The algorithm journey map: a tangible approach to implementing AI solutions in healthcare

Area of Science:

• Health Informatics
• Artificial Intelligence in Medicine
• Clinical Decision Support Systems

Background:

• Integrating AI/ML in healthcare presents challenges due to complex user-technology interactions.
• Lack of understanding hinders AI/ML adoption and the development of best practices for AI solutions.

Purpose of the Study:

• To document the complete lifecycle of an AI solution, SepsisWatch, within a healthcare system.
• To identify generalizable insights for overcoming barriers to AI/ML adoption in healthcare.
• To inform best practices for operationalizing ethical principles in healthcare AI systems.

Main Methods:

• Conducted 20 interviews with engineers and scientists involved in the SepsisWatch project.
• Developed an "Algorithm Journey Map" detailing social and technical activities.
• Documented the AI solution's procurement, development, integration, and lifecycle management.

Main Results:

• The "Algorithm Journey Map" provides a comprehensive overview of the AI solution's lifecycle.
• Identified key "lessons learned" including modeling assumptions, stakeholder engagement, and organizational structure.
• Highlighted generalizable insights for recognizing and navigating AI/ML adoption barriers in healthcare.

Conclusions:

• The study provides a practical framework for understanding and managing AI/ML integration in healthcare.
• Findings contribute to the development of best practices for sustainable and ethical AI implementation.
• The "Algorithm Journey Map" serves as a valuable tool for healthcare organizations adopting AI solutions.