breast cancer”, with 134,588 posts retained after preprocessing. The discourse analysis involved, first, a lexicon-based sentiment classification of the text data, and subsequently, the application of Latent Dirichlet Allocation (LDA) topic modeling of each identified sentiment.
Results
This study identified three dominant emotional categories—positive (40.5%), neutral (34.5%), and negative (25%)—along with distinct thematic patterns associated with each. Salient topics included “personal experiences” and “symptoms,” with positive posts emphasizing recovery narratives, clinical advances, and prevention; neutral discourse focusing on health management and support systems; and negative discourse centered on emotional distress, treatment challenges, and risk factors. Discourse analysis reveals the significance of emotional support, access to medical and insurance resources, and reliable scientific information in fighting breast cancer. Furthermore, breast cancer is discursively constructed through three key frames such as “a women’s issue”, “a public health crisis”, and an issue of “individual responsibility”. Various stakeholders—including media, government, medical institutions, and patient groups—played distinct roles in the discursive construction of breast cancer.
Conclusion
Our findings provide critical insights into public needs on breast cancer, informing the development of targeted strategies to enhance health awareness, shape attitudes, and improve stakeholder-led education and advocacy.
1. Introduction
Breast cancer is the most commonly diagnosed malignancy among women globally, accounting for approximately 25% of all female cancers and representing a leading cause of cancer-related mortality in this population.1,2 Incidence rates in men remain substantially lower. Beyond its physical burden, breast cancer profoundly impairs patients’ psychological well-being, social functioning, and quality of life. 3 Patients face heightened psychological distress and social stigma, which may impede treatment adherence and recovery. Socio-cultural factors—including health beliefs, habits, and healthcare perceptions—significantly shape screening practices. Given that breast cancer intersects the biological body, self-identity, and social structures, its socio-cultural dimensions merit critical scholarly attention. 3
Social media has become integral to public communication in the digital age, with its expanding user base offering a novel methodological avenue for observing social phenomena and public opinions. 4 Within health domains, social media serves as a significant information source. 5 Breast cancer, given its global prevalence, generates substantial discussions on platforms including Facebook, Twitter, Instagram, and Weibo, where patients, professionals, and the public share disease-related experiences and perspectives. 6 This public discourse both facilitates patient access to support and information 5 and provides rich data for analyzing societal perceptions and attitudes toward breast cancer.7–10
Existing literature on breast cancer discourse has predominantly focused on Western social media platforms such as Twitter, Facebook, and Reddit.10–13 In contrast, studies based on Eastern contexts—particularly Chinese social media—remain limited,6,7,14 even though China ranks first worldwide in the absolute number of breast cancer cases. Notably, Basch et al. 7 explored breast cancer-related content on TikTok using a cross-sectional content analysis of the top 100 videos bearing the hashtag #breastcancer. Drawing on manual categorization, they examined video characteristics and content across several thematic domains, including support, coping mechanisms, body image, surgery, cancer education, and awareness raising. The results indicated that patients or their family members produced most of the videos, with support and coping emerging as the most prevalent themes. To examine how breast cancer discourse on Chinese social media differs according to poster-patient relationships, Zhang et al. 6 analyzed 10,322 Weibo posts using fine-tuned Chinese BERT models to classify relationships into five categories: post_user, family_members, friends_relatives, acquaintances, and heard_relation. Sentiment analysis using the LIWC lexicon revealed significantly higher negative emotions and anger in the “no_patient” category, while the “family_members” category showed greater sadness and anxiety. Topic modeling further indicated that the “no_patient” group focused on fears and anger toward cancer, whereas the “family_members” group emphasized cancer treatment. These findings underscore that emotional tone and discussion topics vary meaningfully according to the poster’s relationship to the patient. Zhao et al. 14 analyzed the themes sharing discourse among 40 Chinese breast cancer patients on Weibo platform, and examined thematic variation across disease stages. The findings indicated that patients’ sharing discourse concentrated on disease information, personal emotions, and social relationships. Across the four stages of breast cancer, orientation toward personal emotions and social relations increased progressively, whereas orientation toward disease information declined. The study indicates that digital sharing discourse can capture patients’ cancer-related attitudes and beliefs, while also serving as empathy-mediated support mechanisms for cancer patients throughout the disease trajectory.
Regarding methodologies for analyzing social media discourse on breast cancer, text mining methods such as topic modeling and sentiment analysis, informed by data obtained through web crawling, function as an effective big data approach to the portrayal of digital communication about breast cancer. Topic modeling can identify latent themes or topics present within a corpus of text. 15 By employing algorithms such as Latent Dirichlet Allocation (LDA), researchers can uncover various themes related to breast cancer that users discuss, ranging from treatment options to personal experiences and societal attitudes.8,10,16 Sentiment analysis can identify and extract the emotions or subjective information expressed in the social media text, revealing individuals’ feelings toward breast cancer screening, treatment experiences, and healthcare perceptions.13,14,17,18 The combination of topic modeling and sentiment analysis can provide a comprehensive approach to examining public attitudes and emotional tendency regarding particular topics, products, or events relating to breast cancer.
Taken together, while existing social media discourse studies on breast cancer have predominantly focused on Western platforms, they remain limited in two respects: first, they lack big-data investigation of the discussion content, emotional expression of the Chinese public, and the educational effects of Chinese social media; second, they remain at describing themes and emotions without conducting an in-depth exploration of the social meanings behind digital discourse using social cognitive perspectives, thereby lacking interdisciplinary integration of text mining and discourse analysis. In this regard, the present study employed sentiment analysis and topic modeling to examine Chinese public discourse on breast cancer using data collected from Sina Weibo. Sina Weibo is an effective platform to delve information related to breast cancer. It is one of the most widely used and popular social media platforms in China, often referred to as the “Chinese version of Twitter”. 19 Among social media platforms in China, Weibo ranks in the top three globally by total number of users. 20 The platform is not restricted to Chinese users or subject to regional limitations; international users are able to register and access it globally. It supports multi-language settings and diverse registration methods, thereby overcoming language and geographic barriers. Information dissemination on the Weibo platform is predominantly text-based, and user-generated content exhibits a high density of textual information, which facilitates the construction of a large-scale, real-world corpus. Posts published on Weibo frequently reflect current social concerns, sentiments, and policy issues in China. 21 Thus, substantial discussions and topics related to breast cancer can be found on Weibo. Due to its large user base and diverse content, Weibo data can serve as a valuable corpus for studying communication about breast cancer from a Chinese perspective. 6
In the present study, sentiment analysis was used to identify public emotional tendencies, while topic modeling revealed core discussion themes and their variation across different sentiment categories. Furthermore, drawing on Critical Discourse Analysis (CDA), which investigates how power dynamics, ideologies, and social structures influence language use and shape public perceptions,
22
this study examined how social media discourse on breast cancer shape social cognition within Chinese socio-cultural context. By combining text mining with CDA, this study contributes to a deeper understanding of how social factors influence individual experiences and perceptions of breast cancer, ultimately can shape public health awareness and response in digital context. The study aimed to answer the following questions: (1) What are the sentiments toward breast cancer among the Chinese public? (2) What are the topics of public discourse toward breast cancer among the Chinese public under each sentiment classification? (3) Combined with the identified sentiments and topics, what social cognition do public discourses reflect in the context of Chinese society and culture?
2. Theoretical framework
Critical Discourse Analysis (CDA) constitutes a theoretical and methodological framework that conceptualizes discourse as an integral form of social practice, actively constructing societal understandings of reality.23,24 In the context of public health, this approach is particularly salient for examining how media representations, including social media, shape prevailing conceptions of “health” and “disease”. 25 By systematically analyzing empirical communication data—from mass media to social media posts—CDA seeks to uncover the ideological underpinnings and social consequences embedded within language use. 26
Among the major CDA approaches, Fairclough’s model22,23,27 has proven effective in analyzing health discourse on social media.
28
It has been widely applied across social sciences, linguistics, communication studies and related fields,
29
helping to reveal the complex relationship between discourse and social reality. Compared with van Dijk’s “socio-cognitive model,” which emphasizes cognitive representation, and Halliday’s systemic functional linguistics approach, which focuses on linguistic systems,
30
Fairclough’s model explicitly conceptualizes discourse as a trinity of text, discursive practice, and social practice, thereby enabling a systematic linkage between micro-level linguistic features and macro-level social structures. This study applies CDA to public discourse on breast cancer from Weibo, drawing on Fairclough’s Three-Dimensional Framework,22,23,27 which conceptualizes discourse as a tripartite structure encompassing text, discursive practice, and social practice (see Figure 1). The framework incorporates “linguistic description of the language text, interpretation of the relationship between the (productive and interpretative) discursive processes and the text, and explanation of the relationship between the discursive processes and the social processes”.
27
This model treats discourse not merely as a linguistic product but as a dynamic social process embedded within power relations and ideological contexts.
31
Three-dimensional framework of discourse.
22
To be specific, text analysis is the stage of description, involving systematic linguistic analysis of vocabulary, grammar, rhetorical devices, coherence, and textual structure. 22 Discursive practice is the stage of interpretation regarding the relationship between text and interaction. This dimension places discourse in the context of social practices and social relations, emphasizing how discourse is used, by whom, and for what purpose. It examines processes of text production, distribution, and interpretation, focusing on how social cognition shapes these processes.22,27 Finally, the social practice analysis is the stage of explanation. This dimension situates discourse within broader ideological frameworks, analyzing how social structures both constrain discursive practices and are reconstituted through them.22,31 The three dimensions are fundamentally interconnected, operating dialectically to reveal how language reflects and constructs social reality.
Employing Fairclough’s three-dimensional framework, this study analyzed breast cancer discourse on Weibo through text analysis (sentiment, and topics), discursive practice analysis (Weibo dissemination), and social practice analysis (power structures), revealing how breast cancer discourse shapes and is shaped by public cognition. In textual dimension, linguistic features directly reflect the public’s core perceptions of breast cancer. In discursive practice dimension, discourse production (e.g., media coverage, interactions) and dissemination pathways (e.g., retweet chains on Weibo) can reveal the generative processes of these perceptions. In social practice dimension, analyzing the interplay between discourse and power structures, institutions, and culture can explain how specific cognitive frameworks serve power dynamics. The three-dimensional analysis can reveal the dialectical relationship between how discourse constructs cognition and how cognition reciprocally shapes discourse on social media, offering a comprehensive lens to understand breast cancer discourse and user purposes on Weibo.
3. Materials and methods
3.1. Research design
This is a mixed-methods study, which collected publicly available Weibo posts on the Sina Weibo platform using keyword-based web crawling techniques. The crawler employed Weibo’s advanced search function with “breast cancer” as the search term, retrieving data for the period spanning from January 18, 2022, to June 30, 2024. The data was collected during September 21 to 24, 2024 in Chonqing, China. The text data underwent cleaning and preprocessing. A lexicon-based approach was employed for sentiment analysis, followed by Latent Dirichlet Allocation (LDA) for topic modeling. Based on the sentiment classification (negative, neutral, positive), separate topic modeling was conducted for posts under each sentiment category. Finally, discourse analysis was performed based on the results of the data analysis. Figure 2 illustrates the detailed methodological flow. Methodological flow chart.
3.2. Data collection
This study employed web scraping techniques to collect publicly available posts from Weibo. The program simulated user access to the official Weibo website and conducted searches to gather relevant data. A web scraper was constructed using the requests library and the Scrapy framework in Python, specifying the keyword “breast cancer” for advanced search and setting the time range from January 18, 2022, to June 30, 2024. The Breast Cancer Screening Work Programme was issued by the General Office of the National Health Commission of China on January 18, 2022. This plan aimed to standardize breast cancer screening practices, improve screening quality and efficiency, and promote early diagnosis and treatment of breast cancer. Following this time point, discussions regarding breast cancer on Weibo were likely to increase significantly, facilitating an understanding of public reactions and discussions related to this new policy. It also allows for a better comprehension of public attitudes, behavioral changes, and information needs regarding breast cancer. Thus, data collection was conducted for the period starting from January 18, 2022, and extending to June 30, 2024. A total of 185,486 Weibo posts were collected. After duplicate removal, 134,588 original posts (word count: 30, 417, 201) were retained for data analysis. Additionally, user information was retrieved, including user ID, username, gender, posting time, follower number, following number, number of shares, number of comments, and number of likes, for subsequent analysis. The user demographics of the platform covered most provinces and regions across mainland China, including municipalities and autonomous regions. This confirms the platform’s extensive geographic reach and diverse regional user participation. The breast cancer discourse on Weibo represents a diverse, multi-layered user base, including health-interested general users, breast cancer patients, and caregivers (e.g., family and friends). In addition to this core group, a substantial proportion of discourse was generated by medical professionals and consultative users, such as doctors, hospitals, and media outlets.
3.3. Ethics consideration
This study utilized exclusively publicly available data from which all personal identifiers were absent; therefore, it was exempt from institutional ethics review. To protect anonymity, all posts included in analysis had been paraphrased to prevent traceability to original users. No identifying information—including usernames, IDs, or pictures—was used in the analysis or supplementary materials. The content was accessed in compliance with Weibo platform’s terms of service.
3.4. Data preprocessing
The preprocessing of the text data included text cleaning, Chinese word segmentation and stopword removal. (1) Text cleaning: The Weibo data, characterized by complex linguistic structures and unstructured elements like emojis and special characters, required normalization to ensure analytical accuracy. The preprocessing involved removing user mentions, URLs, punctuation, numbers, and extraneous whitespace. Using Python’s pandas library for data management and the Lib/re.py module for regular expressions, we systematically cleaned the text to create a uniform corpus for subsequent analysis. (2) Chinese Word Segmentation: Chinese text requires word segmentation for computational analysis due to its character-based, unspaced nature. This study employed the Jieba toolkit, a standard Python library in Chinese NLP, to segment text into discrete lexical units. Using its precise mode, which prioritizes segmentation accuracy, we applied jieba. cut (text) to divide sentences into meaningful units or tokens accurately. Taking the sentence “愿这个世界上所有女性不会患有乳腺癌” (Tanslated: May all the women in the world be free from breast cancer) as an example, the sentence is segmented into tokens: “愿” (may), “这个” (this), “世界” (world), “上” (in), “所有” (all), “女性” (women), “不会” (won’t), “患有” (have), “乳腺癌”*(breast cancer). Tokenization concerted raw text into analyzable units and facilitated the accurate extraction of the key terms for topic modeling. (3) Stopword Removal: Following segmentation, stopwords were removed using a comprehensive list compiled from the Harbin Institute of Technology, Baidu, and Sichuan University stopword dictionaries. The filtering process also excluded tokens shorter than two characters. This was implemented programmatically via list comprehension to retain only meaningful, multi-character tokens, thereby refining the dataset for subsequent analysis.
The data processing involved three sequential categorization stages. First, sentiment analysis was conducted on the complete dataset. Second, the data were disaggregated by month to observe temporal sentiment distribution trends. Third, following sentiment classification into negative, neutral, and positive categories, separate topic modeling analyses were performed for each sentiment category.
3.5. Data analysis
Sentiment analysis was performed to classify breast cancer-related posts as positive, neutral, or negative, and to track the monthly evolutionary trends of these sentiments. The analysis utilized the Chinese DLUT-Emotion Ontology as a reference emotion dictionary to identify emotional tendencies in Weibo posts. This emotion dictionary was developed by Dalian University of Technology’s Information Retrieval Laboratory, comprising a lexicon of over 27,000 emotion words.32,33 Statistical analysis of the dictionary’s lexicon across sentiment categories yielded proportions of 40.7% positive, 39.8% negative, and 19.5% neutral words, indicating a relatively balanced distribution between positive and negative terms.
Specifically, sentiment analysis for breast cancer posts was conducted using a Python program. First, the Chinese emotion dictionary was loaded using the pandas library, containing sentiment polarity categorized as positive (1), negative (2), or neutral (0). 33 Each Weibo post was tokenized into lexical units using the Jieba segmentation tool. A function, calculate_sentiment_score, was then used to compute each post’s sentiment score: for each word found in the dictionary, the score was incremented by 1 if the polarity of the word was positive, decremented by 1 if negative, and left unchanged if neutral. After processing all words in a post, the final sentiment score for each post was classified as positive for scores greater than 0, negative for scores less than 0, and neutral for scores equal to 0.
Next, sentiment-specific topic modeling was conducted on negative, neutral, and positive post subsets to examine thematic variations by emotional tone. The study employed Latent Dirichlet Allocation (LDA 34 ), an unsupervised generative model widely used in text mining, to uncover latent thematic structures in Weibo posts. LDA represents documents as mixtures of topics, each characterized by keyword distribution, enabling the extraction of representative features and identification of coherent themes in large-scale textual data. 35
The LDA topic modeling procedure comprised several steps. First, preprocessed text data were read from an Excel file. A dictionary (mapping words to unique IDs) and a corpus (bag-of-words representation of each post) were then constructed. The number of topics was set to five, a choice that avoided overly broad or excessively granular thematic extraction, and the number of passes was set to 15 to balance model accuracy with computational efficiency. After model training, keywords for each topic were extracted using the lda_model.print_topics () method. For each document, the model generated a topic probability distribution, from which the topic with the highest probability was selected as the document’s primary topic; the number of posts assigned to each topic was subsequently tallied. To facilitate intuitive interpretation of thematic patterns, an interactive visualization was generated using the pyLDAvis library. In this visualization, topic proportions in the corpus were represented by circle size, and the relevance metric λ—which controls term ordering within a topic—was set to 1 to prioritize core topic keywords. Finally, each topic was manually summarized and labeled based on combined keyword inspection and visualization results.
Integrating sentiment analysis and topic modeling, this study examined the relationship between public attitudes and themes in breast cancer discourse, while also using critical discourse analysis to investigate how discourse structure and contextual factors on Weibo reflect social cognition.
4. Results
4.1. Sentiments toward breast cancer
Sentiment analysis was performed on a corpus of 134,588 Weibo posts concerning breast cancer. The analysis revealed that 54,551 posts (40.5%) were positive, 46,441 posts (34.5%) were neutral, and 33,596 posts (25.0%) were negative, as detailed in Figure 3. The distribution demonstrates a generally positive discourse surrounding breast cancer on Weibo. The predominance of positive sentiment likely reflects public support for affected individuals and an endorsement of breast cancer prevention and treatment initiatives. The substantial proportion of neutral posts suggests that a significant volume of content is dedicated to informational or educational purposes, rather than personal emotional expression. Conversely, the persistent negative sentiment (25.0%) articulates the psychological distress, fear, and lived challenges associated with the disease. Overall sentiment distribution.
Figure 4 illustrates the monthly trends in sentiments, with separate lines tracking the volume of positive, neutral, and negative posts over the study period. It is evident that the number of posts of all kinds of sentiments rose sharply in March 2022, August 2023, October 2023, and November 2023. In November 2023, posts for neutral and positive sentiments continued to rise and peaked, while posts for negative sentiments begun to plummet. Monthly sentiment trend of Weibo posts.
Overall, positive sentiment consistently dominated the discourse, followed by neutral and negative expressions. Sentiment fluctuations correlated with holidays, policy updates, public health campaigns, and trending news. Notably, during Breast Cancer Awareness Month (October-November 2023), positive and neutral posts rose significantly while negative sentiment sharply declined, indicating the efficacy of focused health campaigns. Furthermore, increased public engagement compared to October 2022 suggests that sustained health education has successfully heightened awareness and promoted preventive health behaviors.
4.2. Topics of breast cancer discourse
To examine sentiment-specific discourse, we stratified the Weibo posts by sentiment (negative, neutral, positive) and performed separate LDA topic modeling on each category. Furthermore, representative posts were selected for critical discourse analysis using Fairclough’s three-dimensional framework—analyzing text, discursive practice, and social practice—to elucidate the social and ideological underpinnings of the breast cancer discourse.
4.2.1. Topics in negative sentiment category
The topics of negative posts and high-frequency keywords.
Topic 1 reveals public concerns regarding the unknown risks and potential causes of breast cancer. Such uncertainty may trigger fear and anxiety. Topic 2 reflects the challenges patients may face during treatment, such as the uncertainty of treatment effectiveness, long treatment durations, economic pressure, or side effects, all of which clearly exacerbates negative emotions. Topic 3 presents the emotional fluctuations and physical discomforts that breast cancer patients may experience, such as pain, changes in body image, and anxiety, which reflects the public’s worries and feelings of helplessness regarding the disease. In topic 4, personal experiences shared by patients or their families are an important component of the negative discussion. These narratives are often emotionally charged, conveying the difficulties and emotions individuals face when dealing with the disease. In topic 5, the diagnosis of breast cancer has a profound emotional impact on both patients and their family members, leading to societal fear of breast cancer and health anxiety.
Here is an example related to the topics from the negative posts: “Fourteen years ago, I was diagnosed with breast cancer. Surgery, nearly ten months of chemotherapy and radiotherapy, and five years of endocrine treatment, the dual imbalance in my body and mind made me feel that those were the darkest years of my life. Over the past year, facing significant choices at home, this feeling of being directionless, worried, confused, and helpless has made me anxious, painful, depressed, and irritable. Once again, I thought of the term ‘darkest years’, but now I timidly hesitate to use the word ‘darkest’. It feels like there is no end, only more challenges. Life is a cruel battlefield, and I can only get through one ordeal after another, hoping that when I look back, I could have been calm and peaceful.”
4.2.1.1. Text analysis
The text illustrates the personal experiences of a patient, by employing emotionally rich vocabulary such as “darkest years,” “anxiety,” “pain”, “depression”, and “irritability”. These words not only depict the author’s inner struggles but also reflect a sense of helplessness regarding breast cancer. The use of metaphors (e.g., “life is a cruel battlefield”) vividly illustrates the difficulties and challenges of breast cancer. The text presents the complete narrative structure, moving from past experiences (breast cancer treatment) to present confusion and anxiety, and finally to a hopeful outlook for the future (“to be calm and peaceful”).
4.2.1.2. Discourse practice analysis
The author shares personal experiences to express reflections on life and the struggles faced, possibly seeking resonance from readers or looking for understanding and support. Today, even with advances in medical technology, breast cancer still has a significant impact on the lives and psychology of patients and the treatment process remains a huge challenge for patients and their families. This post received some likes and comments, which meant that readers might feel sympathy and understanding for the author’s experiences, or be prompted to reflect on their own experiences, creating an emotional connection.
4.2.1.3. Social practice analysis
By narrating personal experiences, the author partially reshapes her identity, transitioning from an individual with illness to one who continuously strives for peace and balance in life. The text potentially reveals the social pressures and cultural expectations individuals face when confronted with significant life challenges. The author’s experiences and feelings may intertwine with societal views on women’s health, family responsibilities, and personal value, reflecting existing gender and social role expectations in these areas.
4.2.2. Topics in neutral sentiment category
The topics of neutral posts and high-frequency keywords.
Topic 1 indicates that the public demonstrates substantial interest in foundational knowledge and preventive measures related to breast health, suggesting that the popularization of health education may have played a role. Topic 2 reflects increased public attention to early detection and prevention, which could also be linked to promotional efforts by governments, medical institutions, or nonprofit organizations. Topic 3 highlights that breast cancer impacts not only individuals but also their familial and social support systems. The sharing of individual illness experiences on social platforms is conducive to mutual support and access to experience information. Topic 4 focuses on treatment methods and medical research, including discussions on scientific advancements, targeted drugs, and novel therapies, may aid in decision-making, support patients, or enhance public confidence in breast cancer treatment. Topic 5 reveals that public discourse of breast cancer not only involves disease prevention and treatment but also encompasses body image and aesthetic concerns, potentially reflecting societal attention to the relationship between breast appearance and health. Overall, neutral sentiment posts cover a wide range of topics, including breast health, prevention, patient support, medical research, and cosmetic surgery. These neutral discussions likely serve as critical roles for disseminating objective information and elevating public awareness of breast health.
Here is an example posts related to the neutral posts: “[Breast cancer prevention, adjust lifestyle and emphasize disease screening] 1. For the non-high-risk breast cancer population, screening is not recommended for those aged 20-39. For individuals aged 40-70, annual mammography is recommended, and for those with dense breast tissue (where mammography indicates breast tissue type as C or D), it is advised to combine this with ultrasound examination. For individuals aged 70 and above, mammography should be conducted every 1-2 years. 2. In contrast, for the high-risk breast cancer population, it is recommended to start breast screening at an earlier age (<40 years), with annual mammography, ultrasound every 6-12 months, and, if necessary, annual breast MRI with contrast.”
4.2.2.1. Text analysis
The post used more formal and concise language to express screening recommendations for breast cancer prevention. The language indicates an objective and authoritative feeling. The use of technical terms and the clear expression of time highlight the medical nature and professionalism, aiming to provide clear medical screening guidance.
4.2.2.2. Discourse practice analysis
The Weibo post, released by Beijing Fangshan Health Education, aims to provide standardized information on breast cancer screening to the public, so its authority is guaranteed. The post is aimed at the general public, especially those who don’t know much about breast cancer screening. Through concise and clear screening recommendations, the post effectively conveys the message of health management. The neutral and professional language strengthens this educational and guiding function, avoids over-emotional description, and makes the public treat breast cancer prevention more rationally. Such advice is usually issued by the medical community and health organizations. The dissemination of such information reflects the dominance of the medical institutions in health management and disease screening.
4.2.2.3. Social practice analysis
The popularization of breast cancer screening is not only a part of public health policy, but also a manifestation of the promotion of healthy lifestyles and disease prevention in modern society. The language of the post presents a sense of social responsibility in health education, reflecting the modern society’s emphasis on health management. By popularizing the relevant knowledge of breast cancer screening, the society strives to reduce the incidence of the disease and advocate the health concept of early detection and early treatment. This discourse reflects the guiding and normative role of health management institutions and medical experts, serving to direct and shape public health behaviors and attitudes.
4.2.3. Topics in positive sentiment category
The topics of positive posts and high-frequency keywords.
Topic 1 focuses on the positive aspects of the patient experience that individuals with breast cancer, as well as their families, find strength and encouragement in sharing their experiences and receiving emotional or practical support. This topic underscores the importance of social and emotional support in breast cancer journey. Topic 2 suggests that a significant portion of the public posts is focused on proactive measures for breast cancer prevention, including maintaining a healthy lifestyle, nutrition, and exercise. This reflects the growing awareness and commitment to prevention as a means to reduce breast cancer risk. Topic 3 highlights the large number of posts promoting knowledge about breast cancer screening, as well as the public’s recognition of the importance of early detection, which contributes to early diagnosis and improved treatment outcomes. Topic 4 reflects the public’s belief in the ongoing progress of medical research and treatment options. The optimism surrounding treatment advancements suggests a sense of hope and confidence in the future of breast cancer care. Despite the challenges posed by breast cancer, topic 5 puts emphasis on hope and resilience indicating that many individuals draw strength from both their own experiences and the collective efforts of others to overcome these challenges. The focus on hope underscores the resilience of the community and the importance of continued support. Positive discussions reflect a focus on hope, empowerment, and active engagement in prevention and treatment. These topics highlight an overall optimistic outlook, with a particular emphasis on the importance of support systems, lifestyle choices, and medical advancements.
Here is an example related to the above topics from the positive posts: “Familiar paths, but a different mindset! I remember when I was first diagnosed, I hurriedly rushed from Xi’an to Beijing in a panic. Time has flown by so quickly; it’s hard to believe it's already been six months. I’m on my last treatment now, and as fate would have it, I had to travel from Xi’an to Beijing again for a biopsy. Along the way, I also dealt with a little hiccup of having my train tickets canceled and unable to change them. “But luckily, I spent some money to resolve it through a scalper. One sister said that spending money to avoid disaster sounds very pleasant,” I don’t know why, but despite the difficulties this time, I feel it’s a good sign. The journey feels similar to the first one, yet it’s the last one. I get the feeling that things will work out smoothly, leading to a positive outcome because it feels like a TV drama or a story with a beginning and an end, almost perfect. I hope this final hurdle goes well, and I wish for the heavens to bless me with a result that isn’t too cruel for all the hardships I’ve faced over these years. I will keep pushing myself too! Sisters, please send me your blessings! #Breast Cancer Super Topic”
4.2.3.1. Text analysis
The post is written in relatively casual, natural language, and the tone is cordial and sincere, with a hint of humor. For example, the statement, “spending money to avoid disaster”, “work out smoothly”, and “leading to a positive outcome” reverses the conventional negative emotion and makes the text seem light and positive. Through the vivid narration of her own experience of disease, the author reflects the positive attitude in the course of disease treatment.
4.2.3.2. Discourse practice analysis
The author of the post is apparently the patient herself. She shared her personal experiences through Weibo, not only to vent her emotions, but also to obtain support and comfort through social media platforms. On the platform of Weibo, the reader group usually has a high interaction by the behavior of commenting, reposting and liking. This interaction not only resonates with the author’s emotions, but also may encourage and support the process of fighting the disease.
4.2.3.3. Social practice analysis
As the attention to women’s health issues, such as breast cancer, has significantly increased, patients’ self-expression through social platforms has become not only a way to cope with illness but also an exploration of the social support system. In Chinese culture, a diagnosis of breast cancer or other cancers often carries a certain social stigma, which causes many people to feel ashamed or fearful about revealing their illness. However, in an increasingly open society, with the acceleration of information flow and the formation of online communities, patients can find broader support online, share their experiences, and help break through these cultural taboos. Additionally, the mention of the saying “spending money to avoid disaster” reflects the common Chinese concepts of “feng shui” (geomancy) and “fate”, which shape the optimistic tone of the emotional expression in this post.
5. Discussion
5.1. Sentiments and topics in text dimension
5.1.1. Negative sentiments and associated topics
In negative posts, the topic “risks and influencing factors” reflects public anxiety about breast cancer etiology. Such posts were primarily disseminated by news media and health organizations, highlighting their authoritative role in defining risk knowledge—a manifestation of Foucault’s knowledge-power dynamic. While this outreach raises awareness, it may amplify public anxiety, particularly when risk factors are oversimplified. This finding aligns with Modave et al., 10 who reported that laypeople discuss “risk” less frequently than institutions. To mitigate unintended distress, public health communication should offer evidence-based information about breast cancer, as credible social media support has been shown to reduce anxiety and empower individuals. 11
The topic “challenges in treatment and patient experiences” underscores the multifaceted struggles of breast cancer treatment, encompassing difficult medical decisions, treatment side effects, and the burden of ongoing care. These discussions reflect a sense of isolation, particularly regarding uncertain outcomes and barriers to care, consistent with research on the financial and emotional distress experienced by patients and caregivers.6,13
The topic “patients’ negative emotions and physical symptoms” highlights the interplay between psychological distress and physical suffering in breast cancer. 6 Notably, a substantial portion of this discourse reflects anger from non-patients using breast cancer as an outlet for general life stress, echoing Zhang et al.'s observation that social media serves as a platform for catharsis. 6 This suggests that public discourse often conflates generalized emotional venting with patient-centered experiences.
The topic “personal experiences” reveals the nuanced emotional landscape of breast cancer, contrasting with prior literature. While previous studies reported these narratives as primarily positive through stories of support and awareness 12 or as negative due to policy and financial stressors, 13 the present study found that personal experiences reflected negative sentiment in weibo discourse, conveying fear and grief related to diagnosis and treatment. This diverged from Clark et al., 13 which attributed distress to insurance, suggests that the emotional tenor of personal narratives is highly context-dependent.
The topic “emotional impact and social responses associated with diagnosis” captures the public’s affective reactions to breast cancer diagnoses, ranging from personal shock upon a family member’s diagnosis to collective grief following celebrity cases (e.g., Coco Lee). These posts, characterized by fear and anger, align with Zhang et al., 6 who identified “diagnosis” as an emotionally charged theme. Our findings extend those of Modave et al., 10 who noted public attention to “diagnosis” without characterizing its emotional dimension.
5.1.2. Neutral sentiments and associated topics
Neutral discourse primarily serves an informational and educational function. The first topic “breast health and disease prevention” reflects public demand for authoritative knowledge on breast cancer. Unlike findings from Thackeray et al. 36 on Twitter, where preventive behaviors were seldom promoted, discussions on Weibo advocate scientific preventive measures against breast cancer. The topic “cancer screening and prevention awareness” highlights early detection and health management. The neutral tone of these posts suggests an objective, informational style concerning concrete preventive actions. The topic “patient experiences and family social support” underscores the essential role of emotional support from both family and the broader community. These insights are critical for healthcare providers and policymakers, highlighting the well-established importance of psychosocial support in the recovery process. 37 The topic “clinical treatment and research advancements” reflects public interest in disseminating scientific updates such as targeted drugs. This contrasts with prior studies that identified “treatment” as a topic but not its neutral sentiment.6,10,13 The neutral tone may be strategic, as medical research dissemination often flows through industry-funded platforms (e.g., clinical trials, sponsored conferences), suggesting ostensibly objective information may still serve pharmaceutical interests. The topic “breast health and cosmetic surgery” reveals the medicalization and commodification of the female body in the context of breast cancer. This creates a power imbalance positioning women as “good consumers” of medical technology. 24 Despite its neutral tone on Weibo, this topic echoes international concerns, 16 confirming body image as a universal patient concern shaped by commercial and clinical narratives.
5.1.3. Positive sentiments and associated topics
In positive posts, the topic “patient experiences and support systems” underscores emotional and institutional support in breast cancer recovery, by story sharing and professional advice on Weibo platform. This aligns with research indicating that online communities help patients seek connection and mutual support.38,39 The establishment of such robust support networks is critical, as it can significantly enhance psychological resilience and improve treatment outcomes. 40
The topic “prevention and healthy lifestyle choices” reflects a public shift toward proactive health awareness, with medical and media discussions that may encourage preventative behaviors and reduce cancer incidence. Our findings contrast with prior research, such as Basch et al., 7 which identified but did not find prevention to be a dominant theme on TikTok, highlighting a potential platform-specific difference in public health discourse.
The topic “screening and prevention of the two cancers” reveals that health promotion in breast cancer context is perceived as empowering rather than anxiety-inducing. Our finding contrasts with prior research,10,41 such as Modave et al., 10 who associated screening with negative sentiment due to regional access barriers. The positive sentiment regarding screening and prevention in our study may thus reflect successful and accessible public health initiatives in Chinese society.
The topic “treatment advancements and clinical research” was characterized by optimism about new therapies and clinical trials. This public engagement with scientific progress on social media aligns with findings from Basch et al., 7 who identified treatment-related content as a form of advocacy on TikTok.
The topic “challenges and hope renewing” underscores resilience in adversity through social support. This aligns with Zhang et al., 6 who identified hope as a central positive theme. Such discourse serves a critical psychosocial function: sharing successful experiences provides vicarious hope, reduces loneliness, 39 and helps patients reframe their life values, while simultaneously raising broader societal awareness about breast cancer.
5.2. Discursive practice underlying breast cancer discourse
Following Fairclough’s three-dimensional framework, we examine intertextuality and discourse distribution at the level of discursive practice, based on the analysis of sentiment and thematic patterns.
First, medical/authoritative discourse predominantly appears in neutral topics (e.g., “clinical treatment and research advancements”) and positive topics related to screening. Users generally adopted strategies of directly reposting or citing medical texts, maintaining a neutral or cautiously optimistic emotional stance. This intertextual operation suggests that authoritative medical discourse plays a role in emotional distribution on Weibo, as users appropriate scientific discourse to legitimize their own speech, which once again corroborates the findings of McGannon et al. 42 Second, themes such as prevention, patient experiences, and support systems exhibit marked intertextual features—users extensively reproduce the “anti-cancer” templates circulated in public service announcements and on social media. This type of discourse forms explicit intertextual relationships with promotional texts from campaigns such as “Pink Ribbon” and “Breast Cancer Awareness Month”. The findings showed that the positive impacts of breast cancer awareness campaigns were also present in the Chinese social media context, suggesting that this phenomenon extends beyond Western societies.12,36 It is worth noting that this patterned intertextual practice leads to a high degree of homogenization of positive emotions, namely, the “hope” expressed by users rarely refers to concrete life expectations but instead becomes a decontextualized, ritualized form of speech.
At the level of distribution of breast cancer discourse, high-circulation posts in this study exhibit explicit intertextuality with public service slogans, celebrity stories, and breast cancer prevention measures—a finding that resonates with previous studies12,36—indicating that formulaic inspirational discourse holds a great communicative advantage on the Weibo platform. At the same time, Weibo’s communication infrastructure exerts a substantial influence on discursive practices. For example, posts bearing official or public welfare hashtags such as #BreastCancer# and #PinkRibbon# occurred in high frequency. This suggests that the platform’s hashtag mechanism favors and amplifies positive discourse that aligns with the public welfare narrative framework, thereby rendering the platform an implicit filter in discourse reproduction.
The discursive construction of breast cancer involves a hierarchy of interest groups (e.g., media, government, medical institutions, patient groups, and family members). Government and medical institutions hold dominant power, shaping the narratives through policy and a scientific discourse that often masks commercial influences.42,43 This “knowledge-power” dynamic positions their framings as objective facts. In contrast, patient and family narratives center on personal experience, emotional response, and the embodied struggle of the disease,44,45 while media amplifies messages through reporting and celebrity influence. 46 While previous studies noted the prominence of lay voices on social media, 47 the present study demonstrates that the overarching discursive framework—defining issues from screening to treatment—remains largely controlled by institutional power, thereby marginalizing patient-centered perspectives.
5.3. Social cognition reflected in breast cancer discourse
The social significance of breast cancer is constructed through the power relations of different interest groups within Weibo discourse. The topic “breast health and cosmetic surgery” and keywords such as “mom”, “children”, and “women” may frame breast cancer as a “women’s issue”. Breast cancer, as a “women’s issue”, is associated with gender roles, female identity, and societal expectations. The emphasis on “breast health and cosmetic surgery” highlights the concern with body image after mastectomy, the appearance of women, and reinforces the connection between breast cancer and femininity. This finding is in line with the research of Gibson et al. 43 and McGannon et al., 42 which revealed the association between breast cancer and femininity, and how the media has tied the disease to female identity through discourse. In the present study, the use of relational terms such as “mom” and “children” frames women within their social roles, while a focus on cancer “risks”, “prevention”, and “screening” emphasizes both risk literacy and early detection. Such discourse structure, although aligned with public health objectives, can generate anxiety by constructing breast cancer as a pervasive health crisis. This is consistent with the findings of McGannon et al.’s study, 42 which found that the news media created public anxiety through coverage of controversial screening guidelines while shaping breast cancer as a public health crisis requiring “universal attention”. The topics concerning “prevention, screening, treatment, medical advancements, and healthy lifestyles” frames breast cancer as a manageable, long-term condition. However, this framing simultaneously individualizes the disease by shifting responsibility onto patients through an emphasis on lifestyle and self-management, a finding consistent with Gibson et al. 43 and McGannon et al. 42 Discussions concerning breast cancer treatment and prevention in our study reflect institutional communications by government, health, and media organizations, which predominantly exhibit positive or neutral sentiment. This strategic framing aims to promote public health literacy, encourage screening, and highlight therapeutic advancements. Such positive messaging aligns with social media algorithms that favor and amplify uplifting content,48,49 thereby enhancing disease awareness and fostering therapeutic optimism among patients and families. Furthermore, this study corroborates with the literature that celebrity advocacy, known to stimulate public engagement in knowledge-seeking activity, 50 serves as a potent catalyst for discussion. This underscores the efficacy of leveraging celebrity influence in Chinese society as a conduit for disseminating positive health education on breast cancer.
Our Weibo discourse was characterized by posts offering or soliciting support for those affected by breast cancer. This aligns with established research affirming the benefits of social support for patients and families,51–53 a function facilitated by digital platforms. 54 Our findings reaffirm that social media such as Weibo enables effective health advocacy through discourse strategies—such as positive framing, evidence-based data, and celebrity influence—while simultaneously fostering a community for exchanging social support through shared personal narratives of lived experience.
The prevalence of negative discourse in our dataset highlights the critical yet often overlooked issue of psychological distress among those affected by breast cancer, including significant life changes, worries about recurrence and survival, painful treatments, and altered body image—factors that can lead to anxiety, fear, and depression.55,56 The psychological distress associated with breast cancer in China is deeply embedded within a specific socio-cultural fabric. 57 A woman’s identity is intertwined with her role as a family caregiver; a breast cancer diagnosis that impairs this capacity can therefore precipitate a crisis of self-worth, triggering significant anxiety. 58 Furthermore, patients’ pervasive fears of recurrence are rooted in a sense of powerlessness, exacerbated by inadequate health literacy and a traditional cultural reluctance to openly discuss illness and mortality. 57 This anxiety is compounded by the threat to bodily integrity. As breasts and hair are culturally significant markers of femininity, treatments such as mastectomy and chemotherapy-induced alopecia are not merely medical procedures but profound assaults on identity.58,59 The resulting changes in body image can evoke feelings of shame and loss, directly impacting psychological well-being and complicating the reconstruction of self-identity post-diagnosis. 60
5.4. Practical implications for enhancing public awareness of breast cancer
In practice, various stakeholders can take the effective measures to enhance public awareness of breast cancer and to foster a more supportive environment for those affected by the disease.
First, media organizations can orchestrate a multi-faceted strategy to enhance public engagement and awareness of breast cancer. This involves actively disseminating scientifically accurate information and promoting positive discourse through optimistic, emotionally resonant content. For rural areas or populations with lower education levels, a three-part format consisting of short videos, dialect narration, and simplified graphics should be adopted. Furthermore, media can provide crucial emotional support by facilitating the sharing of survivor and patient narratives, while simultaneously strengthening mental health advocacy by publishing content on coping with anxiety and depression. For instance, collaborations with top-tier hospital breast departments to produce popular science short videos and articles, alongside establishing official accounts dedicated to breast cancer education, can effectively centralize and disseminate information on screening, treatment, and healthcare security. Regarding breast cancer myths (such as “breast massage can cure” and “steel bra underwear causes cancer”), a dedicated rumor-busting column is suggested to be established, and the credibility level of the information will be jointly labeled with professional medical platform. Platforms should adjust their recommendation logic so that when users search for breast cancer-related symptoms, authoritative screening guidelines are prioritized over folk remedies or advertisements.
Second, policymakers should adopt a comprehensive strategy to strengthen the breast cancer care continuum. This entails promoting early screening and diagnosis through targeted initiatives, enhancing public awareness via robust health education campaigns, and ensuring appropriate funding and policy support for patient support organizations. Clear qualification requirements for posting breast cancer-related content on social media should be established (e.g., content must be reviewed by licensed physicians or professional institutions), along with penalty mechanisms for violations. For women with a family history or other high-risk factors, the age for free screening should be lowered from 40 to 35 or even 30.
Third, healthcare professionals should adopt an integrated care model that addresses both the physical and psychological dimensions of breast cancer. This entails providing patients with comprehensive and accurate medical information while systematically monitoring their mental well-being through the incorporation of psychological counseling services within healthcare institutions. For instance, a Narrative Medicine module can be integrated into continuing education to enhance physicians’ ability to provide emotional support within limited outpatient consultation time. Hospital-operated WeChat groups or dedicated sections within the hospital’s official app should be established, where survivors are regularly invited to share their experiences and experts answer questions, thereby forming a sustained support network outside the hospital setting.
5.5. Limitations and future directions
This study has several limitations. The findings are based solely on data from Weibo, whose user base and content dynamics may not fully represent other major platforms like Xiaohongshu or TikTok. Methodologically, the reliance on dictionary-based sentiment analysis and LDA for topic modeling is constrained in that the former may overlook contextual and emerging emotional expressions, while the latter is sensitive to hyperparameter settings and lacks semantic contextualization. Furthermore, the restricted time frame (2022–2024) limits insight into long-term discourse trends.
Future studies could expand data collection across multiple social media platforms and extend the temporal scope to better capture longitudinal patterns. Employing advanced techniques such as transformer-based models (e.g., BERT) for sentiment and topic analysis could enhance accuracy and semantic nuance. Additionally, combining Dynamic Topic Modeling (DTM) with causal inference methods may help trace policy impacts and public attention shifts. Cross-regional or cross-national comparisons could further reveal sociocultural variations in breast cancer discourse, ultimately supporting more responsive and targeted public health communication.
6. Conclusion
This study combined sentiment analysis and topic modeling with critical discourse analysis to examine large-scale breast cancer discourse on Weibo. Positive posts focused on recovery, treatment advances, prevention, and healthy lifestyles, suggesting that emotional support and health education can raise public awareness and boost confidence in breast cancer treatment. Negative posts centered on diagnosis-related distress, treatment challenges, and risk factors, highlighting the need for emotional support and symptom management to improve patient well-being. Neutral posts provided objective information on screening, patient experiences, and clinical progress, helping to demystify the disease and reduce fear. Together, these findings underscore the dual role of social media in both supporting patients and educating the public.
In Weibo context, public discourse constructs breast cancer as a women’s issue, a health crisis, and a matter of individual health management. While the public demonstrates strong awareness of prevention and early screening, concerns persist regarding treatment challenges and body image. Within the Chinese socio-cultural context—shaped by family values, filial piety, and traditional introverted norms—social awareness of breast cancer is particularly pronounced around family support, body image, and treatment-related concerns.
Our findings confirm that social media platforms are valuable data resources for uncovering public perceptions of disease such as breast cancer, monitoring opinion shifts in real time, and identifying evolving patterns of public attention to health issues. These insights can inform targeted public health strategies—particularly for breast cancer—to enhance awareness, shape attitudes, and improve stakeholder-led education and advocacy.
Footnotes
Ethical considerations
All study methods were approved by the ethics committee of Sichuan International Studies University (202400002). This research was performed in line with the principles of the Declaration of Helsinki. The text data was accessed in accordance with the Weibo platforms’ Terms of Service at the time of data collection. All data extracted were in the public domain and excluded any nonpublic user information.
Author contributions
Conceptualization: Jianan Li and Yu Deng; Data curation: Jianan Li; Formal analysis: Yu Deng; Funding acquisition: Juanjuan Chen; Investigation: Jianan Li; Methodology: Jianan Li and Yu Deng; Project administration: Jianan Li and Yu Deng; Resources: Jianan Li; Software: Jianan Li; Supervision: Yu Deng; Validation: Yu Deng; Visualization: Jianan Li; Writing – original draft: Jianan Li and Yu Deng; Writing – review and editing: Yu Deng and Juanjuan Chen
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by Graduate Education and Teaching Reform Research Project of Sichuan International Studies University (yjsjg202301).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Declaration of AI-assisted technologies
The authors used DeepSeek for language polishing and grammar checking during the writing and revision process. The authors subsequently reviewed all content and take full responsibility for the final work.