ISSUE:

• Unlike other industries, use of AI/ML in HTA is widely considered to be in its infancy
• HTA bodies are beginning to recognize the value of AI/ML, such as in the selection of covariates in survival models, but use cases are few
• Existing concerns around the use of AI/ML in HTA include questions on appropriateness, transparency, and accessibility to non-experts
• However, AI/ML methods have proven advantages over traditional approaches in cases of multi-variable datasets often seen in healthcare
• How then can the benefits of these methodologies be integrated with concerns of decision makers for a balanced approach in HTA and HEOR?

OVERVIEW:

(5 min) Moderator Ines Guerra will introduce the topic of AI/ML in HTA, explain the current situation and important background. She will outline the advantages and disadvantages of AI/ML methods and key vocabulary for the audience.

(10 min) Pearl Gumbs (Industry) will build on this to explain how AI/ML may benefit manufacturers in value demonstration, noting that clear guidance on acceptability of AI/ML is required to reduce rejection risk.

(10 min) In response, Mariam Sood (Payer) will acknowledge the positives of AI/ML in improving process timeliness but explain concerns for AI/ML in HTA. She will detail what needs to change for better applicability in HTA and conclude with a future perspective on use of AI/ML methods from the view of NICE.

(10 min) Finally, Noemi (Academic) will explain how typical challenges of AI/ML (like lack of transparency) can be overcome, and the role of academia in this advancement. She will explain how academic innovation will help bridge Industry and decision makers to embed AI/ML in HTA.

(25 min) Facilitated audience discussion, with opportunities to ask questions and propose solutions to accelerate use of AI/ML in HTA

Stakeholders: AI/ML developers, Payers, Industry HEOR/RWE, Market Access professionals