The willingness to adopt artificial intelligence-driven Clinical Decision Support Systems among doctors at different hospitals in China: A fuzzy set qualitative comparative analysis of survey data (Preprint)

Abstract

Artificial intelligence (AI) driven Clinical Decision Support Systems (CDSS) play an important role in assisting doctors in diagnosis and treatment and in improving the efficiency and quality of medical services. However, not all doctors trust AI technology, and many remain sceptical and unwilling to adopt these systems.

Our study’s aim is to explore in depth the factors influencing different doctors’ intentions to adopt AI-CDSS, along with the causal relationships among these factors, to gain a better understanding to promote their clinical application and widespread implementation.

Based on the Unified Theory of Acceptance and Use of Technology (UTAUT) and Technology-Organisation-Environment (TOE) models, we propose and design a framework for doctors’ willingness to adopt AI-CDSS. We conducted a nationwide questionnaire survey in China and fuzzy set qualitative comparative analysis (fsQCA) to explore the willingness of doctors in different types of medical institutions to adopt AI-CDSS along with the factors influencing their willingness.

The survey was distributed to doctors from different medical institutions in China, categorised as tertiary hospital and primary/secondary hospital. We distributed 578 questionnaires and received 450 valid responses with good reliability and validity. The effective response rate was 77.85%. We analyse the influencing factors and pathways of willingness from three perspectives: the technology, the organisation, and the individual. Doctors in tertiary hospitals were found to have six pathways to AI-CDSS adoption willingness, categorised as technology, individual, and technology-individual dual-driven; doctors at primary/secondary hospital had three pathways to AI-CDSS adoption willingness, categorised as technology-individual dual-driven and organisation-individual dual-driven. There were both commonalities and differences in the pathways among the doctors at the different medical institutions in terms of the factors influencing AI-CDSS adoption. Among the commonalities, AI technology and individual doctor factors played a dominant role in the willingness to adopt, implying that all the doctors believed that AI-CDSS could provide efficient diagnostic and treatment support; thus, they were willing to accept and try these new technologies. Among the differences, conditions of convenience (namely, facility support and resources) had a greater impact on doctors at primary/secondary hospital. For these doctors, only sufficient support would encourage their active adoption of AI-CDSS.

Conclusions: From the perspective of the six configurations among the doctors at the tertiary hospitals and the three configurations among the doctors at primary/secondary hospital, performance expectancy and individual innovativeness were two indispensable and core conditions in the pathways to achieving a favourable intention to adopt AI-CDSS.