Artificial intelligence (AI) can play a role in systematic literature reviews but should not replace human involvement, according to an analysis presented at the AMCP 2025 Annual Meeting. This study was included in part of the Meeting’s Moderated Poster Tours.
Systematic literature reviews are an unbiased process for assessing efficacy, safety, and pharmacoeconomic evidence, although they are resource-intensive endeavors, so the researchers sought to understand how AI may be an opportunity for improvement.
A targeted literature review was conducted in Embase and MEDLINE from January 2019 to October 2024 for terms such as AI, natural language processing, large language model, and machine learning combined with terms from reimbursement authorities. The researchers then developed a survey that was sent to reimbursement decision-makers in Europe and the United States, including members of the Institute for Clinical and Economic Review (ICER) and a health plan, to determine how reimbursement authorities regard the use of AI in systematic literature reviews.
The targeted literature review found that most reimbursement authorities do not mention the use of AI for systematic literature reviews, except for the National Institute for Health and Care Excellence and the Institute for Quality and Efficiency in Health Care.
Most survey respondents were somewhat familiar with the use of AI in systematic literature reviews and thought AI could improve efficiency.
In the United States, the level of familiarity was greater with ICER than with the health plan. However, reimbursement decision-makers were not convinced that AI would improve the quality of systematic literature reviews and thought manufacturers should not be primarily responsible for developing or validating AI tools for systematic literature reviews.
Transparency and reliability were reportedly key factors to consider when incorporating AI in systematic literature reviews for reimbursement decision-making.
“A multistakeholder effort will be required to ensure quality and realize efficiency gains associated with AI-supported [systematic literature reviews] for reimbursement decision-making,” the authors concluded.
Reference
Fox G, Torres Ames J, Santpurkar N, Arca E, Nass P. Artificial intelligence–supported evidence synthesis for reimbursement decision-making. Abstract U7 presented at: AMCP 2025; March 31-April 3, 2025; Houston, TX.
