AI in Radiology and Diagnostic Accuracy: The Moderating Role of Radiologist Experience and Patient Load

Abstract

Artificial Intelligence (AI) has emerged as one of the most transformative technologies in modern healthcare, particularly in the field of radiology where image interpretation plays a critical role in disease diagnosis and treatment planning. AI powered radiological systems utilize machine learning, deep learning, and computer vision algorithms to analyze medical images, identify abnormalities, and support clinical decision making. These technologies have demonstrated significant potential in improving diagnostic accuracy, reducing interpretation errors, and enhancing healthcare efficiency. Diagnostic accuracy is essential for ensuring appropriate treatment decisions, minimizing medical errors, and improving patient outcomes. While AI technologies are increasingly integrated into radiological practice, their effectiveness may depend on contextual and human factors. Radiologist experience reflects the knowledge, expertise, and clinical competence accumulated through education and professional practice, whereas patient load refers to the volume of patients and diagnostic cases managed within a specific period. Experienced radiologists may be better equipped to interpret and validate AI generated recommendations, thereby maximizing diagnostic performance. Conversely, high patient load may increase time pressure and cognitive fatigue, potentially influencing the effectiveness of AI supported diagnostic processes. Therefore, this study investigates the relationship between AI in radiology and diagnostic accuracy while examining the moderating roles of radiologist experience and patient load. Drawing upon Sociotechnical Systems Theory and Human Capital Theory, the study proposes that AI in radiology positively influences diagnostic accuracy and that radiologist experience and patient load significantly moderate this relationship. A quantitative research design was employed using survey data collected from radiologists, radiology technicians, healthcare administrators, and medical imaging professionals. Data were analyzed using Partial Least Squares Structural Equation Modeling (PLS SEM) through SmartPLS software. The findings reveal that AI in radiology significantly improves diagnostic accuracy. Furthermore, radiologist experience strengthens, while patient load weakens, the relationship between AI in radiology and diagnostic accuracy. The study contributes to radiology, healthcare technology, and medical informatics literature by identifying key factors influencing AI effectiveness in clinical practice.