Data guides health discovery: An interview with NLM Director Patricia Flatley Brennan

Patricia Flatley Brennan, R.N., Ph.D., leads the National Library of Medicine (NLM), part of the National Institutes of Health (NIH). NLM’s hardworking experts and vast repositories of information have been crucial to global efforts to combat COVID-19. Dr. Brennan shared the latest on NLM’s work to help foster and accelerate discovery, link people to clinical trials, and more.

What is the mission of NLM?

NLM is a platform for biomedical discovery. We bring together literature, genomic databanks, and researchers to serve science, scientists, clinicians, and the public. And we support discoveries globally. There’s not a biomedical discovery, a public health advance, or a clinical care action in the past 30 years that hasn’t benefited from our resources in some way.

NLM is a leader in biomedical and health data science research, and the world’s largest biomedical library. We conduct and support research in computational biology and computational health sciences. We also provide one of the most important functions that libraries serve in society—we are an archive of medical knowledge across the ages. Our holdings range from books and journals over 10 centuries old to the latest electronic genomic descriptions of the virus causing this awful pandemic.

We think of NLM as a door through which people pass and connect to data, literature, information, expertise, and sophisticated mathematical models, or images that describe a clinical problem. This intersection of information, people, and technology allows us to foster and accelerate discovery. That has been critical in the last year as we helped the country and NIH respond quickly to the coronavirus pandemic.

“We bring together literature, genomic databanks, and researchers to serve science, scientists, clinicians, and the public.”

– Patricia Flatley Brennan, R.N., Ph.D.

How is NLM helping NIH and the scientific community improve their understanding of COVID-19?

NLM staff provide special expertise that has been helping the rest of NIH respond to the COVID-19 pandemic. Our data scientists are helping researchers create reliable ways of collecting and managing data associated with complex studies and genetic sequencing efforts. Through our research and databanks, we are working with teams across NIH, other government agencies, and the larger scientific community to ensure that they have the full perspective of biomedical information available related to coronaviruses, which has been invaluable. So, it isn’t just our literature, our databases, it’s our staff in Bethesda that are helpful here.

Here’s one example: One of our NLM researchers is using computer vision and machine learning to identify lung abnormalities to distinguish between pneumonia caused by a bacteria or virus, and identify unique visual features associated with COVID-19. To do this, we collected large numbers of X-ray images—essentially a library of images—and studied their different patterns of light and dark shading using computer tools. By repeatedly doing this, we can sort them into various piles and identify lung X-rays with a similar shading pattern indicating inflammation from coronavirus, which looks very different than those piles associated with signs of viral or bacterial pneumonia. This model helps us understand new X-rays and helps clinicians make a COVID-19 diagnosis faster, because of what we’ve learned from our library of images.

What is NLM’s GenBank, and what role has it played in aiding the global response to COVID-19?

GenBank is the repository for all publicly available genomic sequences in the U.S. It has been crucial both in the process of developing vaccines and treatments, and in tracking emerging virus variants. As early as January 2020, only a month after the first patient was identified as having COVID-19, we already had the full genomic sequence of the coronavirus. This allowed us to provide a reference genome [a reference to what the genetic structure of the virus looks like] to scientists trying to develop new vaccines or new treatments.

We’ve continued to gather samples of coronavirus over the past year, and now we have almost 300,000 sequences of the SARS-CoV-2 virus. The GenBank sequence database helps researchers and public health authorities see if the virus is changing in a way that might require different kinds of treatment or prevention.

Clinical trials help researchers understand, treat, and prevent COVID-19. How does NLM help researchers, health care professionals, and the public learn more about COVID-19 studies?

NLM maintains a database of all the clinical trials around the world: ClinicalTrials.gov. People who are interested in participating in a clinical trial can visit the website to learn about more than 5,000 clinical studies related to COVID-19, ranging from vaccines and treatments to testing. The ClinicalTrials.gov website provides contact information for studies sponsored by NIH, other federal agencies, nonprofit organizations, and private industry. It’s a resource you can use to learn more about participating in a clinical study and find questions to ask the research team if you are interested in learning more. Questions such as, what will I have to do? What tests or procedures are involved? How long will the study last? Who will oversee my medical care while I am participating in the trial? Who will pay for my participation? Will I be reimbursed for other expenses?

How does the MedlinePlus website help consumers learn about COVID-19 and other health issues?

In addition to NIH MedlinePlus magazine, we also have our MedlinePlus website. This resource, which is available in English and Spanish, offers trusted, authoritative information for patients and families on thousands of health topics, including COVID-19. We recently added several new pages about COVID-19, including an overview, COVID-19 testing, and COVID-19 vaccines. These pages provide up-to-date information from NIH and other federal agencies like the Centers for Disease Control and Prevention.

How else is NLM research contributing to the fight against COVID-19?

As part of the NIH intramural program, NLM conducts research on the tools used to understand data: computational biology, bioinformatics, and biomedical informatics. Computational biology tools allow us to look through vast genomic databanks and understand patterns. For example, one of our investigators is trying to understand how viruses change over time and determine why some forms of the virus mutate to become more dangerous and some mutate to become less dangerous. By understanding these pathways of change, we may be able to predict which viruses will be most severe to the public and which are less likely to be severe.

NLM is also collaborating in an effort to link COVID-19 tests from specific manufacturers to their appropriate national code systems for use in medical record systems. This work has been instrumental in providing daily updates about the distribution of FDA emergency authorized COVID-19 tests.

2018 Morris F. Collen Award of Excellence: Patricia Flatley Brennan, RN, PhD, FACMI

NLM is examining hospital claims data related to COVID-19. Why is that?

Hospitals generate a lot of data, sometimes in the course of care, sometimes afterward when your bill is paid. We call this claims data. We have teams of investigators who are looking through claims information and properly approved hospital records to detect patterns, or associations, between the use of medications—such as ACE inhibitors (for blood pressure), and anticoagulants and antiplatelet drugs (for vascular disease)—on hospitalization and death rates among people with COVID-19. Identifying such patterns can be helpful to guide doctors and nurses about when to start or stop certain treatments.

How is NLM getting scientists rapid access to the latest biomedical literature about COVID-19?

NLM’s PubMed is the most heavily used biomedical literature citation database in the world, and it has long been recognized as a critical resource for helping researchers, health care professionals, students, and the general public keep current with rapid advances in the life sciences. Through PubMed, we provide free access to more than 32 million citations, including coronavirus-related literature. We’ve also launched the first phase of a pilot project to make preprints from NIH-funded research related to COVID-19 searchable in PubMed Central. Preprints are research manuscripts that have not been formally published or critiqued and refined through the peer review process. When made publicly available, preprints can play an important role in accelerating the dissemination of emerging research.

In addition, our data scientists can apply powerful computer tools, using natural language processing and machine learning, to essentially read these articles and find relationships that might not have been discovered by a human reader. We can additionally look for patterns in the literature. To do that, we’ve partnered with more than 50 publishers to make coronavirus-related articles open and accessible—what we call “machine readable”—in a way that is accelerating scientific discovery around COVID-19. This approach takes advantage of NLM’s ability not only to organize the literature, but also to establish policies and partnerships with the private sector, to be able to open the literature for the good of public health and science.