By Deborah Borfitz

October 20, 2020 | Sometimes it really does take a village—when it comes to quick answers on anything COVID-19 related, to cite a particularly timely example. In March, a village of date-sharing researchers was assembled in record time for the largest study ever conducted on the safety of hydroxychloroquine.

The particulars were shared by Kees Van Bochove, founder and owner of The Hyve (based in The Netherlands) at the recent Bio-IT World Conference & Expo Virtual. Over four days, 350 participants from 30 countries held 12 global huddles, more than 100 collaborator calls and 13,000 chat messages, reviewed more than 10,000 publications, drafted nine protocols, released 13 study packages, and assembled a distributed data network with 37 partners signed on to execute studies.

The virtual COVID-19 Study-A-Thon, held by the Observational Health Data Sciences and Informatics (OHDSI) collaborative, was established to collaboratively design and execute observational research and generate real-world evidence to inform the global pandemic, says Van Bochove. It accomplished the first large-scale characterization of COVID patients in the U.S. and Asia, and the first prediction model externally validated on COVID patients to support triage to flatten the curve. 

Participants looked at 12 research questions, a subset of which focused on the safety of hydroxychloroquine administered alone or together with azithromycin, Van Bochove says. The team, representing six nations, applied standard analytical tools to one million patients divided into target, comparator, outcome and negative control cohorts using a unified web interface (the open-source OHDSI ATLAS) creating temporal constraints on contributed data. 

The raw data and execution stayed local and results were sent to a local server, Van Bochove says, making it unnecessary to share patient-level data beyond the custodian institution even as evidence was being generated at scale. The self-controlled case series study posted as a medRxiv preprint (DOI: /10.1101/2020.04.08.20054551) within a week. All the data sources, analytical codes and stratification methods are shared with readers.

Better Science

The COVID-19 Study-A-Thon highlights the ability of open science to speed the generation of medical evidence, says Van Bochove. It’s also the fix to a “broken” field marked by potentially dangerous shortcuts (and sometimes outright fraud), lack of scholarly discourse, and data hidden behind paywalls.

The European Health Data & Evidence Network (EHDEN), with funding from the Innovative Medicines Initiative, is helping to reinvent the scientific process by making the various steps (e.g., hypothesis generation, analysis) independent of one another and to get a global network of scientists to collaborate and exchange datasets, he says. The goal is “science for a better world,” meaning it drives policy decisions and is transparent about how conclusions are reached.

EHDEN’s ambitious goal is to harmonize 100 million health records to a common data model. Achieving that goal will require data interoperability, standardized analytics (enabling patient-level predictions), data networks and a strong collaborative community, and the OHDSI collaborative can make it happen, says Van Bochove. A core asset of the community is standardized vocabulary to which medical terms have been mapped.

A 2016 landmark paper published in Proceedings of the National Academy of Sciences (DOI: 10.1073/pnas.1510502113) demonstrated the feasibility of the OHDSI distributed data network to characterize treatments currently being used and identify those that are potentially better. The study enrolled more than 250 million patients and uncovered widespread differences in the treatment of patients with common chronic diseases such as type 2 diabetes, hypertension, and depression.

More recently, in a study published in The Lancet (DOI: 10.1016/S0140-6736(19)32317-7), OHDSI collaborators used insurance claim data and electronic health records from 4.9 million patients across nine observational databases to show the most popular hypertension drug isn’t the most effective. In terms of both safety and effectiveness, thiazide or thiazide-like diuretics were found to outperform ACE-inhibitors. 

Editor’s Note: Even if you missed the event, Bio-IT World Conference & Expo virtual is still live. Register now for on-demand presentations.