2025/12/19 by Gareth J. Hollands, Ian Shemilt, James Thomas · 1 voice
Social Sciences · Health Professions · Environmental Science · #Computational and Text Analysis Methods #Health Policy Implementation Science #Sustainability and Climate Change Governance
paper · doi:10.31235/osf.io/vey6q_v1
openalex created_date 2025/12/19 · openalex publication_date 2025/12/19 · openalex updated_date 2026/07/14
Producing useful evidence syntheses requires clearly formulated research questions linked to an understanding of their optimal scope. When policymakers (or other end users) commission or steer the production of systematic reviews to support policy development, these parameters are rarely specified sufficiently for production to proceed. Instead, these details are typically developed via a collaborative dialogic process involving evidence producers and users to ensure the proposed synthesis will meet users’ needs. This process is informed by scoping and (more-or-less formally) mapping the existing evidence base, which can take many weeks or months, slowed by applying mostly manual processes to identify and characterise the evidence. Recent developments in artificial intelligence (AI) technologies have opened up new and possibly transformative opportunities for far more rapid, iterative, and responsive mapping of relevant evidence to support such early-stage evidence-to-policy dialogues and processes. In this article, we outline and characterise this approach, termed dialogic rapidly expedited evidence maps (DREEMs). We also present two illustrative case studies drawn from first-hand experiences within a programme of evidence reviews production to inform policy in England.