Key drivers and mechanisms of user satisfaction with smart wearable medical products: a mixed-methods study
Abstract
Smart wearable medical products are increasingly used in chronic disease management and health monitoring. However, negative user experiences and discontinuance remain common in real-world use, and the multidimensional drivers of user satisfaction (US) and their interaction mechanisms remain insufficiently explained. This study adopts a mixed-methods design integrating qualitative identification, structural equation modeling (SEM) validation, and XGBoost–SHAP explanation. It aims to identify the core dimensions of user concern regarding smart wearable medical products, reveal their pathways of influence on US, and characterize the importance and synergistic effects of key factors within an explainable framework. First, based on 5200 valid reviews from e-commerce platforms and 13 semi-structured interview transcripts, procedural grounded theory was used to identify the key factors influencing US. Second, 391 valid questionnaire responses were collected, and SEM was applied to test the relationships among latent variables, while XGBoost–SHAP was used to analyze feature importance and nonlinear interaction effects. The SEM results show that human–AI interaction experience (HAIX), medical value factor (MV), hardware design factor (HD), and quality and performance factor (QP) all have significant positive effects on US ( p < 0.01), whereas commercial marketing factor (CMM) is not significant ( p = 0.14). The SHAP results further indicate that HAIX and QP have relatively high importance and exhibit a strong synergistic interaction. These findings suggest that US with smart wearable medical products is not determined by a single product attribute, but is jointly shaped by HAIX, MV, HD, and QP. Methodologically, this study develops a mixed empirical pathway integrating qualitative identification, SEM validation, and XGBoost–SHAP explanation. Theoretically, it introduces HAIX into satisfaction research on medical wearables. Practically, it provides actionable guidance for firms to optimize interaction design, performance reliability, and the realization of medical value.
// Source
Authors: Hao Fang, Guohui Cun, Zhou Ya, Qian Bao, Sisi Xie, Minjian Hong, Yaxi Wang
Institutions: China University of Geosciences, Wuhan Institute of Technology