Sage Journals: Discover world-class research

Abstract

Aim

To explore the attitudes of healthcare professionals and non-healthcare professionals on anorexic behavior on social media.

Background

The significant function of attitude in alleviating anorexic behavior has been widely recognized. However, traditional methods often fail to capture patients’ hidden emotions due to stigma and fear of judgment. Social media provides a novel platform for anonymously examining these behaviors and emotions, offering insights into anorexic behaviors that can enhance intervention strategies.

Design

This study has a qualitative design based on machine learning.

Methods

Data was collected from Zhihu, Weibo, and Xiaohongshu social media platforms up to 1 September 2024. This study method consisted of five steps: data collection, data cleaning, validation of relevance, sentiment analysis, and content analysis using the K-means algorithm.

Results

This study comprised 1099 comments, comprising 277,793 words. Non-healthcare professionals had seven emotions (good, happy, surprise, anger, disgust, fear, and sad) for anorexic behavior, and negative emotions were predominant. Healthcare professionals had three emotions (happy, good, and sad), and negative emotions were predominant. Healthcare professionals have a role deficit in recognizing negative emotions.

Conclusion

The study emphasizes the need for healthcare professionals to improve the recognition of negative emotions expressed by non-healthcare professionals/patients and develop data-driven interventions that address psychological barriers, fostering holistic patient care and improving outcomes.

Keywords

Artificial intelligence attitude medicine qualitative study social media

Introduction

Anorexia is a psychological disorder characterized by extreme restriction of dietary intake, intense fear of weight gain, and distorted body image perception. 268.7 anorexia deaths per 100,000 people will occur globally in 2019.¹ Anorexia poses a serious threat to the functioning of bodily systems and entails the risk of several complications and dysfunctions, ultimately affecting the quality of life.^2,3 Therefore, anorexia has become an increasingly serious global public health problem. Effectively intervening in anorexic behavior is essential for enhancing patients’ health and significantly alleviating the medical burden on families and Society.

The significant function of attitude in alleviating anorexic behavior has been widely recognized.^4,5 To provide a theoretical framework, attitude is generally understood as a psychological construct comprising cognitive (beliefs), affective (emotions), and behavioral components. In this study, we operationalize attitude by focusing on the affective (emotional) component as derived from sentiment analysis and the cognitive component represented by the content analysis clusters (public concerns and beliefs). Studies have shown that positive emotional support might assist patients in alleviating negative feelings and augmenting their trust and adherence to therapy, thus improving their psychological well-being and behavioral performance.^6,7 However, the prevailing methodology for examining emotions in anorexic behavior predominantly relies on theoretical studies. Fewer empirical studies’ results are also limited to using health professional face-to-face interviews and observation.^6–10 Patients exhibiting anorexic behaviors may hesitate to provide information to health professionals due to societal stigma and apprehension of judgment or misinterpretation, while in-person interviews may intensify this anxiety, hindering patients’ openness and honesty.^8,11,12 Therefore, analyzing the emotions hidden in patients’ anorexic behaviors will provide new research directions for the optimization of intervention strategies.

The prevalence of social media has created new opportunities for examining hidden emotions. While previous studies have failed to capture patients’ hidden emotions, the anonymity and interaction on social media allow anorexic patients to express their inner emotions and share details of their behaviors more freely.¹³ The prevalence of social media, such as China's Weibo, Zhihu, and Xiaohongshu (with hundreds of millions of monthly active users), has created new opportunities for examining hidden emotions. The original, valuable comments on social media from patient-initiated postings are the foundation for the analysis of the resulting attitudes from the non-healthcare professional and healthcare professional communities.¹⁴ Compared with traditional research methods, the Internet's anonymity and interaction allow anorexic patients to express their inner emotions and share details of their behaviors more freely, which, to some extent, reduces social stigma and the fear of being judged or misunderstood.^9,15 Furthermore, the Internet offers patients extensive community support, allowing them to cultivate a sense of identity and emotional connection through online interactions, a beneficial influence that can enhance their rehabilitation.¹⁶ However, the tedious Internet data and the spreading effect of negative emotions remain challenging to address in research. Machine learning methods provide practical support to address this issue.^13,17 Thus, this study uses internet data to study hidden behaviors and an important research perspective for exploring patients’ deeper emotions.

Although social media is valuable for examining anorexic behavior, academics have neglected the significance of social media comments.¹⁶^18–21 The existing social media studies use the Internet as a platform for communication.²² The original, valuable comments on social media from patient-initiated postings were ignored. Social media, as a primary platform for patients’ spontaneous expression, encompasses emotional and behavioral data crucial for developing data-driven intervention models.^9,23 Systematic study of Internet data can reveal the complex features of hidden anorexic behaviors and provide evidence support for public health policymaking. Especially in the early warning and precise intervention of diseases, Internet data analysis will provide a new theoretical basis and practical direction for mental health. Recent studies have highlighted that social media platforms have become important spaces where harmful, stigmatized, or misunderstood health-related behaviors are expressed and interpreted.^16,20 Syed-Abdul et al. demonstrated that video-based platforms such as YouTube can disseminate misleading content related to anorexia, thereby influencing public perceptions and reinforcing unhealthy behavioral norms. Similar patterns have been observed on Chinese platforms, where pro-eating disorder narratives coexist with expressions of distress, concern, and stigma. These findings underscore the need to understand how both non-healthcare and healthcare professionals respond emotionally and cognitively to anorexic behaviors online, particularly in non-Western digital environments where cultural norms may shape the expression and interpretation of health-related content.

This study aimed to conduct a sentiment analysis of anorexic behavior using Internet data to reveal the characteristics of hidden behaviors and their emotional drivers. Specifically, this study will answer the following questions on anorexia: (a) What are the non-healthcare professional and healthcare professional concerns toward anorexia? (b) What are the non-healthcare professional and healthcare professional attitudes toward anorexia?

Methods

Data acquirement

The study utilized “anorexia behavior” as the search term, with data collection from database establishment to 1 September 2024. Social media platforms analyzed included Zhihu, Weibo, and Xiaohongshu. Collecting information comprised variables such as date, user type, text content, and the number of likes. User types were categorized into three groups: non-healthcare professional, healthcare professional, and unknown. Within the non-healthcare professional category, subgroups included self, family/friend, celebrity, and media organization. Ultimately, the study designated texts from the non-healthcare professional category to represent public opinions, while texts from the healthcare professional category were used to reflect healthcare professionals’ perspectives.

Data cleaning

The data cleaning process aims to remove irrelevant information, ensuring the dataset is refined for analysis. It consists of four key steps. First, irrelevant data, such as advertisements, spam comments, and duplicate entries, are eliminated to minimize redundancy and prevent interference with the analysis. Second, text is standardized through a six-step normalization process: converting text to lowercase, removing punctuation marks, eliminating special characters and emojis, correcting spelling errors, managing spacing and abbreviations, and deleting URLs and usernames (e.g. @username). Third, language detection filters out comments in languages other than English or Chinese, ensuring relevance. Finally, sensitive details such as phone numbers and email addresses are removed or anonymized to protect privacy.

Relevance verification

A relevance verification process is conducted to ensure the dataset accurately represents the perspectives of healthcare professionals and the public regarding anorexic behavior. While the data cleaning step effectively removes irrelevant and redundant information, it is less capable of identifying subtle advertisements or misinformation. A double manual review is employed, where two trained reviewers independently assess the relevance and usability of the data. This rigorous approach enhances the accuracy and reliability of the dataset, ensuring it aligns with the study's objectives.

Sentiment analysis

The sentiment analysis process evaluates public emotions toward anorexia behavior, involving several steps.¹³ The first step is selecting and expanding the Dalian University of Technology Sentiment Lexicon, which includes word types, sentiment categories, intensity, and polarity. This lexicon has been adapted for the Chinese context based on Ekman's emotional classification system. Next, the text undergoes segmentation, divided into three parts: Chinese word segmentation using Jieba to split the text into individual words while retaining part-of-speech information; handling synonyms by replacing or tagging them with the HOWNET synonym library to improve accuracy; and creating a custom medical terminology dictionary to enhance the analysis of domain-specific texts.

Following segmentation, word matching, and sentiment annotation are performed. Words are matched against the sentiment lexicon to identify their sentiment category, polarity, and intensity. The matched words are then annotated in the text, with adjustments made for contextual factors such as negations or antonyms. Sentiment scoring is conducted using weighting and counting methods, where positive and negative scores are subtracted to determine sentiment polarity. Scores are standardized for specific applications, ranging from −1 to 1. Double manual annotation is performed on a sample of texts by two trained annotators, and reliability is assessed using metrics like Cohen's kappa. Automated sentiment analysis results are compared against manual annotations, achieving an F1 score exceeding 80%, thus confirming the algorithm's accuracy.

Content analysis using the K-means algorithm

The study utilizes the K-means algorithm, an unsupervised machine learning technique, to uncover themes and patterns within the text data.¹⁷ The process begins with initialization, where initial cluster centers from the dataset are randomly selected. The number of clusters (K) is determined through domain knowledge or preliminary analysis, with multiple initializations performed to enhance robustness. Next, iterative optimization refines the clustering process. Data points are assigned to clusters based on proximity to cluster centers, calculated using Euclidean distance. Cluster centers are then updated as the mean of data points within each cluster, and this reassignment and updating process continues until clusters stabilize.

The algorithm terminates when cluster centers show minimal change between iterations or when a maximum number of iterations are reached, ensuring computational efficiency. Evaluation and interpretation follow, using the elbow method to determine the optimal number of clusters and assess clustering quality. Each cluster is examined for dominant topics by analyzing frequent terms, keywords, and sentiment scores. Clusters are validated against known themes to ensure alignment with expected insights. By integrating these clustering results with sentiment analysis, the study identifies meaningful patterns and provides a deeper understanding of the discourse surrounding anorexia.

Results

This study included 1099 comments, comprising a total of 277,793 words. A total of 906 comments were categorized as non-healthcare professionals (82.4% of total comments). Within the non-healthcare professional category, self/family/friend comments constituted 82% (representing the voices of those personally affected, such as patients or their immediate social circle). A total of 184 comments were categorized as healthcare professionals (16.7% of total comments), and nine comments (0.8%) were from unknown users.

Non-healthcare professional emotions and content clusters

Non-healthcare professionals had seven emotions for anorexic behavior. The content clusters represent the underlying beliefs and concerns (cognitive component of attitude) associated with each emotion.

Good emotion includes cognitive restructuring (positive cognition), rebuilding self-worth, and hope. Happy emotion includes emotional stability (inner state) and social support (external validation). Surprise emotion includes diagnostic shock. Anger includes family conflict, anger internalization and self-control, and guilt. Disgust emotion contains body image concerns, social stigma, anxiety, emotional numbness, shame, and physical nausea. Fear emotion contains weight gain and uncertainty. Sad emotion contains despair and family grief transmission (Table 1).

Table 1.

Sentiment analysis and content analysis of non-healthcare professional.

Emotion	Related sentiment analysis word	Content analysis
Good	Positive emotions, active, psychological interventions, and cognitive therapy	Cognitive restructuring
	Value, competence, health management, and self-esteem	Rebuilding self-worth
	Hope, treatment, and future	Hope
Happy	Good, okay, calm, submissive, and satisfied.	Emotional stability
Happy	Friends, social groups, encouragement, support, and money	Social support
Surprise	Why, answer, reason, method, and first time	Diagnostic shock
Anger	Complaint, anger, annoying, temper, lose, and family	Family conflict
	Break out, control, protest, and punish	Anger internalization and self-control
	Guilt, shame, worry, and relationship	Guilt
Disgust	Thin, fat, not eating, body shape, follow, and like	Body image concerns
	Friends, Internet, social media, discrimination, words, and fights	Social stigma
	Perfectionism, anxiety, fear, and worry	Anxiety
	Abuse, disease, repression, emotions, not caring, casual, and coping	Emotional numbness
	Weight, concerns, environment, and external	Shame
	Greasy, picky, protein, picky, and odor	Physical nausea
Sad	Sad, despair, fear, and no way	Despair
Sad	Family, siblings, and communication	Family grief transmission
Fear	Nightmares, cases, weight, and size	Weight gain
Fear	Complications, timidity, fear, and later	Uncertainty

Healthcare professional emotions and content clusters

Healthcare professionals had three emotions for anorexic behavior: happy, good, and sad. Happy emotions include treatment effectiveness, good emotions include new treatments, and sad emotions include compassion (Table 2).

Table 2.

Sentiment analysis and content analysis of healthcare professional.

Emotion	Related sentiment analysis word	Content cluster (beliefs/concerns)
Good	New technologies, breakthroughs, research, and techniques	New treatments
Happy	Success, effective, safety, better, and happy	Treatment effectiveness
Sad	Guilty, suffering, bad experience, and tolerance	Compassion

Discussion

This study's results show the presence of seven emotions toward anorexic behavior by non-healthcare professionals and three emotions by healthcare professionals. This study is the first to focus on the voices of anorexic behaviors on social media, aiming to enhance future health education and psychological support that healthcare professionals provide.

Non-healthcare professionals hold negative attitudes

Non-healthcare professionals had seven emotions for anorexic behavior, and negative emotions were predominant.

Good emotion includes cognitive restructuring, rebuilding self-worth, and hope. Cognitive restructuring involves reconceptualizing anorexic behavior, allowing patients to generate positive cognition from negative experiences. This result is partially consistent with existing studies.^19,24,25 Therefore, the identification of the cognitive restructuring theme in the data suggests that psychological interventions, such as cognitive behavioral therapy or cognitive remediation therapy, are highly relevant for alleviating inner conflict and improving psychological resilience. Rebuilding self-worth reflects the important role of psychological restorative approaches, particularly in anorexic behavior. This result is partially consistent with Alserihi's study.²⁴ Patients progressively rebuild their self-esteem and self-efficacy throughout treatment via effective health management or external recognition of recovery efforts. Hope reflects the patient's reasonable expectations. The patient's optimism for a potential recovery may stem from affirmative feedback regarding treatment efficacy or external support, which offers psychological motivation for ongoing engagement in treatment. This result is partially consistent with Delaquis's study.²⁶

Happy emotion includes emotional stability and social support. Emotional stability reflects the importance of psychological factors. This result is partially consistent with Natali's study.²⁷ Patients may reduce anxiety and depressive symptoms and gradually achieve emotional stability through psychosocial interventions. This result showed that future intervention studies should concentrate on the influence of psychological and emotional factors on intervention efficacy. Social support reflects the emotional and substantive support provided by the family, healthcare team, and social network. This result is partially consistent with Cripps's et al.,²⁸ Saukko et al.,³⁰ and Galiana Camacho et al.²⁹ study. Social support promotes restoring the patient's confidence and alleviates feelings of isolation by increasing belonging.

Surprise emotion includes diagnostic shock. This study verified that the sudden cognitive changes brought about by the first definitive diagnosis triggered the surprise emotion. Diagnosis initiates disease intervention and marks the beginning of the patient's reassessment of their condition and psychological adaptation. Therefore, this study emphasizes the importance of the initial diagnosis. Patients are highly willing to participate in interventions when they have high self-efficacy. The earlier the intervention, the higher the probability of recovery.

Anger includes family conflict, anger internalization and self-control, and guilt. Family conflict reflects the complex impact of family interactions on disorder management. Conflicts arising from family members’ misconceptions about anorexic behavior or a narrow focus on food intake and weight may exacerbate the patient's anger. This finding suggests the continued relevance of systemic approaches, like family therapy, to address these relational dynamics.^28,29 Anger internalization and self-control reflect the intrinsic motivational mechanisms of high stress. This result is partially consistent with the speset's study.³¹ Patients may internalize anger due to unmet self-imposed goals, which can result in excessive control over eating and behavior. Guilt reflects individual moral cognition and intimacy dynamics. This result is partially consistent with Petry's study.³² Patients feel guilty for the consequences of sudden, angry behavior.

Disgust emotion contains body image concerns, social stigma, anxiety, emotional numbness, shame, and physical nausea. Body image concerns reflect social bias. This result is partially consistent with Bachner's study.³³ Patients’ distorted perceptions of their body image were further exacerbated by unrealistic sociocultural expectations of the “ideal body,” resulting in self-loathing. Social stigma is influenced by cultural context. This result is partially consistent with Chubbs-Payne et al.,⁶ Sebastian et al.,¹² and Caslini's study.³⁴ Social stigma is influenced by cultural context. Stigmatization in informal contexts correlates with adverse personality characteristics and diminished social expectations. Therefore, the identification of social stigma as a disgust-related emotion is critical and suggests the need for public health policies and psychoeducational interventions targeting the non-healthcare professional community to reduce stigma and promote better social support for individuals with anorexic behaviors. Emotional numbing has a two-sided effect. This study is the first to validate the sentiment for anorexic behavior. Patients may experience emotional numbing triggered by disease progression or treatment uncertainty. At the same time, emotional numbness may function as a psychological defense mechanism, diminishing patients’ sensitivity to external stimuli. Shame reflects that the patient feels self-perceived inadequacy or social prejudice. This result is partially consistent with Dimitropoulos's study.³⁵ The self-perception of body image in individuals diagnosed with anorexia. Patients may experience shame feelings related to weight gain, bulimic behaviors, or external scrutiny of their condition. This shame may ultimately lead to difficulties in dating and relationships for the patient.³⁶ Physical nausea is the pathology of regurgitation in which the patient is confronted with food. The regurgitation phenomenon can cause the patient to fear eating.

Fear emotion contains weight gain and uncertainty. Fear of weight gain was associated with body image deterioration. This result is partially consistent with Galiana Camacho et al.,²⁹ Espeset et al.,³¹ Achermann et al.,³⁷ and Sauve's study.³⁸ Fear of body image deterioration reflects the patient's excessive preoccupation with body shape within a pathological cognitive framework. Uncertainty reflects the significant impact of anxiety and insecurity on the illness experience. This result is partially consistent with Zickgraf's study.³⁹ Uncertainty about treatment outcomes, social acceptance, and future health status exacerbated patients’ fears.

Sad emotions contain despair and family grief transmission. Despair correlated with diminished self-confidence in recovery and chronic psychological stress. This result is partially consistent with Fitzsimmons-Craft et al.,³⁶ Brockmeyer et al.,⁴⁰ and Eiring's study.⁴¹ Under treatment failure or disease relapse, patients may experience significant hopelessness. Family grief transmission was associated with family emotional dynamics. This result is partially consistent with Galiana's study.²⁹ Family members’ emotions, such as anxiety and disappointment, may affect patients through interactive mechanisms that further exacerbate their sadness.

Health professionals hold positive attitudes

The health professional has three emotions for anorexic behavior; positive emotions were predominant. The condition is related to the health professional's role as a health facilitator. The condition is related to the health professional's role as a facilitator. The health professional communicates positive messages to encourage the patient to have a positive outlook on life. Happy emotion indicates treatment effectiveness, which arises from enhancing the patient's health or progressive recovery. This result is consistent with Bamford's study.⁴² Good emotion arises from trust and expectation regarding the new technology. This study firstly validates this result. New treatments for anorexia nervosa, including advanced nutritional support, cognitive-behavioral therapy, and technological interventions, provide caregivers with optimism and positive emotional responses. Learning and mastering these new techniques also bring fulfillment to their professional growth. Sad emotion reflects the physician's empathy for the patient's suffering and lack of improvement. This result is consistent with Bamford's study.⁴² It is worth noting that a patient's suicide also triggers health professionals’ grief. Therefore, it is important to give health professionals psychological support in advance.

Health professionals have a role deficit in recognizing negative emotions

Health professionals have a role deficit in recognizing negative emotions on the Internet. This result may be related to insufficient training, time, and resource limitations, and ambiguous role definitions for health professionals. Firstly, the current education system emphasizes pathology, physiology, and technical training while insufficiently addressing the development of skills in emotion recognition and psychological support for health professionals. Second, excessive workload and resource constraints hinder healthcare professionals from dedicating sufficient time to addressing patients’ emotional needs. Further, the problem is exacerbated by role ambiguity, where healthcare professionals tend to view themselves as disease treaters rather than emotional supporters. Additionally, an unclear division of labor within the team regarding patient emotion management leads to poor implementation. Therefore, this study suggests three solutions. First, the curriculum of vocational training programmers should be enhanced to include content on emotion management. Secondly, allocating time and resources strategically to enhance the psychological support capabilities of the healthcare team. Finally, clarifying the responsibilities associated with emotional care within multidisciplinary collaboration and establishing a standardized process.

Limitations

This study included three limitations. First, the Internet data was biased, excluding the viewpoints of older persons with limited Internet abilities. The emotional attitudes of the older population as a group with higher health risks are valuable. Future research should concentrate on anorexic behavior among older adults. Secondly, the study's focus on text content meant it did not analyze user engagement metrics such as followers, forwards, or content types such as images-only posts. Furthermore, emojis were removed during the data cleaning process to standardize the text for machine learning. The study could not ascertain the clinical specialty or whether the healthcare professionals in the dataset actively treat eating disorders. Unlike traditional qualitative studies, this machine-learning-based qualitative design focused on identifying large-scale emotional and thematic clusters and did not include literal quotes to represent the high volume of text data. Finally, this study could not ascertain whether the comments referred specifically to a diagnosis of anorexia nervosa or bulimia nervosa, or to other forms of disordered eating, as the search term was “anorexic behavior.” Furthermore, the study primarily used the Chinese term for “anorexic behavior,” which may have excluded content using slang or censored terms (e.g. “ana” or “an0rexia” in other contexts).

Conclusions

This study found that non-healthcareers had negative emotions towards anorexic behaviors, while health professionals had positive emotions. Health professionals inadequately identified negative emotions in non-healthcare on the Internet, which may result in a mismatch between medical interventions and the actual needs, thereby diminishing treatment effectiveness. Identifying negative emotions within the non-healthcare community can enhance their understanding of patients’ psychological needs and potential emotional barriers. Sentiment analysis can help non-healthcare patients express and address their emotions and avoid further mental health problems caused by prolonged emotional neglect. Therefore, this study concludes that health professionals must pay attention to the negative emotions of non-healthcare users regarding anorexic behaviors on the Internet. This action can comprehend hidden behaviors and emotions, promoting holistic health behaviors among non-healthcare professionals.

Footnotes

Acknowledgements

The authors would like to thank Yuqi Xv, who provided advice and assistance with this study.

ORCID iD

Qi Zhou

Ethics approval and consent to participate

The study did not involve human subjects, and the data came from publicly available platforms for academic research. Therefore, after the Institutional Review Board's approval, the researchers only collected comments, which did not involve human subjects’ privacy and did not cause human subjects harm. Ethics approval was waived by the review board of Hangzhou Linping District Integrated Traditional Chinese and Western Medicine Hospital (No. 2025-L-014).

Consent for publication

Not applicable.

Author contributions

DH Q, HQ W, YTY, YKM, JKY, LHY, ZQW, and QZ participated in designing, data collection, data analyze, and drafted the manuscript. QZ critically assessed the manuscript, and all authors read and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

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

The datasets generated and analyzed during the present study are not publicly available, as this data can only be used for academic research, but are available from the corresponding author on reasonable request. If the dataset is needed, support can be obtained from Xaiohongshu, Weibo, and Zhihu platforms.

References

GBD 2019 Mental Disorders Collaborators. Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry 2022; 9: 137–150.

Westmoreland

Krantz

Mehler

. Medical complications of anorexia nervosa and bulimia. Am J Med 2016; 129: 30–37.

Zipfel

Giel

Bulik

, et al. Anorexia nervosa: aetiology, assessment, and treatment. Lancet Psychiatry 2015; 2: 1099–1111.

Hanegraaf

Anderson

Neill

, et al. Treatment targets and strategies for eating disorders recovery: a Delphi consensus with lived experience, carers, researchers, and clinicians. Int J Eat Disord 2024; 57: 2494–2507.

Kamrani

Kachouei

Sobhani

, et al. Nourishing the mind: how the EAT-lancet reference diet (ELD) and MIND diet impact stress, anxiety, and depression. BMC Psychiatry 2024; 24: 709.

Chubbs-Payne

Lee

Isserlin

, et al. Attitudes toward physical activity as a treatment component for adolescents with anorexia nervosa: an exploratory qualitative study of patient perceptions. Int J Eat Disord 2021; 54: 336–345.

Kons

Essayli

Shook

. Comparing the knowledge, attitudes, and practices of pediatric and family medicine clinicians toward atypical anorexia nervosa versus anorexia nervosa. Int J Eat Disord 2024; 57: 993–1001.

Crisafulli

Von Holle

Bulik

. Attitudes towards anorexia nervosa: the impact of framing on blame and stigma. Int J Eat Disord 2008; 41: 333–339.

Ramírez-Cifuentes

Freire

Baeza-Yates

, et al. Characterization of anorexia nervosa on social media: textual, visual, relational, behavioral, and demographical analysis. J Med Internet Res 2021; 23: e25925.

10.

Reas

Isomaa

Solhaug Gulliksen

, et al. Clinicians as a critical link: understanding health professionals’ beliefs and attitudes toward anorexia nervosa, bulimia nervosa, and binge eating disorder. Scand J Psychol 2021; 62: 775–779.

11.

Caslini

Crocamo

Dakanalis

, et al. Stigmatizing attitudes and beliefs about anorexia and bulimia nervosa among Italian undergraduates. J Nerv Ment Dis 2016; 204: 916–924.

12.

Sebastian

Richards

Bilgin

. Education and contact strategies to reduce stigmatising attitudes towards anorexia nervosa among university students. Health Educ J 2017; 76: 906–922.

13.

Zhou

Yang

, et al. Attitudes of the public and medical professionals toward nurse prescribing: a text-mining study based on social medias. Int J Nurs Sci 2024; 11: 99–105.

14.

Zhou

Lei

Tian

, et al. Perception and sentiment analysis of palliative care in Chinese social media: qualitative studies based on machine learning. Soc Sci Med 2025; 379: 118178.

15.

Oksanen

Garcia

Räsänen

. Proanorexia communities on social media. Pediatrics 2016; 137: 1–2. DOI: 10.1542/peds.2015-3372.

16.

Suresh

Pallempati

Saxena

, et al. Exploring YouTube videos about anorexia nervosa on the basis of reliability, popularity, and contributions of healthcare professionals: a cross-sectional study. Cureus 2023; 15: e48095.

17.

Zhou

Lei

, et al. Public concerns and attitudes towards autism on Chinese social media based on K-means algorithm. Sci Rep 2023; 13: 1–8.

18.

Hanegraaf

Anderson

Neill

, et al. Treatment targets and strategies for eating disorders recovery: a Delphi consensus with lived experience, carers, researchers, and clinicians. Int J Eat Disord 2024; 57: 2494–2507.

19.

Praet

Stevens

Casteels

, et al. The association of social media use and eating behaviour of Belgian adolescent girls diagnosed with anorexia nervosa – a qualitative approach. Children (Basel) 2024; 11: 822.

20.

Syed-Abdul

Fernandez-Luque

Jian

W-S

, et al. Misleading health-related information promoted through video-based social media: anorexia on YouTube. J Med Internet Res 2013; 15: e30.

21.

Yom-Tov

Fernandez-Luque

Weber

, et al. Pro-anorexia and pro-recovery photo sharing: a tale of two warring tribes. J Med Internet Res 2012; 14: e151.

22.

Holmes

. ‘My anorexia story’: girls constructing narratives of identity on YouTube. Cultural Studies 2017; 31: 1–23.

23.

Weimann-Saks

Eliash

. Media images of anorexia: effects of moderate vs. extreme anorexia-related messages on young women. Community Ment Health J 2020; 56: 823–829.

24.

Alserihi

Hubayni

Alotaibi

, et al. Effect of cognitive remediation therapy in anorexia nervosa: a systematic review and meta-analysis. Front Psychiatry 2024; 15: 1484457.

25.

Real-Brioso

Estrada

Ruiz-Lee

, et al. Early response in people with anorexia nervosa receiving cognitive-behavioural therapy for eating disorders (CBT-ED): a latent change study. Psychother Res 2024; 35: 1–16.

26.

Delaquis

Godart

Barry

, et al. Perfectionism, self-esteem, and affective symptoms in anorexia nervosa subtypes: a network analysis of French inpatients. J Clin Psychol 2024; 80: 1852–1875.

27.

Natali

Meregalli

Rowlands

, et al. Virtual food exposure with positive mood induction or social support to reduce food anxiety in anorexia nervosa: a feasibility study. Int J Eat Disord 2024; 57: 703–715.

28.

Cripps

Serpell

Pugh

. Processes of change in family therapies for anorexia nervosa: a systematic review and meta-synthesis of qualitative data. J Eat Disord 2024; 12: 104.

29.

Galiana Camacho

León Latorre

Martos Enrique

Análisis de contenido sobre prescripción enfermera en los medios de comunicación

Dissertation/Thesis. Utah State University, Almería, Spain, 2020. https://e-revistas.uc3m.es/index.php/RECS/article/view/4815 (accessed 6 January 2025).

30.

Saukko

Malson

Brown

. The ethoses of (dis)connecting with friends on social media: digital cocooning and entrepreneurial networking among people with eating disorders. Social Media + Society 2024; 10: 20563051241287284.

31.

Espeset

EMS

Gulliksen

Nordbø

RHS

, et al. The link between negative emotions and eating disorder behaviour in patients with anorexia nervosa. Eur Eat Disord Rev 2012; 20: 451–460.

32.

Petry

Vasconcelos

FAG

Costa

LCF

. Feelings and perceptions of women recovering from anorexia nervosa regarding their eating behavior. Cad Saude Publica 2017; 33: e00048716.

33.

Bachner-Melman

Zohar

Elizur

, et al. Protective self-presentation style: association with disordered eating and anorexia nervosa mediated by sociocultural attitudes towards appearance. Eat Weight Disord 2009; 14: 1–12.

34.

Caslini

Crocamo

Dakanalis

, et al. Stigmatizing attitudes and beliefs about anorexia and bulimia nervosa among Italian undergraduates. J Nerv Ment Dis 2016; 204: 916–924.

35.

Dimitropoulos

McCallum

Colasanto

, et al. The effects of stigma on recovery attitudes in people with anorexia nervosa in intensive treatment. J Nerv Ment Dis 2016; 204: 370–380.

36.

Fitzsimmons-Craft

Eichen

Monterubio

, et al. Longer-term follow-up of college students screening positive for anorexia nervosa: psychopathology, help seeking, and barriers to treatment. Eat Disord 2020; 28: 549–565.

37.

Achermann

Günther

Goth

, et al. Body-related attitudes, personality, and identity in female adolescents with anorexia nervosa or other mental disorders. Int J Environ Res Public Health 2022; 19: 4316.

38.

Sauve

Silva

Costa

, et al. Adolescent girls with anorexia nervosa clinical profiles in relation to their social networks use. Eur Eat Disord Rev. Epub ahead of print 30 November 2024. DOI: 10.1002/erv.3157.

39.

Zickgraf

Schwartz

. Broadening the conceptualization of panic disorder to include the fear presentation of avoidant/restrictive food intake disorder: the legacy of panic control therapy. Behav Res Ther 2024; 185: 104677.

40.

Brockmeyer

Grosse Holtforth

Bents

, et al. Lower body weight is associated with less negative emotions in sad autobiographical memories of patients with anorexia nervosa. Psychiatry Res 2013; 210: 548–552.

41.

Eiring

Wiig Hage

Reas

. Exploring the experience of being viewed as ‘not sick enough’: a qualitative study of women recovered from anorexia nervosa or atypical anorexia nervosa. J Eat Disord 2021; 9: 142.

42.

Bamford

Mountford

. Cognitive behavioural therapy for individuals with longstanding anorexia nervosa: adaptations, clinician survival and system issues. Eur Eat Disord Rev 2012; 20: 49–59.

Non-healthcare and healthcare professionals’ attitudes towards behaviors of anorexia: A qualitative analysis of Chinese social media content

Abstract

Aim

Background

Design

Methods

Results

Conclusion

Keywords

Introduction

Methods

Data acquirement

Data cleaning

Relevance verification

Sentiment analysis

Content analysis using the K-means algorithm

Results

Non-healthcare professional emotions and content clusters

Healthcare professional emotions and content clusters

Discussion

Non-healthcare professionals hold negative attitudes

Health professionals hold positive attitudes

Health professionals have a role deficit in recognizing negative emotions

Limitations

Conclusions

Footnotes

Acknowledgements

ORCID iD

Ethics approval and consent to participate

Consent for publication

Author contributions

Funding

Declaration of conflicting interests

Data availability statement

References