RESEARCH · RESEARCH · #2872
Preprint: Algorithmic Bias from the Perspectives of Healthcare Professionals
This arXiv preprint reports survey and interview data from healthcare professionals (clinical and administrative roles) about algorithmic bias in ML/AI-driven healthcare information systems. The study finds that views on causes of bias, assessment criteria, perceived fairness, likely acceptance, and mitigation approaches vary by job role, task, and algorithm characteristics, and calls for further research to ensure fairness in healthcare.
KEY POINTS
- This arXiv preprint reports survey and interview data from healthcare professionals (clinical and administrative roles) about algorithmic bias in ML/AI-driven healthcare information systems.
- The study finds that views on causes of bias, assessment criteria, perceived fairness, likely acceptance, and mitigation approaches vary by job role, task, and algorithm characteristics, and calls for further research to ensure fairness in healthcare.
- Healthcare professionals' differing views on bias and fairness can affect assessment, adoption, and mitigation of ML/AI tools in clinical settings, shaping real-world impacts on care equity.
WHY IT MATTERS
Healthcare professionals' differing views on bias and fairness can affect assessment, adoption, and mitigation of ML/AI tools in clinical settings, shaping real-world impacts on care equity.