2021 Alternative Spring Break Program

Laura Bachor

This week, I was fortunate to participate in the University of Kentucky’s Alternative Spring Break program at the National Library of Medicine. While the program was conducted remotely this year due to the COVID-19 pandemic, my classmates and I were able to participate in many great activities to enhance our understanding of NLM and gain valuable hands-on experience. The bulk of the week was spent working on our projects; my particular project was focused on creating quality datasets for consumer health AI systems. I assigned metadata and tags to videos that could help answer specific consumer health questions, allowing these videos to be tested in an AI question answering service. It was fascinating to learn more about how these systems are designed and tested, and it felt very rewarding to know that this work will help NLM build more powerful tools for improving consumer health literacy.

My classmates and I were also able to participate in several educational sessions with experts in a variety of LIS fields during the week-long program. These sessions included engagement and training, cataloging and metadata, digital humanities, and the services and awards offered through NNLM. It was so interesting to hear about the wide range of opportunities available through NLM, as well as how passionate everyone was about their work. I appreciated that everyone we spoke with was so generous with their time and encouraging to us as new professionals and students. The Alternative Spring Break program was a wonderful experience from start to finish. I know the combination of hands-on experience and education that I gained during this program will be invaluable moving forward in my new career. I am very grateful that the program was offered this year, even under the difficult circumstances. It was a wonderful way to cap off my time in the LIS program!

 

Annabelle Smith

I’m so thankful that I had the opportunity to participate in the University of Kentucky’s Alternative Spring Break program at the National Library of Medicine (NLM) during the week of May 24th – 28th, 2021! This week I worked on a Preliminary Collection Assessment on Epidemics with my classmate Carly Jessup. During this assessment, we examined titles from University of Toronto Libraries’ Bibliography, “Epidemics in History, Literature, and Art.” A total of 123 titles were searched in the National Library of Medicine’s collection using LocatorPlus. Of those 123 titles, the National Library of Medicine held 110 of them. The missing titles were then further researched, and recommendations were made as to whether these would be beneficial to add to the National Library of Medicine’s collection. Additionally, some titles were referred to the appropriate liaisons to further analyze the titles as needed.

Being able to work on a fascinating project that is incredibly timely during the Coronavirus pandemic has been one of the best experiences I’ve had during the Master’s of Library Science program so far! I also had the opportunity to meet and learn from many professionals in the field of health sciences librarianship. This week my classmates and I participated in several discussions with experts to learn about many important topics in health sciences librarianship. These sessions covered topics in digital humanities, metadata, engagement, outreach and training, and awards offered through the Network of the National Library of Medicine (NNLM). Also, library professionals at NLM were eager to share volunteer and fellowship opportunities for students both during school and after graduating from LIS programs.

This week It has been a privilege to work with professionals that work to enhance information services locally and globally. The Alternative Spring Break program is a great opportunity for students to connect with professionals, learn new skills, and get hands-on library experience! I recommend that all students apply!

 

Susan Koenig

 

It felt a bit odd tuning in for an Alternative Spring Break (ASB) program at the end of May, after I had technically graduated. This program was something I looked forward to during my time at UKY, I was excited for the opportunity. Even though the program itself looked a bit different than it had in the past, the enthusiasm everyone at NLM brought was abundantly apparent.

I worked on the Creating High-Quality Datasets for Consumer Health AI project with Asma Ben Abacha and Shweta Yadav. My main task was annotating a set of over 500 health-related questions pulled from Yahoo Answers. I would read the question, provide a summary of what the person was asking, and classify the type and focus of the question. I was surprised to find additional opportunities to contribute to the project. After annotating a shorter set of questions, I met with Shweta and Asma to go over any questions I had about the annotation process. Based on my questions and suggestions, they updated the annotation guidelines for future use and edited the annotation portal. At the end of the week, I had learned so much about what goes into preparing datasets for natural language processing (NLP) and how complex these systems really are.

As I worked through the dataset, I kept thinking about how similar this process is to the work I do as a reference librarian, in particular the reference interview. So much of this process was reading paragraphs of information and pulling out the threads of what this post was actually asking. Quite often as librarians we tend to silo out our contributions to data analysis - we provide support for data management, information on best practices for storage and accessibility, and occasionally instruction on data literacy. But we have valuable contributions to make to data analysis and creation. At the end of the week, Shweta invited me to return to the NLM sometime this summer, and even offered to help me work on a publication based on the work I did during the ASB. I plan to take her up on that offer, and hope to write something about the parallels between dataset annotation for NLP and the reference interview.

I am so incredibly grateful for the opportunity to work at the NLM on this project. It was such a unique experience, and one that I will surely remember throughout my career. I made connections with other professionals in my field, learned a great deal, and have the opportunity to potentially present or publish something on the work I accomplished. I hope the ASB program continues to provide these same opportunities to future students.

 

Carly Jessup

During my week as a virtual intern for the National Library of Medicine, I had the opportunity to complete a Mini Collection Assessment on Epidemics. My research included examining a bibliography of epidemic related titles, determining whether or not NLM held these titles in their collection, and making recommendations as to whether the titles not held would be valuable additions to their collection. First, I learned how to use their online catalog. Next, I would research titles not held to determine whether or not they would be a valuable addition to the collection. At times, this could be challenging when looking for rare books with little information about them online. Additionally, I learned to make considerations concerning author credibility, relevance of content, etc. Lastly, we made a final summary of our findings and passed along titles to consider to the appropriate subject liaison. Considering the events of this past year, this research was both fascinating and rewarding, as it was an extremely relevant subject. Additionally, I had the opportunity to explore other areas by attending several virtual information sessions on a variety of topics such as Outreach and Engagement, Awards, Metadata, and Digital Humanities.

NLM did an excellent job of making this a rewarding experience, despite it being virtual due to COVID-19. I had the opportunity to learn from librarians in a variety of departments. This internship also allowed me to connect with my fellow UK peers, including both new and familiar faces. I am truly grateful that NLM welcomed us virtually, as many events this past year have had to be cancelled without alternatives. Additionally, I am grateful that the University of Kentucky and NLM have this valuable partnership, as it allows students to gain hands-on experience. It has been of great value to have been able to explore my library interests and learn from the caliber of librarians such as those at the National Library of Medicine.

 

Leah Everitt

For my week-long virtual internship with the National Library of Medicine (NLM) I worked with the team that helps maintain RxNorm. RxNorm is a standard vocabulary for drugs, which helps medical doctors, pharmacists, and researchers in insuring they are all talking about the same drug. For the internship project I worked to evaluate citations submitted to the RxNorm project through an online form. We were interested in knowing how RxNorm is used and how medical professionals and researcher might use RxNorm in the future.

Through the project I learned that RxNorm is most often used for ePrescriptions within medical system software. Currently several research groups are also creating software that can aid medical doctors and researchers in evaluating electronic health records, several standard medical terminologies are currently used in these records including RxNorm. This software should help researchers glean relevant information from electronic health records much faster than would be possible without the aid of computers. They would be able to use this information for research studies on particular diseases, subpopulations, and clinical research study cohort formation.

During the project I created summaries for the research papers I read, as well as powerpoint presentations that could be used as internal RxNorm case use reviews. Additionally, I was able to work with another student on the project, Jeff, and we collaborated finding the best method to evaluate the citations. Additionally, Jeff created some fantastic bibliometric data visualizations that we discussed the best method to produce and explain to the RxNorm team.

In addition to the RxNorm project all of the students participating in the alternative spring break also got to hear presentations from several different departments of the NLM. I hadn’t realized how far-reaching the NLM’s outreach efforts went, from helping rural communities in the US to US allies abroad. My favorite presentation was on metadata, presented by Nancy Fallgren, and went into detail about the way NLM transforms their metadata so that it is machine-readable, open-access, and interoperable in nature. The number of different forms of metadata that bibliographic records at NLM go through is dizzying and I found the presentation fascinating. Overall I feel that I not only learned a lot from my project about RxNorm and electronic health records, but also learned a lot about NLM as an organization.

 

Jeff Griffin

For the 2021 Alternative Spring Break project, I opted for the 2019 RxNorm Use Case Review. As a component of the Unified Medical Language System (UMLS) it too was developed by the National Library of Medicine. I became interested in UMLS in LIS 602 Knowledge Organization, using it as the subject of my final term paper. It seemed to me then and now an elegant solution to a uniquely fragmented and disjointed US healthcare system. But even more than that, it offers an avenue to something like natural language processing by distilling a hoard of controlled vocabularies into the minimum number of concept identities. NLM offers it freely for whatever use, and because its purpose is unspecified, what exactly it is being used for is a matter of interest, which is where this project comes in.

Over the course of the week, Leah Everitt and I went through and summarized RxNorm-related articles listed in Medline. It was far more challenging than I anticipated, since these articles represented highly specific and novel uses in various technical fields. The goal was to create easily digestible summaries of each article, outlining the ways in which RxNorm was used in each study. Since RxNorm, like UMLS, was designed to supplement a primary application, often it played a secondary, albeit essential, role to the studies in question. The articles were interesting. A common theme was the use of RxNorm on Electronic Health Record (EHR) databases for the purpose of retrospective cohort identification and construction. Because EHR systems have no shared structure or terminological convention, it is imperative to yield useful and consistent information from unstructured data (like those found in clinical notes). RxNorm greatly eases this task, allowing for machine processing of natural language by linking any term to all the affiliated ancestor and child terms via a concept unique identifier, which is just a semantically empty numerical string (it’s very beautiful in a Platonic sense).

The second component involved the creation of high-quality visualizations, but we were told at the start that they did not expect us to get to these because of the shortened amount of time allotted. But I remembered that, in LIS 662 Data Visualization, we briefly went over this bibliometric visualization program called VOS Viewer, which is open-source, creates nice bibliometric visuals, and actually has a user interface—so no programming required. Since I had a little time when I finished, I made some visualizations on author relations, keyword relations, MeSH relations, etc. This seems to me to be where bibliometrics is most useful, since it affords a visual summation of the interrelatedness of seemingly disparate articles. I would have done a topic-modeling analysis in R but time constraints and the dearth of articles kind of obviated this approach. It would be interesting to do it on UMLS in general, I think, since there is a significantly greater number of publications on it (UMLS predates RxNorm). Because UMLS and RxNorm often play a secondary role in the studies that cite them, treating the collection of said publications as a single corpus and using text mining to tease out topical themes would be ideal for this.

Elsewhere, they were generous enough to schedule talks by representatives of different departments, highlighting what they do and opportunities available for LIS graduates. This was immensely valuable to me since this field—along with its keystone institutions—is somewhat opaque and is singular in every sense. Although the project was virtual this year, I still feel it a rewarding experience. I think the National Library of Medicine is more exemplar of health science librarianship than, say, the Library of Congress is to general librarianship, because the field is defined by a single

matter of interest—that is, human health—which is every bit as collective as it is individual. The interests are consistent throughout. So I do feel grateful that I caught a glimpse of the inner workings of the NLM, if only from afar.