Validity and transferability of Model for ASsessing the value of Artificial Intelligence (MAS-AI)

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Healthcare Decision-Making

Background:

• A multidisciplinary group developed the Model for ASsessing the value of AI (MAS-AI) in 2022.
• MAS-AI aids decision-makers in prioritizing AI solutions in medical imaging.
• Assessing AI's value is critical for informed healthcare technology adoption.

Purpose of the Study:

• To evaluate the face validity and transferability of the MAS-AI tool.
• To investigate perceptions of MAS-AI's content importance among workshop participants.
• To assess the applicability of MAS-AI across different international healthcare contexts.

Main Methods:

• A Delphi process was employed, incorporating insights from four workshops.
• Participants included decision-makers from hospitals and the healthcare sector, patient partners, researchers, and experts.
• Participants rated the importance of MAS-AI domains and subtopics using a 0-3 Likert scale.

Main Results:

• 95 participants from Denmark, Italy, and Canada contributed to the study.
• Face validity of all MAS-AI domains was confirmed, with over 70% rating them as important.
• All subtopics were rated above the 70% threshold, except for five in Italy.

Conclusions:

• The MAS-AI tool demonstrated confirmed validity in Denmark, Italy, and Canada.
• The study identified areas for improvement in study design and data collection.
• Future research should analyze participant-specific ratings for deeper insights into AI value assessment.