Acceptance of Artificial Intelligence Augmented Systematic Reviews by Health Technology Assessment Bodies
Speaker(s)
Umapathi K1, Nevis I2
1ICON plc, Tiruvannamalai, TN, India, 2ICON plc, Tucson, AZ, USA
Presentation Documents
OBJECTIVES: To review the acceptance of the use of artificial intelligence (AI) in the conduct of systematic reviews (SRs) among Health Technology Assessment (HTA) bodies.
METHODS: We conducted an environmental scan to obtain the list of national and regional HTA bodies by referring to "'INAHTA'—a network that connects HTA agencies," “’EunetHTA'— for HTA across Europe,” and "AdHopHTA_toolkit_tol24—the list of HTA agencies”. We visited the respective websites, to analyze the methodological guidelines for conducting a systematic review specifically focusing on the acceptance of AI.
RESULTS: We retrieved a list of 59 HTA bodies and reviewed the guidelines and recommendation section to understand the acceptance of the AI in the SR methodology. Only four HTA bodies and a network of HTAs provided some recommendations on the use of AI. National Institute for Health and Care Excellence's recommended priority screening techniques using AI to increase screening efficiency. The National center for Pharmacoeconomics stated the possibility of AI use in the Health Research Board-Collaboration in Ireland for clinical effectiveness reviews, but provided no information in the methodological guidelines, whereas the Scottish Medicine Consortium of Scotland refers to the NICE guidelines for evidence synthesis. Institute for Quality and Efficiency in Health Care recommends search algorithms for SRs of clinical evidence that make use of verified classifiers based on machine learning. Similarly, European Network for Health Technology Assessment's suggest the acceptance of AI/Machine Learning based classifiers in limiting literature search. But most of the HTA evidence review guidelines suggest that the manufacturer or authors who are submitting reports can refer to Cochrane guidelines for building evidence.
CONCLUSIONS: Currently there is very few guidelines and recommendations available on the use and acceptance of AI among HTA bodies. HTA agencies should start incorporating these strategies in their guidelines for systematic review methodologies.
Code
MSR70
Topic
Methodological & Statistical Research
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics
Disease
No Additional Disease & Conditions/Specialized Treatment Areas