New publication from AI-Mind on Framework for Evaluating AI-Based Health Technologies
The AI-Mind project is excited to announce the publication of “Health Technology Assessment Framework for Artificial Intelligence-Based Technologies” by partners from The Graduate School of Health Economics and Management (ALTEMS) in the International Journal of Technology Assessment in Health Care. This research is a significant advancement in evaluating AI technologies in healthcare, particularly in ensuring they are safe, effective, and ethically sound and the outcome of the activities within the work package 6 (WP6) led by Rossella di Bidino.
Developed with input from 47 experts through a two-round Delphi survey, the study prioritizes critical factors in assessing AI tools. It emphasizes the importance of addressing bias in data (85%), model accuracy (88%), and trustworthiness (85%) to ensure AI-driven technologies meet the highest standards of reliability. Among traditional domains from the EUnetHTA Core Model, clinical effectiveness (82%) and ethical considerations (81%) were rated most critical, followed by cost-effectiveness (77%). Beyond traditional HTA domains, additional areas like data transparency, algorithm appropriateness, and robustness emerged as crucial. These findings suggest that current HTA models require updates to accommodate AI’s unique characteristics.
We asked our expert, what was the most challenging of this task.
The primary challenge was closely linked to the survey’s objective, which goes beyond the development of an HTA framework. The overarching aim is to establish and promote a multidisciplinary approach to AI in healthcare. This requires not only fostering a shared language among diverse experts and stakeholders but also cultivating an understanding of the interdependence and needs of each of them. As a result, the key difficulty was identifying the full spectrum of competencies involved in AI development, adoption, and assessment, and thoroughly exploring which topics hold the greatest relevance for each area.
Relevance to AI-Mind and the next steps
This publication supports AI-Mind’s goal of developing AI tools for early diagnosis and risk stratification in neurodegenerative disorders. The framework ensures that technologies developed by AI-Mind are assessed rigorously, addressing not only their clinical utility but also their ethical and societal impacts. As AI-based healthcare tools grow more complex, this framework provides a foundation for transparent and effective evaluation, aligning AI innovation with patient-centred care.
To learn more, access the full publication here: Cambridge Core.