Comparing general-purpose and domain-specific AI in confounder identification for HTA submissions
The increasing use of AI in health technology assessment (HTA) raises important methodological and regulatory questions, particularly where transparency, reproducibility, and evidentiary standards are central to decision-making. One such area is systematic confounder identification, a mandatory and labor-intensive step in benefit assessments conducted by HTA bodies such as the German Institute for Quality and Efficiency in Healthcare (IQWiG).
The study, “General-purpose versus domain-specific AI for systematic confounder identification in multiple sclerosis: A comparative methodological study using IQWiG assessments as ground truth,” presented at ISPOR 2026, examines the performance of general-purpose large language models compared with a domain-specific retrieval-augmented system.
In this interview, Florian Woeste, Co-Founder and Managing Director of PHAROS Labs GmbH, explores this, asking one of the most pressing methodological questions in HTA: how can AI be used to improve the efficiency of systematic confounder identification without compromising transparency and evidentiary rigor?
For readers working in HTA, HEOR, market access, or regulatory evidence generation, this interview provides a timely look at where AI may deliver genuine value and where caution is still required.
Interviewee
Florian Woeste
Co-Founder & Managing Director PHAROS Labs GmbH
Florian Woeste is Co-Founder and Managing Director of PHAROS Labs GmbH, a Hamburg-based health technology company developing Regulaido, an AI-powered platform for medical writing, regulatory evidence generation and systematic literature research in the pharmaceutical and biotech sector. Florian has a bachelor’s degree in economics and software development with a master’s degree in AI.
Disclaimers
The opinions expressed in this feature are those of the author and do not necessarily reflect the views of The Evidence Base® or Becaris Publishing Ltd.
Sponsorship for this Peek Behind the Poster was provided by PHAROS Labs GmbH.