Study on the impact of recommendation algorithms on user perceived stress and health management behaviour in short video platforms

Short videos have become an ideal platform for global health information dissemination. Past research has tended to overlook the potentially negative experiences associated with recommendation algorithms and their positive coping mechanisms, while focusing too much on short-term effects at the expense of long-term effects. This study integrates multiple theoretical frameworks and proposes a new theoretical model aimed at determining the impact of this technostress from recommendation algorithms on users' health stress perceptions, health-related short videos usage, and health self-management behaviours. This paper builds a theoretical model based on the stressor-strain-outcome (SSO) framework, combining the health belief model and self-validation theory. Data was collected and analyzed using a diary study to understand the mechanism of the influence of daily stress perception on users' health habits and behavioural changes through multilevel modelling. The findings show that external stressor variables have no significant effect on the continuous usage intention of health-related short videos under algorithm recommendation. Cognitive validation and affective validation among the internal stressor variables have a positive effect on users' continued intentions. Notably, despite the presence of continuous usage intention, it was negatively related to health self-management behaviours. These findings provide new insights into the composition of internal and external health stressors driven by recommendation algorithm technostress, constructing a multilevel theoretical model of health-related short video usage, and tapping into the potential for the application of short video platforms in health self-management. © 2024 Elsevier Ltd