This paper examines algorithmic bias in machine learning and artificial intelligence applications within healthcare information systems. Using quantitative data and qualitative comments gathered from surveys and interviews with healthcare professionals across various roles (e.g., clinical versus administrative), the study investigates the relationships among algorithmic bias, perceived fairness, and the willingness to accept and adopt ML algorithms and their outcomes.
Findings reveal that professionals' opinions on the causes of bias, assessment criteria, perceived fairness, and mitigation approaches vary according to job role, perspective, tasks, and algorithm characteristics. Clinicians tend to prioritize diagnostic accuracy and patient safety, while administrators focus on workflow efficiency and cost-effectiveness.
The authors call for more research to uncover bias sources and to develop role‑specific fairness evaluation frameworks and mitigation strategies, aiming to ensure fairness and equality in healthcare.
Review