DIGITAL SPACES AND EATING DISORDERS: A LITERATURE REVIEW ON THE IMPACT OF SOCIAL MEDIA ALGORITHMS AND MOBILE HEALTH APPLICATIONS ON RECOVERY

Abstract

Background: Digital spaces function as an active, structural determinant of eating disorder (ED) psychopathology, representing a modern "digital exposome" that poses significant psychiatric and systemic medical risks. While commercial social media algorithms often act as risk factors, the rapid evolution of digital psychiatry simultaneously offers new scalable therapeutic options. Objectives: This literature review aims to examine the dual impact of modern digital environments, critically contrasting the pathogenic mechanisms of engagement-driven social media algorithms with the clinical efficacy of mobile health (mHealth), virtual reality (VR), and artificial intelligence (AI) interventions. Methodology: A narrative review of recent literature (2019–2026) was conducted utilizing PRISMA-informed search strategies across primary medical and psychological databases. The synthesis integrated epidemiological meta-analyses, bibliometric trends, and randomized controlled trials (RCTs) evaluating both algorithmic harm and e-mental health therapeutics. Results: Findings indicate that engagement-driven algorithms (e.g., the "TikTok paradox," mukbang trends) systemically exacerbate self-objectification and frequently precipitate severe somatic complications, including nephrological, cardiovascular, and metabolic crises, across diverse demographics, extending to physically active males. Conversely, adaptive mHealth applications, VR cue-exposure protocols, and Large Language Model (LLM) chatbots demonstrate significant preliminary efficacy in cognitive restructuring and symptom reduction. However, widespread clinical adoption remains hindered by poor user experience (UX) retention rates and severe ethical dilemmas regarding AI "hallucinations" and medical safety. Conclusion: Mitigating the digital ED crisis necessitates a  coordinated, multidisciplinary clinical approach, and highlights the urgent need to integrate routine "digital anamnesis" into primary and psychiatric care. Future imperatives include structural algorithmic regulation, the establishment of standardized AI safety frameworks, and the development of user-centric digital therapeutics to ethically redesign the digital exposome.