The physician-AI relationship: partnering for precision medicine via clinical decision support

Abstract Background Clinical decision support systems (CDSSs) can enhance physician decision-making, but successful adoption depends on positive user attitudes. Understanding Israeli physicians’ attitudes toward CDSSs is crucial for overcoming implementation barriers and refining training. Methods Semi-structured interviews explored knowledge, experiences, perceived benefits/drawbacks, and potential barriers/facilitators to CDSS use among 25 Israeli physicians across specialties, roles, settings, and experience levels selected via stratified sampling. Individual 45-minute video interviews were conducted in January-July 2023. Results Familiarity and CDSS experience varied substantially across medical specialties. While risk calculators, information systems, and expert systems disseminating contemporary research insights are widespread, AI-based system adoption is relatively low outside radiology. Physicians expressed positive attitudes, recognizing CDSSs’ potential to enhance clinical work, reduce errors, and alleviate burdens. However, concerns included AI systems’ impact on workload, patient relationships, skill erosion, and overreliance. Key implementation barriers involved technological infrastructure, organizational processes, physician perceptions, and system features. Facilitating factors included tailored design, focused training, continuous accuracy/reliability validation, and clear autonomy/liability guidelines. Conclusions Addressing key concerns around transparency, trustworthiness and responsible use is crucial for AI-CDSS to fulfill its promise of enhancing quality and outcomes. Insights from Israeli clinicians provide guidance on priorities like transparency of data and algorithms as well as ensuring AI behaves ethically. Collaboration among technology designers, healthcare professionals and stakeholders can help develop systems that physicians trust to improve care while safeguarding patients. Ongoing evaluation is also needed to optimize AI-CDSS over time. Key messages • Understanding physician attitudes is key to overcoming barriers to clinical decision support system adoption and optimizing training protocols. • Addressing transparency, trustworthiness, and ethical considerations around AI clinical decision support systems is crucial for enhancing quality care.