Artificial Intelligence and Machine Learning Applied at the Point of Care

Area of Science:

• Digital Health
• Medical Informatics
• Regulatory Science

Background:

• Healthcare data and big data analytics are advancing Artificial Intelligence (AI) and Machine Learning (ML) applications.
• Despite advancements, point-of-care AI/ML implementations in medicine remain limited.

Purpose of the Study:

• To review AI/ML capabilities in healthcare settings.
• To analyze regulatory requirements for AI/ML in the US, Europe, and China.

Main Methods:

• A targeted narrative literature review of AI/ML digital health tools.
• Analysis of scientific publications and grey literature from regulatory agencies.

Main Results:

• AI/ML solutions are often classified as Software as a Medical Device (SaMD).
• Case studies include chronic disease management, diagnostic imaging support, and clinical decision support for treatment and precision dosing.
• Regulatory requirements encompass administrative, software, and clinical evidence, with variations in China regarding population applicability and third-party evaluation.

Conclusions:

• Promising AI/ML technologies are emerging, but widespread point-of-care adoption is hindered.
• Key obstacles include the need for external validation, implementation logistics, and data privacy concerns.