Abstract
BACKGROUND: Digital health technologies (DHTs) are transforming global health care delivery, yet physician adoption remains highly variable and influenced by a complex interplay of individual, institutional, and technological factors. In China, despite national initiatives such as the "Healthy China 2030" strategy promoting DHT integration, understanding physicians' heterogeneous perceptions is essential for effective implementation. OBJECTIVE: This study aimed to identify distinct latent profiles of Chinese physicians based on their perceptions of DHT benefits, barriers, and behavioral intention, and to examine the demographic and occupational factors associated with profile membership. METHODS: A cross-sectional survey was conducted among 4851 physicians (female, n=2994, 60.69% ; mean age 38.4, SD 8.7 years; 51.58% (n=2502) with more than 10 y working experience) from 46 hospitals in Shaanxi Province, China, between October and December 2023. Assessment included nine indicators across three domains: Perceived Benefits (4 items), Adoption Barriers (4 items), and Behavioral Intention (1 item). Latent profile analysis was used to identify distinct subgroups of physicians based on their response patterns. Multinomial logistic regression examined predictors of profile membership, and results were reported as odds ratios (ORs) with 95% CIs. RESULTS: The latent profile analysis identified 5 distinct profiles: Reform-Adaptable (n=516, 10.64%; 95% CI 9.76%-11.52%), Negative (n=1003, 20.68%; 95% CI 19.50%-21.86%), Neutral (n=2276, 46.92%; 95% CI 45.50%-48.34%), Reform-Conservative (n=545, 11.23%; 95% CI 10.33%-12.13%), and Positive (n=511, 10.53%; 95% CI 9.66%-11.40%). Significant intergroup differences were observed in demographic and occupational characteristics. For instance, compared with the Negative profile, male physicians were less likely to belong to the Neutral (OR 0.76, 95% CI 0.64-0.90; P=.001) and Reform-Conservative (OR 0.67, 95% CI 0.54-0.84; P=.001) profiles. Compared to the Neutral profile, physicians with a master's degree or above were less likely to be in the Reform-Conservative profile (OR 0.75, 95% CI 0.59-0.96; P=.052). Those working in tertiary hospitals were less likely to belong to the Positive group (OR range 0.56-0.66, P=.001). High-income physicians were more likely to be in the Reform-Conservative group (OR range 1.83-2.38, P=.001). In addition, higher occupational stress was associated with a greater likelihood of Positive profile membership (OR range 1.12-1.26, P=.001), while better work satisfaction predicted higher odds of Positive profile membership (OR range 1.04-1.16, P=.001). CONCLUSIONS: This study introduces a novel, person-centered approach by identifying five distinct perceptual typologies among physicians, moving beyond traditional variable-centered analyses. This typology provides an evidence-based foundation for tailored interventions. For instance, the Reform-Adaptable group may need barrier reduction, while the Reform-Conservative group may require clearer value demonstrations. This nuanced understanding can help healthcare systems enhance the impact and scalability of digital health technologies in real-world clinical practice.