How do medical institutions co-create artificial intelligence solutions with commercial startups?

Area of Science:

• Medical Informatics
• Artificial Intelligence in Radiology
• Health Technology Co-creation

Background:

• Radiology departments increasingly adopt artificial intelligence (AI) solutions.
• Commercial AI applications often do not align with specific departmental needs, necessitating co-creation.
• The process of co-creating AI solutions in healthcare is under-researched, lacking guidance frameworks.

Purpose of the Study:

• To examine the process of co-creating AI solutions in radiology.
• To identify challenges and develop a framework for monitoring, assessing, and guiding collaborations between medical and technological parties.
• To provide insights into effective co-creation strategies for AI in clinical practice.

Main Methods:

• A multi-case study of three co-creation projects at a Dutch academic medical center.
• Analysis of co-creation processes based on levels of "resourcing," "adaptation," and "reconfiguration."
• Development of a relational framework to guide and assess AI co-creation collaborations.

Main Results:

• A relational framework was developed to monitor, assess, and guide AI co-creation collaborations.
• The framework facilitates the discovery of novel use-cases and reassessment of existing assumptions and practices.
• Distinct co-creation journeys with varied outcomes were identified and mapped onto the framework for diagnosis and guidance.

Conclusions:

• Co-creation outcomes in AI development can be highly variable.
• The proposed framework assists medical institutions and technology companies in assessing challenges and adjusting collaborations.
• This framework helps steer the co-creation process toward desired clinical and technological goals.