Not yet human-reviewed
This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.
A comprehensive overview of barriers and strategies for AI implementation in healthcare: Mixed-method design
Citation. Nair, M., Svedberg, P., Larsson, I., & Nygren, J. M. (2024). A comprehensive overview of barriers and strategies for AI implementation in healthcare: Mixed-method design. PLOS ONE, 19(8), e0305949. https://doi.org/10.1371/journal.pone.0305949
Summary. This mixed-method study inventories barriers to and strategies for implementing AI systems in healthcare across planning, implementation, and sustained-use phases. It synthesizes 38 empirical cases drawn from six reviews and uses direct qualitative content analysis of 69 interviews with healthcare leaders and professionals to confirm and extend the findings. The resulting framework identifies organizational, sociotechnical, legal, data, and maintenance conditions that shape whether AI enters routine clinical practice.
Key claims
- Barriers and strategies span eleven concepts—leadership, buy-in, change management, engagement, workflow, finance and human resources, legal, training, data, evaluation and monitoring, and maintenance—while the interview analysis added ethics as a twelfth concept.
- AI implementation needs process-specific planning and ongoing support rather than a one-time technical deployment, because barriers and strategies differ across planning, implementation, and sustained use.
- Buy-in depends on locally relevant problems, leadership support, usable workflow integration, opportunities for clinicians to test system quality, and clarity about clinician liability and final decision authority.
- Evaluation and monitoring should include staged pilots, baseline and comparison measures where useful, patient-safety risk and impact analysis, and cost-benefit assessment before scaling.
Evidence & limitations
- The study combines an overview of six scoping/systematic reviews with inductive coding of 38 empirical implementation cases and direct qualitative content analysis of 69 interviews with healthcare leaders and professionals.
- The evidence concerns healthcare implementation and draws on heterogeneous underlying cases; its inventory identifies recurring barriers and strategies rather than estimating causal effects or establishing that any individual strategy will transfer unchanged to other sectors.
- This is an open-access, peer-reviewed PLOS ONE article published under CC BY 4.0.