Can AI reliably replace human reviewers in literature reviews?
The EU Joint Clinical Assessment (JCA) is expected to require considerable time and resources from both countries and manufacturers, especially in developing systematic literature reviews (SLRs) for dossier submissions. This increase in SLR work requires a new approach to manage the volume and resource requirements under short timelines. Our whitepaper examines whether current AI technology can help address this growing workload.
We tested ChatGPT, a large-language model, on several key tasks associated with SLRs, including defining PICO criteria, developing search strategies, screening abstracts and full texts, and extracting data. Our study evaluated the accuracy of its performance on these tasks and considered whether it could replace human efforts or reduce their workload.
Download the whitepaper here to learn more about these findings.
About the authors
Manca has over 9 years’ experience in Health economic outcomes research and market?access consultancy across global and local markets. She specializes in innovative approaches?to value demonstration, including how best to use new and emerging technologies to expedite?market access
Emma is a health economic outcomes research and market access consultant, with experience?conducting SLRs across a range of disease areas. She has a particular interest in how artificial?intelligence can be leveraged to assist in the SLR process, particularly in the context of the evolving?landscape of health technology assessment.