Health Disparities: Big Data to the Rescue?

Guest post by Dr. Fred Wood, Outreach and Evaluation Scientist in the Office of Health Information Programs Development.

Socially disadvantaged populations have fewer opportunities to achieve optimal health. They also experience preventable differences when facing disease or injury. These inequities, known collectively as health disparities, significantly impact personal and public health.

Despite decades of research on health disparities, researchers, clinicians, and public health specialists have not seen the changes we were hoping for. Instead many health disparities are proving difficult to reduce or eliminate.

With that in mind the National Institutes of Health (NIH) National Institute on Minority Health and Health Disparities (NIMHD) launched a Science Visioning Process in 2015 with the goal of producing a scientific research plan that would spark major breakthroughs in addressing disparities in health and health care. NIMHD defines health disparities populations as including racial or ethnic minorities, gender or sexual minorities, those with low socioeconomic status, and underserved rural populations.

Through a mix of staff research and trans-NIH work groups—of which the National Library of Medicine is a part—NIMHD is gathering input on the current state of the science on minority health and health disparities.

Prompted in part by the NIH All of Us precision medicine initiative, one key visioning area—methods and measures for studying health disparities—includes big data.

We expect big data to bring significant benefits and changes to health care, but can it also play a part in reducing health disparities?

Last month the journal Ethnicity & Disease published a special issue focused on big data and its applications to health disparities research (Vol. 27, No. 2).

The issue includes a paper co-authored by the current NIMHD director, several NIH researchers (including me), and several academic partners. Titled “Big Data Science: Opportunities and Challenges to Address Minority Health and Health Disparities in the 21st Century,” (PDF | 436 KB) the paper identified three major opportunities for big data to reduce health disparities:

  1. Incorporate social determinants of health disparities information––such as race/ethnicity, socioeconomic status, and genomics–in electronic health records (EHRs) to facilitate research into the underlying causes of health disparities.
  2. Include in public health surveillance systems environmental, economic, health services, and geographic data on targeted populations to help focus public health interventions.
  3. Expand data-driven research to include genetic, exposure, health history, and other information, to better understand the etiology of health disparities and guide effective interventions.

But using big data for health disparities research has its challenges, including ethics and privacy issues, inadequate data, data access, and a skilled, diverse workforce.

The paper offered eight recommendations to counteract those challenges:

  1. Incorporate standardized collection and input of race/ethnicity, socioeconomic status, and other social determinants of health measures in all systems that collect health data.
  2. Enhance public health surveillance by incorporating geographic variables and social determinants of health for geographically defined populations.
  3. Advance simulation modeling and systems science using big data to understand the etiology of health disparities and guide intervention development.
  4. Build trust to avoid historical concerns and current fears of privacy loss and “big brother surveillance” through sustainable long-term community relationships.
  5. Invest in data collection on area-relevant small sample populations to address incompleteness of big data.
  6. Encourage data sharing to benefit under-resourced minority-serving institutions and underrepresented minority researchers in research intensive institutions.
  7. Promote data science in training programs for underrepresented minority scientists.
  8. Assure active efforts are made up front during both the planning and implementing stages of new big data resources to address disparities reduction.

Big data, it seems, is the classic double-edged sword. It offers tremendous opportunities to understand and reduce health disparities, but without deliberate and concerted action to address its inherent challenges and without the active engagement of minority communities in that process, those disparities could widen, keeping the benefits of precision medicine—including improved diagnosis, treatment, and prevention—from millions of those who need them.

How do you think big data will inform health disparities research? And what else might we do to ensure the disparities gap continues to close?

Head Nurse of the Library?

How a nurse’s way of knowing helps me do my job.

Many of you know I am a nurse. I began my career with a BSN from the University of Delaware (1975) and then completed an MSN at the University of Pennsylvania (1979). I also hold a license as a registered nurse.

I think of myself as a psychiatric nurse, although I spent a few years in the mid-1970s on the 3-to-11 shift in a busy surgical shock-trauma ICU. I have also served on the faculty of nursing at Case Western Reserve and at the University of Wisconsin-Madison.

It helps to understand nursing to see why being a nurse is such a great asset in my role as director of the National Library of Medicine.

To me nursing is fundamentally the diagnosis and treatment of the human response to disease, disability, and developmental challenges.

Note that phrase “human response,” i.e., how people react to the challenges in their lives.

Nurses focus on the human response, while the biomedical knowledge we have, like pathology or anatomy, helps us understand what a person is coping with and what kinds of complementary or supplementary supports are needed.

As noted nurse and author Virginia Henderson observed, nurses must actively engage with the patient to help him or her perform “those activities contributing to health or its recovery (or to a peaceful death) as he [or she] would do unaided had he [or she] the necessary strength, will or knowledge,” with independence from the nurse the common goal.

Nursing addresses the whole person and helps that person live to the fullest extent.

Along the way, nurses come to know people differently than the other clinical disciplines. And that knowledge, I’ve discovered, helps me as Library Director.

It affords me special insight into the public patrons who use our services. I can imagine a young mother using MedlinePlus in the middle of the night to figure out how to comfort a feverish child. I can anticipate the information needed by someone with a late-stage cancer diagnosis and recognize the need to complement journal articles on treatment options with literature on comfort measures and death with dignity. And I can appreciate the challenges of navigating the health care environment, from its specialized vocabulary to its unique culture.

A nurse’s way of knowing helps me set policies for integrating into the Library’s formal standards and language systems the terminologies that address the social and behavioral domains of health. Nurses live in those domains, as much if not more than the technical or scientific.

Perhaps most importantly, my experience as a nurse has taught me that each person has his or her own strengths, and that the Library’s resources should build upon those strengths to help the person make healthful choices, not just explain deficits.

But the benefits of being a nurse and a library director do not run only one way. Directing a library also lets me fulfill my nursing role, as I help others achieve the highest level of wellness possible.

As NLM director, I advocate for those in need, ensuring the literature is sufficiently inclusive. I improve patient care by guiding the fields of data science and biomedical informatics toward a future where professionals and patients interact to achieve care goals. And I model for younger nurses a career path that engages all that I am as a nurse while collaborating meaningfully across disciplines.

Being a nurse is not a job requirement for directing the National Library of Medicine, but it certainly is an asset. Advanced education as a health professional gives me an appreciation for how complex health care is and how important it is to engage all disciplines toward addressing that complexity. Engaging patients as partners in care motivates me to build resources that foster full participation of people in health. And experience as a team player in psychiatric services enables me to join with my colleagues from library science, information and computer science, linguistics, medicine,  and other disciplines to make the NLM the foundation of the future of health and discovery.

NLM’s Contributions to Healthy Communities

Public health lies at our core.

I’ve been on the job for eight months now, and I’m still learning more about the National Library of Medicine and its vast resources. These days, I’m broadening my understanding of how NLM  supports communities, population health, and public health.

The library has long had public health as an area of focus, both for its collection and its work. In 1993, the NIH Revitalization Act formalized some of our efforts by creating at NLM the National Information Center on Health Services Research and Health Care Technology (NICHSR).

Health services research (HSR) helps improve the quality and equity of health care for individuals and communities. NICHSR (pronounced “nik-H-S-R”) supports and promotes these efforts through its tools and by disseminating high-value information for the public health and health services workforce.

For example, in support of Healthy People 2020, a national program of specific objectives to improve the health of all Americans, NICHSR helps those in public health easily find the most up-to-date published evidence through a suite of preformulated PubMed search strategies. Their HSR and public health web portals deliver high-quality datasets, analyses, guidelines, and news, and the HSRProj database provides information about in-progress research before published results are available. And thanks to the newly-released NICHSR ONESearch, you can search across all NICHSR databases from a single point.

Committed to integrating resources for research and practice, NICHSR works closely with many of our sister federal agencies, such as the Centers for Disease Control and Prevention (CDC) and the Agency for Healthcare Research and Quality (AHRQ). Through the Partners in Information Access for the Public Health Workforce (PHPartners), which is a consortium of US government agencies, public health organizations, and health sciences libraries, NICHSR also works with national organizations such as the American Public Health Association, the National Association of County and City Health Officials, and the Association of Schools and Programs of Public Health. Together these collaborations deepen NLM’s work across the areas within public health while also promoting NLM’s public health resources to the communities who might benefit from them.

But NICHSR is not alone at NLM in working to ensure the health of communities and the people who live in them.

The NCBI Pathogen Detection System uses whole-genome sequencing to uncover potential sources of contamination during the outbreak of foodborne diseases. By comparing sequence data for bacterial pathogens obtained from food, the environment, and human patients to sequences in the NCBI database, we can determine if the bacteria are from the same outbreak, helping public health scientists trace the source of the contamination and target their response to the outbreak.

TOXNET, the flagship resource on toxicology and environmental health from the NLM Specialized Information Services (SIS) Division, delivers a collection of databases covering chemicals, drugs, occupational safety and health, poisoning, and toxicity risks, while SIS’s Disaster Information Management Research Center (DIMRC) supports disaster preparedness, response, and recovery by connecting people to quality disaster health information and fostering a culture of community resiliency.

Our landmark exhibition, Native Voices, preserves the healing cultures of American Indians, Alaska Natives, and Native Hawaiians and explores, among other things, how community lies at the center of Native conceptions of health. Through audio, video, photos, and stories, the exhibition—which is now traveling across the country—looks at how wellness of the individual is inseparable from harmony within the family and community and pride in one’s heritage.

This list, while far from comprehensive, gives a sense of the myriad ways this great library contributes to healthy communities. In fact, one could argue that everything we do ultimately has that impact.

And now we are looking for new and different ways to impact community health, this time through data and data-driven discovery. What ideas do you have about the types of data we should be collecting?

Illustration (top): Giovanni Maki, Public Library of Science (CC BY 2.5) | cropped, text removed

“And the Altman Goes to…”

NLM’s Role in Driving Progress in Translational Bioinformatics

Guest post by Dr. Valerie Florance, Director of the NLM Division of Extramural Programs.

How can you discover the cool stuff happening in translational bioinformatics?

Check out Russ Altman’s “Year in Review.”

Each year since 2007, Altman, a professor of bioengineering, genetics, medicine, and biomedical data science at Stanford, has hosted a popular plenary session at the annual joint summit on Clinical Research Informatics and Translational Bioinformatics sponsored by AMIA, the American Medical Informatics Association. Entitled “Translational Bioinformatics: The Year in Review,” this lively talk provides an overview of scientific trends and publications, celebrating progress and highlighting opportunities in research focused on informatics and data science methods that link biological entities to clinical entities.

It’s not exactly the Emmys, but Altman’s talk does bestow minor celebrity status on those acknowledged—and more importantly, it draws attention to new data collections, new software tools, and new directions for a field that increasingly impacts real problems in biology and medicine.

To build the list of candidate papers, Altman solicits recommendations from scientific colleagues. He then enlists volunteers from AMIA’s student working group to review and score the articles based on three basic criteria: informatics novelty, application importance, and presentability.

This year the student team scored over 285 articles, out of which Altman chose 28 to present at the joint summit and another 32 to receive a shout-out. This and past years’ honorees are enshrined on Altman’s blog.

After each meeting, NLM’s grant program staff comb through Altman’s presentation to identify work that builds on or highlights NLM-funded research. The process is not exact, relying on the articles’ acknowledgements of support to identify funding sources, but, like Altman’s talk itself, it’s something to go on, a back-of-the-envelope way of seeing where NLM dollars are having an impact and driving the science forward.

Of course, these papers commonly have multiple authors and multiple sources of funding. Furthermore, for the data scientists and informaticians supported by NLM, their methodological work may not be the focus of the article, but that work nevertheless contributes fundamentally to the reported results.

Of the 28 articles Altman presented this year, five acknowledged NLM grant support, while five additional NLM grantees secured one of Altman’s 32 shout-outs [total: 10 of 60 (16.7%)].

The following three examples from Altman’s list give a sense of the kind of work—and impact—NLM grants and grantees are having.

This research was funded by NIH grants from four different NIH Institutes, and by the Intramural Research Program of the NIH National Human Genome Research Institute.

Although Sean Mooney is not listed as a co-author, the article acknowledges his NLM-funded grant, “Informatic Profiling of Clinically Relevant Mutation” (1R01LM009722). Initially issued in 2007, Dr. Mooney’s award is still active and has produced 53 publications with more than 780 citations (Thomson Reuters).

The funded research develops novel methods to identify patients at risk for complex trait disorders, with a long-term goal of creating a whole genome interpretation engine based on data in public resources. Dr. Mooney’s academic background is computational modeling and biochemistry.

This paper acknowledges support from three different NIH Institutes and the National Science Foundation.

Co-author Marylyn Ritchie received NLM funding from 2009-2013 for the grant “Analysis Tool for Heritable and Environmental Network Associations” (1R01LM010040). Her work resulted in 33 publications with more than 151 citations (Thomson Reuters).

The project helped develop the ATHENA framework, which uses machine learning to incorporate biological information from databases with diverse data types to detect disease susceptibility driven by gene-gene and gene-environment interactions. Dr. Ritchie’s academic background is applied statistics and statistical genetics.

This work was funded by five different NIH Institutes and Pfizer.

The paper acknowledges Russ Altman’s NLM grant currently focused on “Text Mining for High-Fidelity Curation and Discovery of Gene-Drug-Phenotype Relationships” (5R01LM005652). There are 98 publications attributed to this grant, with 620 citations (Thompson Reuters).

Altman, an MD, PhD with academic training in molecular biology and medical information sciences, uses computational natural language processing to extract semantically precise knowledge about drugs, genes, and phenotypes.

In 1997, Altman’s research earned him the Presidential Early Career Award for Scientists and Engineers (PECASE), the first NLM grantee to earn that distinction.

Now, twenty years later, Dr. Altman is encouraging other scientists by featuring some of the innovative work happening in translational bioinformatics.

Whether funded by NLM or not, we expect this research to move the science forward, and who knows? Maybe something more than an Altman lies in the future for these researchers. Stay tuned.

Get Ready for Tornado Season

Information and readiness build resilience.

Guest post by Stacey Arnesen and Florence Chang, NLM Specialized Information Services.

The red pickup anxiously plowed through the corn field, heading toward the dark, ominous funnel, an F5-class tornado. Despite their many failed attempts, Jo and Bill were ready this time. On the truck bed was “Dorothy IV,” Jo and Bill’s invention, carrying thousands of data-collecting sensors. As the truck plunged into the swirling monster, Dorothy IV released the sensors, twirling and blinking like pixie dust. Streams of data collected by the sensors started flowing into a computer designed to make sense of the millions of data points regarding wind speed, air temperature, rotational force, and other elements of the storm.

This scene from the 1996 movie “Twister,” starring Helen Hunt and the late Bill Paxton, might have seemed a bit futuristic twenty-one years ago, but it was grounded in the science of the day. NASA and NOAA (National Oceanic and Atmospheric Administration) have collected data, shared data, and used data to predict the weather and make meteorological discoveries for decades before “Twister” brought storm-chasing and the work of the National Severe Storms Lab (NSSL) to the public’s attention.

Thanks to that research, today’s Doppler radar system uses sophisticated algorithms and modeling to monitor the weather and detect tornadoes. Advances in mobile technology and social media have also opened new channels for data exchange and communication, resulting in such apps as Wireless Emergency Alerts from NOAA and NSSL’s mPing, which crowdsources weather observations. Together, these early warning systems save lives and prevent injuries by giving people more time to take shelter before a tornado strikes.

Even with these advanced detection and warning systems, preparedness for tornadoes is still crucial. FEMA recommends that families start by assembling an emergency preparedness kit and creating a household evacuation and shelter plan. These will help you stay safe during the storm and survive on your own afterward until help can arrive.

NLM’s Disaster Information Management Research Center (DIMRC) suggests going one step further and packing a “digital go-bag” on your mobile device. Apps from FEMA, the American Red Cross, the National Weather Service, and NLM will ensure you have helpful and current information at your fingertips, from up-to-the-minute, localized weather warnings to first aid instruction and post-storm tips. You can also help with situational awareness and emergency response by submitting photos of damaged areas to FEMA.

Librarians as Information First Responders

Information is key during an emergency, which highlights an important role for libraries and librarians: as information first responders.

New Jersey State Librarian Mary Chute noted that disasters and other emergencies can turn a library into a “a safe haven where librarians, skilled in customer service and effective communications, can help those struggling to cope with unusual and stressful situations…and  offer critical services to help support police, firefighters, and medical personnel.”

In 2011, that role played out in Springfield, Massachusetts where, following a tornado, the libraries offered a space where community members without electricity or with damaged homes could gather to connect with each other and with essential services. A year later, following Hurricane Sandy, libraries were often the go-to place, and then-FEMA Administrator Craig Fugate tapped into the idea of library-as-community-builder by encouraging parents to read to anxious children. Readers tweeted their top picks using the hashtag #StormReads.

To foster the role of librarians as partners in disaster management, NLM, together with the Medical Library Association, developed a Disaster Information Specialization for those looking for ways to support their institutions or communities during such emergencies.

Who knows? Maybe the sequel to “Twister” will have librarians in the lead. (Hmmm, now we’re wondering who should play us.)

Florence Chang and Stacey Arnesen smilingGuest bloggers Florence Chang and Stacey Arnesen work in the NLM Division of Specialized Information Services.

Photo credit (tornado, top): Brent Koops, NOAA Weather in Focus Photo Contest 2015 [Flickr (CC BY 2.0)]