JMIR Publications

ABSTRACT

Background:

The integration of artificial intelligence (AI) into healthcare systems is fundamentally transforming medical practices. Understanding medical staff’s attitudes toward AI adoption is pivotal for developing effective implementation strategies.

Objective:

To investigate attitudes and perceptions regarding medical AI among doctors and nurses in China, and to identify the influencing factors of adoption.

Methods:

A national cross-sectional survey was conducted online in China from December 12 to 26, 2024. Participants were recruited from the Chinese Medical Association and the Chinese Nursing Association. The structured questionnaire assessed demographic characteristics, knowledge and attitudes about medical AI, experiences and insights in using medical AI, and perceptions and factors influencing the adoption of AI in healthcare based on the Unified Theory of Acceptance and Use of Technology (UTAUT) model. Multiple linear regression and Karlson-Holm-Breen (KHB) mediation analysis were employed to identify influencing factors. Sample weighting by regional distribution was applied for sensitivity analysis.

Results:

The survey included 991 doctors and 1714 nurses. 92.4% (916/991) of doctors and 84.2% (1443/1714) of nurses reported awareness of medical AI application, only 22.8% (226/991) of doctors and 17% (291/1714) of nurses had practical experience. 82.6% (819/991) of doctors and 80.2% (1375/1714) of nurses held optimistic views about AI’s prospects. After adjusting for covariates, performance expectancy (doctors: β=0.144, 95%CI [0.092 - 0.197]; nurses: β=0.292, 95%CI [0.245 - 0.338]), effort expectancy (doctors: β=0.681, 95%CI [0.562 - 0.800]; nurses: β=0.440, 95%CI [0.342 - 0.538]), social influence (doctors: β=0.264, 95%CI [0.187 - 0.341]; nurses: β=0.098, 95%CI [0.045 - 0.152]), and facilitating conditions (doctors: β=0.098, 95%CI [0.030 - 0.165]; nurses: β=0.158, 95%CI [0.105 - 0.212]) had significant positive impacts on willingness to use. Perceived risk showed no significant effect on doctors’ intention (β=0.012, 95%CI [-0.022 - 0.045]), but negatively impacted nurses’ intention (β=-0.041, 95%CI [-0.066 - -0.015]). Performance expectancy and effort expectancy partially mediated the relationship between facilitating conditions and intention to use. Age, educational level, hospital grade, work experience duration, and personal views also significantly influenced willingness. Weighted analysis results were consistent with unweighted analysis, confirming the robustness of research findings.

Conclusions:

Significant disparities exist between high willingness and low utilization of medical AI among Chinese healthcare professionals. System optimization focusing on utility enhancement, workflow integration, and risk mitigation for medical staff, along with support for infrastructure and training, could accelerate AI adoption in clinical practice.