Use of AI Methods in NICE Technology Appraisals: What Is the Current Status Quo?
Author(s)
Luis Val Maranes, MSc, Rikal Bhaila, MPH, Louise Heron, MSc, Fiona Pearson, PhD, Zoe Blumer, PhD.
Adelphi Values PROVE™, Bollington, United Kingdom.
Adelphi Values PROVE™, Bollington, United Kingdom.
OBJECTIVES: Artificial Intelligence (AI) is playing an increasingly prominent role in health economic (HE), systematic review (SR) and evidence synthesis (ES) methodology. This raises important questions about its implementation within health technology assessment (HTA). As HTA agencies like National Institute for Clinical Excellence (NICE) and Canada’s Drug Agency (CDA-AMC), release position statements aiming to guide its use it is important to understand the status quo alongside the transparency, validity, and impact of its application.
METHODS: A scoping review (ScR) was conducted following the JBI guidance and reported following the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for ScR. All completed Technology Appraisals (TAs) and Highly Specialised TAs published on the NICE website from 28th May 2024 to 6th June 2025 were included in this scoping review. Data were to be extracted by a single reviewer using a pre-specified form with an independent check by a second reviewer and findings were to be narratively and visually summarised.
RESULTS: Of the 90 TA reports published by NICE over the last year, none reported on or discussed AI use in the conduct of HE (e.g. to dynamically adapt simulation models), SR (e.g. machine learning to facilitate screening) or ES methods (e.g. data extraction using Natural Language Processing).
CONCLUSIONS: No NICE TA report explicitly mentions use of AI in HTA. Recently published Cochrane guidance and position statements from NICE and CDA-AMC signal a growing acceptance of AI methodology and commitment to incorporate use within HTA. However, currently established rather than innovative approaches are being maintained. Broader and more consistent integration of AI in HE, SR, or ES methods could be supported by clear guidance, standards for practice and case examples. This would help to ensure AI is applied within HTA effectively, consistently, ethically and its use reported upon transparently.
METHODS: A scoping review (ScR) was conducted following the JBI guidance and reported following the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for ScR. All completed Technology Appraisals (TAs) and Highly Specialised TAs published on the NICE website from 28th May 2024 to 6th June 2025 were included in this scoping review. Data were to be extracted by a single reviewer using a pre-specified form with an independent check by a second reviewer and findings were to be narratively and visually summarised.
RESULTS: Of the 90 TA reports published by NICE over the last year, none reported on or discussed AI use in the conduct of HE (e.g. to dynamically adapt simulation models), SR (e.g. machine learning to facilitate screening) or ES methods (e.g. data extraction using Natural Language Processing).
CONCLUSIONS: No NICE TA report explicitly mentions use of AI in HTA. Recently published Cochrane guidance and position statements from NICE and CDA-AMC signal a growing acceptance of AI methodology and commitment to incorporate use within HTA. However, currently established rather than innovative approaches are being maintained. Broader and more consistent integration of AI in HE, SR, or ES methods could be supported by clear guidance, standards for practice and case examples. This would help to ensure AI is applied within HTA effectively, consistently, ethically and its use reported upon transparently.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
HTA349
Topic
Health Technology Assessment
Topic Subcategory
Systems & Structure
Disease
No Additional Disease & Conditions/Specialized Treatment Areas