Presenting artificial intelligence, deep learning, and machine learning studies to clinicians and healthcare stakeholders: an introductory reference with a guideline and a Clinical AI Research (CAIR) checklist proposal

Area of Science:

• Orthopedic and Medical Research
• Artificial Intelligence (AI)
• Machine Learning (ML)
• Deep Learning (DL)

Background:

• AI, ML, and DL are increasingly used in medical research, particularly orthopedics.
• A gap exists between AI engineers and healthcare professionals due to differing expertise and goals.
• Effective communication and knowledge exchange are crucial for advancing medical AI applications.

Purpose of the Study:

• To provide clinicians with a foundational understanding of AI research in medicine.
• To facilitate better communication between AI creators, researchers, clinicians, and readers of medical AI literature.
• To equip healthcare professionals with the knowledge to critically evaluate AI research.

Main Methods:

• Discuss common tasks, methodological considerations, and ethical pitfalls in AI research.
• Cover key topics such as data labeling, missing data handling, model training, testing, and overfitting.
• Present standard performance metrics (e.g., accuracy, precision, recall, F1, Dice, AUC, ROC curves) and ethical considerations (privacy, fairness, autonomy, safety, liability).

Main Results:

• Developed guidelines for reporting medical AI research tailored for clinicians.
• Introduced a Clinical Artificial Intelligence Research (CAIR) checklist.
• Proposed specific performance metrics guidelines for evaluating AI components in research.

Conclusions:

• The CAIR checklist and performance metrics guidelines aim to improve the clarity and evaluation of medical AI research.
• These resources are intended for researchers, engineers, clinicians, and stakeholders to enhance understanding and application of AI in healthcare.
• The guidelines support a broader consensus process for standardized reporting of AI research in medical settings.