Big Data/Machine Learning/AI
Leveraging artificial intelligence to gather epidemiologic data on infectious diseases: the Global Repository of Epidemiological Parameters (GREP) project Melanie Sterian* Melanie Sterian Kusala Pussegoda Tricia Corrin Bing Hu Emmalie Tomini Nauman Shakeel Lizaveta Vasileuskaya Stephanie Brazeau Lisa Waddell
Background:
The emergence of new infectious diseases and the increasing threat of epidemics and pandemics highlight the importance of improving efficiencies in acquiring and analyzing epidemiological parameter data to inform public health response.
Objective:
The Global Repository of Epidemiological Parameters (GREP) project, developed within the World Health Organization Collaboratory, aims to create an accessible, living database of epidemiological parameters on priority infectious diseases to serve as a global public health resource.
Methods:
Several projects are underway to progress on GREP’s five workstreams: prioritization of pathogens and key parameters; standardized data extraction methods; tools for storage and use of data; data validation and continuous maintenance of the database; and scientific recognition of contributors to the database. The data pipeline under development leverages artificial intelligence (AI) technologies to automate the maintenance of GREP in real-time.
Results:
Extraction of over 25 parameters for diseases such as measles is underway using systematic review methodology. In tandem, large language models are being used to automate literature searches and deduplication of search results, relevance screening, obtaining full texts, data extraction of parameters and data validation. Data storage and analyses are also being developed by other collaborators. The presentation will focus on specific approaches and performance related to automation of the data pipeline.
Discussion:
Using AI to modernize the way parameter data is gathered and synthesized will reduce the burden of repetitive tasks in the identification and extraction of data from the literature and allow human resources to be focused on assessment and analysis. GREP will be a centralized resource that helps to reduce global duplication of effort in curating evidence and facilitate faster evidence-based decision-making for better public health outcomes and epidemic preparedness.