JMIR Publications

ABSTRACT

Background:

The scale and quality of the global scientific response to COVID-19 has unquestionably saved lives. However, COVID-19 has also triggered an unprecedented “infodemic”; the velocity and volume of data production has overwhelmed many key stakeholders such as clinicians and policy makers who have been unable to process structured and unstructured data for evidence-based decision making.

Objective:

Current solutions aiming to alleviate this multi-disciplinary data synthesis challenge and minimise the deleterious impact of misinformation, are unable to capture heterogeneous web data in “real-time” for the production of contemporaneous answers and are not based on the identification and interpretation of high quality information in response to a freetext query.

Methods:

To realise an infrastructure that can addresses these shortcomings, the PanSurg Collaborative at Imperial College London has developed a unique data pipeline based on a webcrawl methodology.

Results:

REDASA is now one of the world’s largest and most contemporaneous COVID-19 evidence sources consisting of 104,000 documents and counting. This data pipeline converges with a novel curation methodology that adopts a “human in the loop” methodology for the characterisation of quality, relevance and key evidence across a range of scientific literature sources. By capturing curator’s critical appraisal methodology as discrete labels and rating information, REDASA has rapidly developed a foundational data science dataset of over 1400 articles in the COVID-19 problem space; that represents ∼10% of the papers written worldwide on COVID-19 in under 2 weeks.

Conclusions:

This dataset can act as ground-truth for future implementation of live, automated systematic review.