MUSC joins national patient registry that will answer crucial questions about COVID-19

August 04, 2020
Big data analytics
Big data enables use of artificial intelligence algorithms. Image by Macrovector. Licensed from iStockphoto.com.

Size matters when it comes to research, as illustrated by recent debates about whether hydroxychloroquine is effective against COVID-19.

Some small studies showed it was effective while others showed it wasn’t. The studies couldn’t arrive at a definitive answer because they didn’t enroll enough patients to reach sufficient statistical “power.” 

More relevant results are only now beginning to emerge as higher-powered data has become available from studies enrolling more patients.

“A national database makes it possible to have a much larger number of patients and represent the different ways the pandemic is evolving in different regions and populations.”  

--Dr. Stephane Meystre
N3C principal investigator for MUSC 

Definitive answers about COVID-19 will require much larger datasets than those available in any one hospital system. Comprehensive national data will be needed to track the pandemic's spread, assess treatments and identify hot spots.

To meet that need, the National COVID Cohort Collaborative, or N3C, is creating a centralized national registry of patients who have been tested or treated for COVID-19. Participating hospitals will provide COVID-19 patient data twice weekly to keep the registry current.

The project is funded by the National Center for Advancing Translational Sciences (NCATS). NCATS has invited the more than 60 Clinical and Translational Awards (CTSA) hubs, whose mission is to speed research breakthroughs into the clinic, to submit data into the registry

The South Carolina Clinical & Translational Research (SCTR) Institute, which has an academic home at MUSC, was among the first five CTSA hubs to sign on to N3C, and MUSC has already begun sharing its data. 

Stephane Meystre, M.D., Ph.D., is the principal investigator of N3C for the MUSC site. Meystre is the SmartState Chair of the Translational Biomedical Informatics Center for Economic Excellence.

Dr. Stephane Meystre of the Biomedical Informatics Center at MUSC 
Dr. Stephane Meystre is N3C principal investigator at MUSC.

“A national database makes it possible to have a much larger number of patients and represent the different ways the pandemic is evolving in different regions and populations,” said Meystre.

Beyond tracking the evolution of the pandemic, the national registry will help researchers to study some of the rarer complications of COVID-19, according to Meystre. Researchers could use the registry, for instance, to identify enough pediatric patients who were diagnosed with multisystem inflammatory syndrome in children (MIS-C), which can result from COVID-19, to begin to understand and treat that disease better. 

Perhaps most exciting of all, the enormous dataset will enable artificial intelligence (AI) to unravel some of COVID-19’s mysteries. 

Given access to large quantities of data, AI can see trends and patterns that the human eye cannot. It can even make predictions about what will happen next.

For example, in a separate project, researchers at the MUSC Biomedical Informatics Center (BMIC) and the Data Analytics team have developed an AI algorithm that can analyze data on patient symptoms to predict which patients are most likely to have COVID-19. These predictions will be used to help give more accurate advice to patients monitoring their symptoms and will allow for more efficient COVID-19 testing.

Such AI algorithms feed on data, and the bigger the dataset, the more accurate its predictions will be. N3C will provide AI algorithms the robust data that is needed to begin predicting COVID-19’s next move and to identify the therapies that will most effectively combat it. 

N3C is initially requesting two pieces of protected health information – dates and zip codes – from participating institutions. This information is crucial for obtaining a comprehensive view of how the pandemic has progressed over time and spread geographically. It also helps researchers to assess the effectiveness of mitigation strategies and to make predictions about where the next hot spots might be.

“It is remarkable how quickly we were able to go from a crucial problem we had to help address ... to a very effective system extracting, transforming and exporting this information.”

-- Dr. Stephane Meystre
N3C principal investigator for MUSC

N3C wants to make this vital information available to researchers while taking every precaution to preserve patient privacy. Institutions encrypt the patient data before sharing it. N3C has created a synthetic dataset that shares statistical similarities with the actual data in the registry but does not contain any specific patient information. This synthetic dataset will be open to all registered researchers, and all researchers will be asked to begin their queries of the registry here. Should researchers require actual patient data, which will be available either fully deidentified or as a limited data set, they must request additional approvals and apply to N3C’s Data Access Committee for access.

MUSC was among the first to join N3C because BMIC researchers collaborate closely with investigators at Johns Hopkins University and Oregon Health Sciences University, who are leading N3C, and the University of North Carolina, which is heading up data acquisition for the project. Meystre and other BMIC researchers continue to work with N3C to develop tools using AI for detailed clinical information extraction to help to characterize patients with COVID-19. Investigators can then access the detailed clinical information that is relevant to their research question.

"Because our [CTSA] hub is so collaborative, we can expedite new collaborations and partnerships quickly.”

-- Stephanie Gentilin, director
SCTR SUCCESS Center 

Also key to the early activation of N3C at MUSC was SCTR’s SUCCESS Center and Regulatory Knowledge and Support (RKS) teams and the Office of Clinical Research’s Research Opportunities and Collaborations (ROC) team. Regulatory coordinator Toni Mauney quickly obtained MUSC institutional review board approval and ROC program coordinator Elizabeth Szwast shepherded the project through the activation process. Together, the teams were able to activate the project within two weeks of learning about the opportunity.   

 "Because our hub is so collaborative, we can expedite new collaborations and partnerships quickly,” said Stephanie Gentilin, director of SCTR’s SUCCESS Center and RKS. 

Meystre is proud of what the SCTR and BMIC teams have achieved.

“It is remarkable how quickly we were able to go from a crucial problem we had to help address, enabling secure and privacy-preserving access to detailed information from patients evaluated or treated for COVID-19, to a very effective system extracting, transforming and exporting this information,” said Meystre.

If N3C proves a success, it could be a model for how CTSA hubs could share patient data to speed the development of new therapies for any number of diseases, thereby improving patient care.

About the Author

Kimberly McGhee

Keywords: COVID-19