Knowledge, attitude, and practice of primary care physicians toward clinical AI-assisted digital health technologies: Systematic review and meta-analysis.

BACKGROUND

The landscape of digital health technologies is evolving rapidly, with clinical artificial intelligence increasingly integrated into primary care. Successfully adopting these technologies depends on the users' knowledge, attitude, and practice.

AIM

This systematic review and meta-analysis aims to assess primary care physicians' knowledge, attitude, and practice toward clinical artificial intelligence and to uncover the key determinants influencing its implementation in primary care.

METHODS

PubMed, Web of Science, Scopus, and Institute of Electrical and Electronics Engineers (IEEE) were searched on 18.10.2023 and 03.05.2024 to systematically review quantitative and qualitative relevant primary studies. Three authors independently reviewed and appraised the studies using the Mixed Methods Appraisal Tool. Thematic analysis and proportion meta-analysis of the addressed domains were performed, with results aligned with a recent integration framework.

RESULTS

24 publications, including 4074 primary care physicians, suggested that knowledge levels were generally low, with passive opportunistic learning (pooled proportion 0·33, 95 % Confidence Interval (CI) 0·16-0·50, n = 6 studies, 2358 physicians). Attitudes varied, with concerns about losing jobs and rejecting new technologies (0·53, 95 %CI 0·42-0·64, n = 11, 2988). Practice experience was positive with AI simulation/prior training or negative with infrastructure and electronic medical records limitations (0·52, 95 %CI 0·36-0·68, n = 12, 3459). The risk of bias was low in 14 studies and moderate-high in ten, with significant heterogeneity between studies.

CONCLUSION

This review underscores the importance of effectively integrating clinical artificial intelligence-assisted digital health technologies within primary care. Acknowledging the current knowledge, attitude, and practice state and identifying gaps and opportunities, a physician-driven artificial intelligence implementation process with sustainable adoption might be possible. More attention is needed to counterbalance the concerns hindering the effectiveness of advanced tools in primary care practice.