Speaker
Description
Authors: Pamela Buckley (University of Colorado Boulder), Diana Fishbein (National Prevention Science Coalition to Improve Lives; Frank Porter Graham Child Development Institute, University of North Carolina-Chapel Hill; Human Development and Family Studies, The Pennsylvania State University), Neil J. Wollman Fishbein (National Prevention Science Coalition to Improve Lives)
Background: Evidence-based decision-making applies empirical evidence to inform policies and involves integrating relevant information from various sources. Online clearinghouses support evidence-based decision-making by synthesizing evidence on what works, though manually updating the literature is incomplete. In addition, passively summarizing evaluations is insufficient for end-users to implement preventive solutions that achieve population impacts. The design of clearinghouses can significantly enhance evidence-based decision-making by building in stepwise, interactive, artificial intelligence (AI)-driven capabilities that augment human expertise and harness machine learning for increased efficiency and comprehensiveness of evidence synthesis.
Methods: We propose a two-part conceptual framework for a clearinghouse platform. First, clearinghouses should adopt a “living” systematic review wherein evaluation summaries get automatically updated and relevant evidence is incorporated into the review. Living evidence has already been embraced globally, with the World Health Organization, Cochrane Collaboration, and United Nation’s Pan American Health Organization all committing to this approach. The second component involves adding a chatbot search engine to support assessment and implementation guidance within a clearinghouse and/or network of clearinghouses. Central to this two-part framework are human touchpoints to avoid unintended negative consequences of AI such as biases and inaccuracies.
Results: AI algorithms can make recommendations vetted by prevention scientists for (1) the provision of all evidence-based preventive interventions (EBPIs) and their key activities, (2) EBPIs shown to achieve equitably distributed outcomes, (3) culturally relevant EBPIs that align with and respect the cultural beliefs, practices, and needs of a target population, (4) implementation support, such as materials, training, and fidelity measures, and (5) delivery costs.
Discussion: Though only in a conceptual phase of development, we present a two-part framework for AI to enhance the speed, scope, and relevance of clearinghouse functionality. We are confident that the resulting platform will lead to more accessible evidence on effective preventive strategies.
| Conflict of interest | None |
|---|
