Can Virtual Companionship Alleviate Health Anxiety? A Longitudinal Study with Health Self-Efficacy as a Mediator

Health anxiety has become increasingly prevalent in recent years, driving demand for new forms of emotional support. The development of AI chatbots has introduced a new form of virtual companionship, characterized by emotionally supportive and information-driven interactions with AI systems. This study investigates the relationships among the frequency of use of virtual companionship (VCF), health self-efficacy (HSE), and health anxiety (HA). A three-wave longitudinal survey was conducted among 657 Chinese college students from May 2022 to May 2023. The results showed that frequent users of virtual companionship reported lower levels of HSE, which in turn predicted higher HA. Cross-lagged analysis revealed a reciprocal relationship between VCF and HA over time, suggesting a reinforcing cycle. These findings extend social support theory and self-efficacy theory, and introduce the reinforcing spirals framework into the context of AI-based human-computer interaction. The study offers insights for improving the emotional and informational design of health chatbots.