Where Is AI in Evidence Synthesis for NICE HTA Submissions?
Author(s)
Barinder Singh, RPh1, Karam Diaby, PhD2, Dilip Makhija, MS3, Deepika Thakur, PhD4, Amit Kulkarni, PhD5, Ritesh Dubey, PharmD1, Pankaj Rai, MS1.
1Pharmacoevidence, Mohali, India, 2Department of Pharmaceutical Sciences, University of Florida, Florida, FL, USA, 3Gilead Sciences Inc., Foster City, CA, USA, 4Hoffman La Roche AG, Mississauga, ON, Canada, 5Otsuka Pharmaceutical Development & Commercialization Inc, Princeton, NJ, USA.
1Pharmacoevidence, Mohali, India, 2Department of Pharmaceutical Sciences, University of Florida, Florida, FL, USA, 3Gilead Sciences Inc., Foster City, CA, USA, 4Hoffman La Roche AG, Mississauga, ON, Canada, 5Otsuka Pharmaceutical Development & Commercialization Inc, Princeton, NJ, USA.
OBJECTIVES: Interest in applying Artificial Intelligence (AI) in medical research and evidence generation has grown significantly in recent years. This surge in interest, along with potential impact of AI application, the National Institute for Health and Care Excellence (NICE) published a comprehensive position statement in August 2024, highlighting NICE’s expectations and guiding principles for the use of AI methods in the generation and submission of evidence within its evaluation programs. This study aimed to investigate the extent to which AI technologies have been utilized in NICE UK Health Technology Assessment (HTA) submissions, specifically for conducting systematic literature reviews (SLRs), over recent years.
METHODS: A comprehensive review of NICE HTA submissions from the past three years was conducted. Committee papers and accompanying guidance documents were systematically screened using relevant keywords to identify the use of AI or automation technologies in the context of SLRs. The identified included documents were reviewed to assess the scope, objective, and context of AI use in these SLRs.
RESULTS: A total of 288 technology appraisals (TAs) were screened in detail, none reported the use of AI, generative or otherwise, for data screening or systematic data collection within the evidence review process. Only two submissions mentioned incorporating AI in a limited capacity. TA1071 used an AI-assisted ad-hoc literature search to help identify potential treatment effect modifiers, while TA962 employed a Natural Language Processing model to assist with search query development by highlighting key terms within publications.
CONCLUSIONS: Although the 2024 NICE position paper acknowledges the potential of AI in evidence generation, its integration into HTA evidence reviews is currently lacking. This gap presents an opportunity to leverage AI to enhance the efficiency, rigour, and timeliness of future submissions, ultimately supporting more informed and accelerated decision-making.
METHODS: A comprehensive review of NICE HTA submissions from the past three years was conducted. Committee papers and accompanying guidance documents were systematically screened using relevant keywords to identify the use of AI or automation technologies in the context of SLRs. The identified included documents were reviewed to assess the scope, objective, and context of AI use in these SLRs.
RESULTS: A total of 288 technology appraisals (TAs) were screened in detail, none reported the use of AI, generative or otherwise, for data screening or systematic data collection within the evidence review process. Only two submissions mentioned incorporating AI in a limited capacity. TA1071 used an AI-assisted ad-hoc literature search to help identify potential treatment effect modifiers, while TA962 employed a Natural Language Processing model to assist with search query development by highlighting key terms within publications.
CONCLUSIONS: Although the 2024 NICE position paper acknowledges the potential of AI in evidence generation, its integration into HTA evidence reviews is currently lacking. This gap presents an opportunity to leverage AI to enhance the efficiency, rigour, and timeliness of future submissions, ultimately supporting more informed and accelerated decision-making.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
SA105
Topic
Study Approaches
Topic Subcategory
Literature Review & Synthesis
Disease
No Additional Disease & Conditions/Specialized Treatment Areas