As healthcare workers, policymakers, and the general public continue to grapple with difficult decisions during the COVID-19 pandemic, it is of utmost importance that these decision-makers have access to complete and comprehensive information. While many resources have been created to track cases and mortality across the globe, it is vital that this evidence is supplemented with scientific research concerning the efficacy of therapies, the implications for at-risk subpopulations, and the clinical outcomes observed across the published literature.
This task of accessing and learning from the literature has rapidly become untenable for even the most dedicated of researchers. The number of COVID-19-related scientific publications is now consistently over 2,000 per week, over 70% of which do not report novel data. This means that any researcher looking to answer simple questions based on a comprehensive review of the literature (e.g. “how do the clinical outcomes of chloroquine compare to other therapies?”) is inundated with opinions, but will have difficulty in finding an up-to-date, evidence-based answer.
That is why Nested Knowledge, in collaboration with infectious disease researchers from multiple hospitals, has launched our COVID-19 StudyViz. This updatable, comprehensive review of the COVID-19 literature enables users to view individual studies—with machine-learning-assisted identification of the population, interventions, and outcomes reported—at the click of a button, but more importantly, the entire site serves as a mechanism to rapidly find the evidence related to COVID-19 research questions.
The map on the right enables users to filter to studies from any given country, while the filters below it allow users to select research based on study type and date of publication. More importantly, the sunburst diagram enables users to identify all studies reporting patient data related to any intervention, including all reported drug therapies, subpopulations of interest, any outcome of interest, or even articles reporting diagnostic data from imaging or lab tests.
The sunburst’s filters are based on a tagging hierarchy custom-designed to the evidence reported in COVID-19 publications, and each tag is nested within the relevant “parent” category to enable users to rapidly find and filter down to the desired level of specificity (e.g. “all studies of pharmacological interventions,” “all studies of antiviral therapies,” and “all studies of lopinavir” are all possible to find with one click).
This presents the medical public with a research capability that has never before existed—a user-customized, instant filter to studies that report data related to vital research questions. Importantly, this is not a search engine—users do not need to parse through the studies to determine if they report relevant data; instead, this is a tagged database, where users can choose what level and type of evidence is desired and simply click to get the finalized list of studies to examine.
Nested Knowledge offers this updatable data hub of COVID-19 research for free to promote evidence-driven decision-making. However, we would like to go beyond just offering the outputs of our own medical research; we would like to offer medical researchers across the globe the ability to build visualizations just like this one, using their own expertise, on any disease state or therapy of interest. To do so, all a researcher needs to do is create an account and follow the steps outlined here. By doing so, a researcher can launch a search, configure inclusion criteria and a customized tagging hierarchy, and apply their criteria and tags to studies, and a StudyViz page will be automatically launched. Not only that, but our system automatically saves the search, inclusion, and tagging activities, so if a researcher wants to go beyond the StudyViz and publish a research article based on their work, our system will generate the methods documentation (including a PRISMA diagram) to accelerate meta-analytical authorship.
Lastly, if any researcher would like to collaborate on our COVID-19 research or believes there is a pressing question that StudyViz should be used to address, contact us and we may provide a team of data gatherers and a developer to collaboratively build a StudyViz and even customize its contents and presentation to the project’s needs. We hope that the medical community recognizes the value in replicable, systematic, updatable review technologies, and will continue building StudyViz and associated technologies to curate the medical literature and democratize medical knowledge. We invite collaborators to reach out at email@example.com. For support/bugs/issues/feature requests by users, please email firstname.lastname@example.org.