Technology and Data in Mental Health: Applications for Suicide Prevention

Guest post by Elizabeth Chen, PhD, Associate Director of the Center for Biomedical Informatics, Associate Professor of Medical Science, and Associate Professor of Health Services, Policy & Practice at Brown University.

Biomedical informatics as a discipline is broadly concerned with the effective use of data, information, and knowledge to improve human health. Since its origins in the 1950s, we have watched this discipline evolve with advances in health information and communications technology as well as the explosion of electronic health data. During this time, we have also seen the emergence of sub-disciplines reflecting areas of specialization. In fact, a 2015 study uncovered almost 300 different “types” of informatics! Among these was mental health informatics, which first appeared in the title of a 1995 article indexed in PubMed.

Using technology to understand and support mental health dates to the 1950s when specialized television broadcasts delivered mental health training. In the 1960s, computers analyzed data for psychological diagnoses and housed “artificial intelligence” systems that simulated communication with a psychotherapist. More recently, with the rapid adoption of electronic health record (EHR) systems that can collect longitudinal patient information such as diagnoses and medications, we are observing the increased use of EHR technology and data for improving health care, including mental health care.

Mental health remains a global crisis. In the United States alone, mental health conditions affect 1 in 5 adults and children. These conditions are among the factors that contribute to making suicide the 10th leading cause of death overall and 2nd leading cause among 10- to 34-year-olds nationally. With suicide rates having increased by nearly 30% since 1999,  the National Strategy for Suicide Prevention calls for a comprehensive and coordinated approach that includes data-driven strategic planning and evidence-based programs.

There are numerous and wide-ranging applications of mental health informatics and EHRs contributing to these efforts, including the following:

  • Two independent datasets, one including EHR and biobank data from the Vanderbilt University Medical Center, have characterized the role of common genetic variants among those who have attempted suicide. These large-scale genetic analyses support a heritable component to suicide attempts and an incomplete genetic relationship with psychiatric and sleep disorders.
  • At the Parkland Health & Hospital System in Texas, a Universal Suicide Screening Program, initiated in 2012, led to implementing the Columbia-Suicide Severity Rating Scale in the EHR system for adults. The integration of this clinical decision support tool into the clinical workflow demonstrates how technology may be used to improve suicide risk recognition.
  • Researchers across the country are developing models for predicting patients’ future risk of suicidal behavior using “machine learning” techniques, state death certificates, and longitudinal EHR data from a range of health systems, including Partners Healthcare in Massachusetts [PubMed], HealthPartners in Minnesota, Henry Ford Health System in Michigan, and five different Kaiser Permanente locations [PubMed]. Implementing these predictive models as clinical decision support tools in EHR systems has the potential to improve screening, detection, and treatment of suicide risk.
  • In Connecticut, EHR data from the statewide health information exchange and five clinical partners are being used to identify patients at risk of suicide. Claims data from the All-Payer Claims Database and mortality data from the State Department of Public Health will be used to assess the outcomes and impact of the quality improvement efforts.

And these are just a few examples.

Technology and data will continue to play important roles in advancing mental health care. We have already seen the contributions of mental health informatics over the years and those of related areas such as behavioral health informatics and computational psychiatry. There is much more to come in the development of effective and innovative solutions for improving diagnosis, treatment, and prevention of mental health conditions, including those related to suicidal thoughts and behaviors.

headshot of Dr. Elizabeth ChenElizabeth S. Chen, PhD is the founding Associate Director of the Center for Biomedical Informatics, Associate Professor of Medical Science, and Associate Professor of Health Services, Policy & Practice at Brown University. She leads the Clinical Informatics Innovation and Implementation (CI3) Laboratory that is focused on leveraging EHR technology and data to improve healthcare delivery and biomedical discovery. Dr. Chen is an elected fellow of the American College of Medical Informatics and is a member of NLM’s Biomedical Informatics, Library and Data Sciences Review Committee.

 


Dr. Chen will deliver the next NLM Biomedical Informatics & Data Science Lecture on Wednesday, November 14, 2018, at 2:00 pm in the Natcher Conference Center (Building 45), Balcony A. Her talk, “Knowledge Discovery in Clinical and Biomedical Data: Case Studies in Pediatrics and Mental Health,” is free and open to the public. It will also be broadcast live globally and archived via NIH Videocast.

Thank a Medical Librarian

Celebrating National Medical Librarians Month

“Get the word out. Tell the world what we do!”

I received that earnest and heartfelt request from those attending the Medical Library Association’s Midwest Regional Conference in Cleveland earlier this month. And though I thanked the conference attendees for all they do for NLM—helping us connect with our constituents in hospitals, academic institutions, and communities across middle America—I realized there was more I could do to thank and acknowledge all medical librarians, starting with this blog.

I believe that quality information is essential for improved health. It improves clinical decision making and patient care, boosts the quality of biomedical research, supports patients, families, and caregivers, and reduces health care costs.

And who is responsible for organizing and delivering that essential information?

Medical librarians and their partners in the health information profession.

For that, they deserve our thanks, but even more, they should be acknowledged for the myriad ways they improve health care and biomedical research.

Medical librarians…
  • Curate diverse and valuable collections.
    Librarians make deliberate and systematic choices to select the books, journals, data, and other resources needed for research and clinical care.
  • Catalog, index, and make available acquired materials.
    They make the needle you need findable in the collection haystack by adding relevant and appropriate subject headings or keywords to books, journal articles, data sets, images, and other items, which you can then locate by searching freely available databases like PubMed.
  • Manage access rights.
    Medical librarians support copyright and help maintain the intellectual property of authors, publishers, and database creators as they acquire and license resources on behalf of those who need them.
  • Support data discovery.
    Medical librarians identify and create pathways to data repositories that bolster genomic and biomedical informatics research.
  • Find the hard-to-find.
    Librarians know the ins-and-outs of online searching. They’ve trained for it, learning how different databases are organized and how best to extract precise results. Their expertise will save you time and improve outcomes.
  • Help authors publish.
    Librarians can help researchers at every stage of the publishing journey, from writing, revising, and formatting the paper to selecting appropriate and trustworthy outlets for publication.
  • Preserve materials for the future.
    They ensure the collections so painstakingly assembled are safe, secure, and available now and in the years to come, digitizing print materials, monitoring storage conditions, and conserving brittle, crumbling works.

Of course, to thank a medical librarian you have to find one. I suggest starting with NLM’s National Network of Libraries of Medicine (NNLM). At  over 7,000 sites strong, this network provides a point of presence for medical librarianship in almost every county in the US. Many NNLM members are academic institutions, health science libraries, or hospital or clinic libraries, but an increasing number (over 1,700 now) are public libraries taking on new ways to serve their communities.

They’re not alone.

Medical librarians have long ago left the desk behind and stepped into new roles, whether in health care institutions, academic libraries, or private industry. They are leading patient-and-family information services, becoming a part of the knowledge management resources of large health care systems, serving on patient safety and quality control committees, and joining teams of investigators to manage publications, locate critical data sets, gauge research impact, or write grants. From embedded librarian initiatives and innovative outreach programs, medical librarians are deepening the connection with the people they serve, bringing them shoulder-to-shoulder to share knowledge and solve problems.

They’re doing all this because they, too, believe that quality information is essential for improved health, and they know their skills and training put them in the best position to deliver that information.

That’s not only worthy of thanks but of shout-it-from-the-rooftops support. And not just because I say so, but because the data say so.

So, to provide better care, make better decisions, and save money, ask—and then thank—your medical librarian. They’re experts in helping you succeed.

Data in the Scholarly Communications Solar System

Guest post by Kathryn Funk, program manager for NLM’s PubMed Central.

The Library of the Future. What will it look like?  The NLM Strategic Plan envisions it partly as “one of connections between and among literature, data, models, and analytical tools.” In this future, journal articles are no longer lone objects drifting in space, but, rather, each a solar system waiting to be explored. Indeed, we’re already seeing the published literature associated with datasets, clinical trials, protocols, software, earlier versions (including preprints), peer review documents, and so on through consistent identifiers and standardized publishing and archival practices.

To help researchers and the public navigate this new solar system, PubMed Central (PMC), NLM’s full-text archive of journal literature, has been collaborating with publishers and funders for the last year to support efficient ways of linking journal articles with associated data. We’re encouraging authors to cite their open datasets and publishers to archive and make available those data citations in a machine-readable format. Though data citations represent only a small percentage of how PMC articles are linked to data (supplementary material continues to be the predominant method for associating data with articles in the archival record), the growth in data citations in the last year has been promising, nearly doubling the previous year’s total (i.e., 850 articles with data citations in 2017 vs.  approximately 440 in 2016). NLM is also supporting the public access policy requirements of our research funder partners by encouraging authors to deposit datasets as supporting documents via the NIH Manuscript Submission (NIHMS) system.

But solar systems, even the metaphorical kind, are meant to be explored, so we’re also working to expose each journal article solar system in a way that promotes discoverability. We want to make it easier to discover articles in PMC with associated data citations, data availability statements, and supplementary data, through improved record displays and new search facets, leveraging the data-related search filters announced earlier this year.

NLM is also looking beyond datasets to archive and expose articles’ key satellites, including, for example, comments generated during the peer review process. As the effort to expand the openness of peer review gains traction, PMC staff have been collaborating with publishers and Crossref on standardized ways to make readily available those peer review materials.

As with any exploration of new solar systems, it’s our hope that taking these steps will help generate new knowledge, and in so doing drive research that is reproducible, robust, transparent, and reusable. And as we move toward becoming the Library of the Future, how we can best support your research needs in connecting the literature with the rest of the research universe? Please let us know.

With thanks to Jeff Beck for the solar system analogy. 

casual headshot of Kathryn FunkKathryn Funk is the program manager for PubMed Central. She is responsible for PMC policy as well as PMC’s role in supporting the public access policies of numerous funding agencies, including NIH. Katie received her master’s degree in library and information science from The Catholic University of America.

Clarity Across Languages

The art and science of translating health information

Guest post by Fedora Braverman, team lead for the MedlinePlus en español website.

Communication can be tricky, regardless of what language you speak. Take, for example, this conversation I had some time ago with a Hispanic acquaintance:

Him (in a panic): “My friend has a tumor. The tumor is benign!”

Me (not understanding why he is panicking): “That’s great. That’s (kind of) good news.”

Him (looking at me like I had two heads and with his eyes wide open): “It was BENIGN!”

He thought “benign” (benigno in Spanish) meant cancer. He thought his friend had cancer.

I work for MedlinePlus en español, and moments like these make my job so rewarding because I can ease that man’s worries. By pointing him to our website with reputable and reliable health information, I can help him understand his friend’s condition.

But the conversation also made me think: if he thought “benign” meant cancer, then others might, too.

We work in the largest biomedical library. We are used to these words. Our audience is not.

Because our goal is to reach out to the Hispanic population as a whole, regardless of health literacy levels, our site needs to address disparities in health literacy by striving for clarity. So, the next day, after relaying this conversation to my team, we updated all instances of the word benigno on the MedlinePlus en español site, clarifying that it meant “non-cancerous.”

We are constantly learning and improving, refining our translation of the MedlinePlus en español site to enhance its cultural sensitivity and accessibility. It’s an art and a science. We use tools—from print dictionaries to Google searches and everything in between—to determine which word is used across Latin America  for a particular ailment, condition, or medical term. But translating text calls for far more than just swapping an English word for a Spanish equivalent. Word choice matters, as we try to accommodate regional linguistic and cultural differences along with the subtlety and nuance inherent in any language. That’s the art, as we not only translate the words but also adapt the text to our audience’s culture. Only then can we expect readers to connect with and understand the information.

Understanding the culture of our audience is imperative to building a site like MedlinePlus en español. Being knowledgeable about how Hispanics talk about their health issues (e.g., referring to diabetes as “this condition that affects the pancreas”), how they deal with certain hot topics or sensitive issues like sexually transmitted diseases, and what their health challenges are is crucial. That’s why my team and I work together, going back and forth until the text is as understandable and as culturally relevant as possible—even if that means re-working text we had previously translated.

So, whether you call it gripe, trancazo, influenza o gripa, benigno or no canceroso, NLM’s MedlinePlus en español is the trusted website for you.

headshot of Fedora BravermanFedora Braverman leads the MedlinePlus en español website’s operations, outreach activities, and social media platforms. She previously worked as a consultant for the US State Department and for the Library of Congress Hispanic Division. She has also served as an information specialist at the US Embassy in Buenos Aires, Argentina.

Communicating and Leading in a Time of Change

Every change management book I’ve ever read stresses the importance of communication during periods of organizational change. Talk to staff often, express your vision, share your excitement, listen to others on the journey. All sage advice.

And we’re certainly following it.

All NLM leaders, from group supervisors to branch chiefs to division heads to me, are communicating a lot—via staff meetings, wiki notices, large town hall meetings, brown-bag discussions, even this blog—as we prepare the National Library of Medicine to meet the future head on. We’re talking and listening, listening and talking, and it helps, opening our eyes to staff concerns and perceptions, along with their hopes (and fears) for the future.

Communicating vision and direction is so important to my leadership responsibilities that I invest a great deal of time and effort in organizing what I plan to say and how I say it. I review talking points with my senior leadership team to make sure I haven’t forgotten anything. I repeatedly rehearse my message, practicing timing and finding the right words to provide reassurance or spark excitement. I solicit staff reactions and allow time in my talks for open discussion (and provide channels for private, written feedback as well). I try to remember all the things I’ve learned about sender/receiver behavior, the importance of eye contact, and the power of nonverbals, including body language and tone of voice.

Despite all that though, what I say is not always what gets heard. After all, even the most carefully honed message has to be heard and understood, and that job falls to the listener. The listener’s daily concerns, tolerance for ambiguity, risk propensity, sense of control, cultural background, and other factors uniquely combine to filter and decode my words into what he or she understands me to be saying. Multiply that by several hundred listeners, and it’s a wonder there’s any alignment at all between what I send out and what is taken in.

To counteract that, I use what in engineering is known as purposeful design to help shape my communication. I carefully consider the range of ways a simple sentence or phrase might be heard by others and think about what aspect of the situation—stress, the presence of others, the nature of the news— might activate various filters and compound the mismatch between what I say and what is heard. I also actively solicit “back talk,” asking people to paraphrase what I’ve said or to tell me what they’ve heard. It takes extra effort, but it can save time in the long run as it uncovers misunderstandings, engages people around the same concept, and ensures we’re all in agreement.

And speaking of “back,” how about walking backward as a communication and leadership strategy? My good friend and mentor John Maeda recently re-shared one of his team leadership messages that featured Pentagon tour guides walking backwards. The guides are trained to lead groups through the building’s maze of hallways while facing them, pointing out important places along the way. John extends the idea to managers, recommending the manager occasionally “walk” facing the team, holding and directing their attention. And, as he wisely notes, “When you are watching where your team’s going, you also need to watch whether they are following.”

So, in this time of change, as we rely on our supervisors and branch chiefs and division heads to communicate vision, listen to staff, and help us move forward, I’m asking them to remember that what they say—what we all say—is often not what is heard and to occasionally walk backwards. The view is enlightening.

Animation credit (walking backwards, top): CC BY Ret Samys