NICE and CDA Assessment of Artificial Intelligence-Enabled Health Technologies: Analysis and Insights
Author(s)
Saeed Anwar, MPharm1, Raju Gautam, PhD2, Ratna Pandey, MSc1, Tushar Srivastava, MSc2.
1ConnectHEOR, Delhi, India, 2ConnectHEOR, London, United Kingdom.
1ConnectHEOR, Delhi, India, 2ConnectHEOR, London, United Kingdom.
OBJECTIVES: The growing interest in artificial intelligence (AI)-enabled technologies in healthcare presents novel opportunities and challenges for health technology assessment (HTA) agencies. The UK’s National Institute for Health and Care Excellence (NICE) and Canada’s Drug Agency (CDA) are at the forefront of evaluating such innovations. However, the unique characteristics of AI such as algorithmic learning, data dependency, and limited clinical validation necessitate adapted or novel assessment frameworks. This study aims to analyze and provide insights on HTA submitted to NICE and CDA for AI-enabled technologies.
METHODS: The websites of NICE and CDA were searched to identify HTA submissions published up to May-2025. Data was extracted on disease indication, utility of AI-enabled technology, HTA decision drivers, key critiques, and final recommendations.
RESULTS: The searches identified only six HTA submissions (NICE, n=5; CDA, n=1). All HTAs were for AI-enabled software used for detection of diseases (cancer detection, n=3; stroke, n=2; fracture detection, n=1). Majority of assessments were early value assessments (68%). Most appraisals resulted in conditional recommendations (5/6, 85%), if the evidence outlined in the evidence generation plan is being generated and after appropriate regulatory approval including NHS England's Digital Technology Assessment Criteria (DTAC) approval; none was fully recommended, and one was not recommended. The main HTA decision drivers were unmet clinical need, time-sensitive decision making, diagnostic accuracy of the technology, clinical evidence, etc. The key critiques or limitations were related to AI-technology (n=3), such as uncertainty around methods used to develop AI models, AI accuracy, uncertainty around AI algorithms, etc.
CONCLUSIONS: The submission of AI-enabled health technologies remains very limited. The significance of AI-enabled technology assessment can be expected to gradually increase as the HTA agencies begin to update their methodological guidelines with novel assessment frameworks for these highly innovative technologies.
METHODS: The websites of NICE and CDA were searched to identify HTA submissions published up to May-2025. Data was extracted on disease indication, utility of AI-enabled technology, HTA decision drivers, key critiques, and final recommendations.
RESULTS: The searches identified only six HTA submissions (NICE, n=5; CDA, n=1). All HTAs were for AI-enabled software used for detection of diseases (cancer detection, n=3; stroke, n=2; fracture detection, n=1). Majority of assessments were early value assessments (68%). Most appraisals resulted in conditional recommendations (5/6, 85%), if the evidence outlined in the evidence generation plan is being generated and after appropriate regulatory approval including NHS England's Digital Technology Assessment Criteria (DTAC) approval; none was fully recommended, and one was not recommended. The main HTA decision drivers were unmet clinical need, time-sensitive decision making, diagnostic accuracy of the technology, clinical evidence, etc. The key critiques or limitations were related to AI-technology (n=3), such as uncertainty around methods used to develop AI models, AI accuracy, uncertainty around AI algorithms, etc.
CONCLUSIONS: The submission of AI-enabled health technologies remains very limited. The significance of AI-enabled technology assessment can be expected to gradually increase as the HTA agencies begin to update their methodological guidelines with novel assessment frameworks for these highly innovative technologies.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
MSR156
Topic
Health Technology Assessment, Methodological & Statistical Research
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics