Surveillant care or benevolent control? A multimodal genre and sentiment analysis of therapeutic mHealth app reviews

The rapid expansion of mobile mental health (mHealth) apps has redefined the landscape of digital mental health interventions, offering unprecedented accessibility and scalability while embedding users within algorithmic infrastructures of care and control. This study adopts a multidisciplinary framework, integrating Natural Language Processing (NLP) and move analysis to interrogate the linguistic, affective, and visual strategies shaping user engagement with mHealth apps. Through an extensive analysis of 30,000 user reviews from Headspace, Calm , and BetterHelp , the paper uncovers a structured rhetorical pattern in user feedback, revealing the interplay of empowerment, algorithmic governance, and neoliberal self-optimization. Key insights include four emerging inter-related repertoires across the dataset. These findings advance the concept of therapeutic surveillance advanced by the current research endeavor, illustrating how mHealth apps function as digital shepherds – nudging users toward self-regulation while extracting behavioral and emotional data.