Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

MLA 2025 Annual Meeting Research Awards

Lindsay E. Blake, EdD, MLIS, AHIPa

aClinical Services Coordinator, UAMS Library, University of Arkansas for Medical
Sciences, Little Rock, Arkansas, https://www.orcid.org/0000-0002-8234-8611 ,
leblake@uams.edu

Cite as: Blake LE. MLA 2025 Annual Meeting Research Awards. Hypothesis.
2025;37(2). doi:10.18060/28904.

Blake. All works in Hypothesis are licensed under a CC BY-NC 4.0
DEED Attribution-NonCommercial 4.0 International. Authors own copyright of their
articles appearing in Hypothesis. Readers may copy articles without permission of the
copyright owner(s), as long as the author(s) are acknowledged in the copy, and the
copy is used for educational, not-for-profit purposes. For any other use of articles,
please contact the copyright owner(s).

1

https://www.orcid.org/0000-0002-8234-8611
https://orcid.org/
mailto:leblake@uams.edu
https://creativecommons.org/licenses/by-nc/4.0/
https://creativecommons.org/licenses/by-nc/4.0/


Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

Congratulations to the award-winning research papers and posters from MLA ’25
Pittsburgh!

The MLA Research Caucus is pleased to announce the winners for best research papers
and posters presented at the MLA 2025 Pittsburgh annual meeting. Thank you to all
the judges who volunteered their expertise to help select these deserving awardees both
in the pre-judging phase and during the conference. To learn more about the awards
and selection process, visit the Research Caucus on MLAnet.org.

MLA 2025 Annual Meeting Contributed Paper Awards

1st Place – Readability and understandability of online patient education in
semaglutide: Recommendations for consumer health librarians.
Authors: Alexandria Q. Wilson, Senior Medical Librarian, R. Eric Heidel, Shauntá
Chamberlin, and Kelsey L. Grabeel, University of Tennessee Graduate School of
Medicine.

Objectives: Media attention and public interest in semaglutide medications like
Ozempic have led to increased numbers of people searching for information about this
drug online. Given that national estimates suggest that many people struggle with
health literacy, we hypothesize that most online semaglutide information will be
written at or above the recommended level of 8th grade for most people. The primary
objective is to assess the readability of online information about semaglutide. The
secondary objectives will be to assess the understandability and quality.
Methods: Web links were collected by searching three terms (‘Ozempic,’ ‘Wegovy,’
and ‘Semaglutide’) on Google in private incognito browsers using three different public
computers. The non-sponsored results on the first five pages for each search were
manually imported into Excel and screenshots of each result page were saved.
Duplicate links were removed and screened against inclusion/exclusion criteria by two
researchers independently. Web links were excluded if they met the following criteria:
(1) main purpose is advertisement, (2) patient testimonials, (3) content for healthcare
professionals, (4) non-text material, and (5) research articles. Inconsistencies were
solved by discussion. The text from included web links will be extracted and evaluated
by two researchers independently. Included web links were evaluated for readability
using the SMOG Readability Formula and Flesch Reading Ease, for understandability
using the Patient Education Materials Assessment Tool (PEMAT) and for quality using
DISCERN. A statistician ran reports for interrater-reliability (Cronbach’s alpha),
medians and interquartile ranges, and frequency statistics of the evaluations.
Results: Three searches resulted in n=143 links. After duplicate removal, n=125 links
were screened for inclusion and n=61 met evaluation criteria. Inter-rater reliability
coefficients ranged poor (a = 0.37) to perfect (a = 1.00) between reviewers. Median
scores for SMOG and FRE were 13th grade level and College, respectively. Fewer than
10% of the documents were found to be at or below the recommended reading grade
level. The documents at or below were from hospital or international government
websites. For understandability, the median score of PEMAT was 62%. For quality, the
median overall score of DISCERN was 4 out of 5.
Conclusions: While the average quality was good, the majority of patient education

2

www.MLAnet.org


Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

available online about semaglutide medications is not written at the appropriate
reading level for the average consumer. Patient education on semaglutide needs to be
rewritten to be at the recommended 8th grade reading level. Additionally, with the
median understandability at 62%, more needs to be done to improve the education to
make it easier for the reader to understand. These findings are likely not unique to
semaglutide medications, but the popularity of these medications suggests an increased
need for patient education that is more understandable and readable. Librarians who
distribute online patient education to the public should be aware of the potential high
reading grade level for this topic.

2nd Place – “Social media is imperfect, but it’s the best we have”:
Understanding and addressing sexual health information seeking challenges
among young adults.
Author: Nick Vera, User Experience and Educational Strategist, Medical University of
South Carolina.

Summary: This session explores the critical role of health sciences librarians in
addressing sexual health information disparities among young adults in South Carolina.
By investigating how young adults use social media platforms as primary sources for
sexual health information, this study identifies key barriers such as misinformation,
algorithmic biases, and challenges in evaluating content credibility. Through a
dual-method qualitative approach, semi-structured interviews, and digital
walkthroughs. This session provides insights into young adults’ strategies and
experiences when navigating social media for sexual health information. It highlights
opportunities for health sciences librarians to develop targeted digital health literacy
resources, including workshops, fact-checking initiatives, and resource guides designed
to equip young adults with the skills needed to critically assess and use online sexual
health information. Attendees will gain practical strategies for implementing digital
literacy programs that enhance health information accessibility and empower young
adults to make informed decisions about their sexual health.
Objectives: Young adults in South Carolina face increasing challenges when seeking
accurate sexual health information due to limitations in traditional and authoritative
sexual health education and healthcare access. This paper investigates the future of
libraries and health science librarians by exploring how young adults use social media
platforms to seek sexual health information. It identifies the primary barriers young
adults face and investigates opportunities for developing innovative digital health
literacy resources.
Methods: This study employed a dual-method qualitative approach, utilizing
90-minute semi-structured interviews and digital walkthroughs with 20 young adults
aged 18-25 from across South Carolina. The interviews explored participants’
experiences and strategies when using social media to seek sexual health information,
while the digital walkthroughs provided real-time observations of their navigation
patterns, platform choices, and interactions. This dual-method approach allowed for an
in-depth analysis of the role that social media functionalities, algorithms, and
community interactions play in shaping the information-seeking practices of young
adults. The data included for analysis consisted of transcribed participant narratives
and screen recordings of participants’ digital walkthroughs. All data were analyzed
thematically using NVivo to identify salient patterns, challenges, and opportunities for

3



Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

interventions such as digital health literacy workshops, collaborative fact-checking
initiatives, and the development of educational resources tailored to guide young adults
in critically assessing and navigating social media for accurate sexual health
information.
Results: Preliminary findings show that while social media offers accessible sexual
health resources, young adults struggle to find credible and relevant information.
Barriers include misinformation, algorithmic biases that obscure accurate content, and
challenges in assessing creators’ credibility. There is also mistrust of medical
professionals due to perceived cultural incompetence in addressing young adults’ needs,
resulting in alternative strategies to navigate these obstacles. These findings emphasize
the need for targeted interventions, such as resource guides and digital literacy
programs, to improve digital health literacy and provide young adults with the skills to
evaluate and effectively use online sexual health information critically.
Conclusions: The study emphasizes the crucial role of health sciences librarians in
connecting young adults with credible sexual health information on social media.
Instead of focusing solely on users’ information gaps, health librarians can shift their
approach by developing resources that align with young adults’ existing social media
habits. This involves creating resource guides and digital literacy programs that teach
fact-checking techniques, identify trustworthy content creators, and explain platform
algorithms. By offering these targeted services, health sciences librarians enhance the
accessibility and quality of health information while positioning themselves as central
figures in addressing sexual health disparities. This proactive approach meets the
evolving information needs of users and highlights the future role of health sciences
librarianship in advancing digital health literacy.

3rd Place – Identifying strengths and gaps in health equity research: A
portfolio analysis of federally funded grants utilizing community engaged
research methods.
Authors: Poppy Krump, Information Scientist, Taneya Koonce, Consuelo Wilkins,
Spencer DesAutels, Mallory Blasingame, Alicia Cavanaugh, Laura Logie, Tiffany
Vassell, Braveheart Gillani, Jasmine Bell, and Nunzia Guise, Vanderbilt University
Medical Center.

Summary: To advance the science of community-engaged research, gaining an
understanding of previously funded research is required. Analysis of federal grants can
provide insights into community engagement trends. This analysis was able to identify
the trends and gaps in community engaged research, presenting opportunities to
improve the way research could be conducted and funded in the future. Information
scientists, given their expertise in information retrieval and data organization, can play
key roles in interpreting, structuring, and implementing grant portfolio analyses.

MLA 2025 Annual Meeting Contributed Poster Awards

1st Place – What does the body know? Nursing students’ epistemic beliefs
about embodied health misinformation.
Author: Brynne Campbell Rice, Librarian for Health Sciences: Nursing, New York
University.

4



Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

Summary: Academic health librarians are deeply invested in helping future healthcare
professionals develop the information literacy skills necessary to navigate an
increasingly fraught information landscape. In that vein, this poster presents initial
results from a qualitative study exploring how nursing students contend with a
particularly complex form of health misinformation. Specifically, this study investigates
nursing students’ perceptions and epistemic beliefs related to ”embodied health
misinformation” - the misinformation that arises in the messy intersection where bodily
experiences conflict with biomedical evidence. In investigating how nursing students
respond to this type of misinformation, the study aims to reveal how they engage with
information outside of traditional scholarly evidence, challenging simplistic binaries in
the study of misinformation. In doing so, this study contributes to the scholarly
conversation around how health professionals’ information literacy practices can
balance epistemic justice with a commitment to evidence-based health care.
Objectives: The objective of this research is to explore nursing students’ perceptions
and evaluation of ”embodied health misinformation” - misinformation that is woven
into bodily experience, where an individuals’ intimate knowledge of their own body is
positioned as more credible than biomedical evidence. Healthcare professionals need
well-developed personal epistemologies to navigate these complexities, yet there’s
limited research into how they perceive this type of information. To fill that gap, this
study asks: (1) What are nursing students’ perspectives on the body as a source of
health information and misinformation? and (2) What are nursing students’ epistemic
beliefs related to embodied health misinformation?
Methods: This study follows a qualitative methodology, employing a series of
in-depth, semi-structured interviews with nursing students. In order to elucidate their
thoughts on how to contend with bodily information when it appears to convey
misinformation, participants are asked to respond to two conflicting information
sources on the same topic: a personal health narrative that incorporates subjective,
affective health experiences, and an evidence-based information source. The audio of
the recorded interviews is transcribed and transcripts are coded inductively for
emergent themes in the nursing students’ perspectives on bodily information, and
deductively for any specific epistemic beliefs they reveal in their responses.
Results: Preliminary results following the analysis of 6 interviews with undergraduate
nursing students reveal key initial themes and topics that include: triangulation of
bodily information with external sources of information; the practice of bodily
listening; issues of individual vs. generalized health claims; and considerations of time
and risk in evaluating health information.
Conclusions: This project is ongoing, but it is anticipated that the results will
contribute to broader conversations about the development of sensitive information
evaluation practices among students who are entering health professions. Specifically,
understanding nursing students’ perceptions and beliefs around embodied information
can inform how librarians help prepare health professionals to contend with the full
complexity of the health misinformation that they will encounter, as both information
consumers and eventually, trusted sources of health information themselves.

2nd Place – Automatic indexing errors in PubMed: Sometimes the apple
does fall far from the tree.
Author: Paije Wilson, Health Sciences Librarian, University of Wisconsin-Madison.

5



Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

Summary: The objective of this study was to explore the prevalence and
characteristics of precision errors in a convenient sample of automatically indexed
MEDLINE records in PubMed. Specifically, the sample was comprised of records
automatically indexed with the MeSH term, Malus (the genus name for apple tree).
Objectives: Since 2022, fully automatic indexing has been used to assign Medical
Subject Heading (MeSH) to MEDLINE records in PubMed. Though automatic
indexing has improved efficiency, there have been concerns relating to its precision. The
objective of this study was to explore the prevalence and characteristics of precision
errors in a convenient sample of automatically indexed MEDLINE records in PubMed.
Specifically, the sample was comprised of records automatically indexed with the MeSH
term, Malus (the genus name for apple tree).
Methods: The following search string was used in PubMed to capture records
automatically indexed with the MeSH term, Malus: ”Malus[mesh] AND
indexingmethod automated”. The search retrieved 1,705 records, which were exported
from PubMed and then imported into EndNote 21. Title/abstract screening was
conducted in EndNote 21, wherein records were categorized as being correctly indexed
(i.e., the record was about Malus, meaning they referred to the literal tree or fruit),
incorrectly indexed, or uncertain (i.e., there was not enough information in the
title/abstract to verify if they were correctly indexed with the MeSH term, Malus).
Records identified as being incorrectly indexed or uncertain were exported into an
Excel sheet, wherein they underwent full text screening. During full text screening,
records were labeled as being correctly indexed, incorrectly indexed, or ”could not
determine” (i.e., the full text couldn’t be accessed for the record). Data extraction was
performed for records labeled as being incorrectly indexed during full text screening,
which consisted of copying quotations where variations of the term, Malus, (e.g., apple,
apple tree, etc.) were used in the record, and pasting them into an Excel sheet. These
quotations were then labeled with the type of indexing error (e.g., figurative language,
names, acronyms, etc.).
Results: 1,705 records were automatically indexed with the MeSH term, Malus. Of
these records, 1,562 (91.6%) were correctly indexed, 135 (7.9%) were incorrectly
indexed, and 8 (0.5%) could not be determined. Of the 135 records incorrectly indexed,
80 (59%) were due to the use of figurative language (e.g., ”comparing apples to
oranges”, ”apple shaped”, etc.). 50 (37%) were due to names (e.g., Apple, Inc.,
Miyake-Apple Technique, apple snails, etc.). 4 (3%) were due to acronyms (e.g.,
”Access to Post Partum LARC in Edinburgh South (APPLES)”). 1 (1%) was due to a
passing reference to Sir Isaac Newton.
Conclusions: With the recent implementation of the National Library of Medicine’s
newest automatic indexing algorithm, MTIX, to PubMed in 2024, studies documenting
automatic indexing errors in PubMed are especially vital. In addition to bringing
attention to automatic indexing errors, such studies may inform future refinements to
the MTIX algorithm to reduce erroneous indexing. Though limited, this study gives
some insight into the prevalence and characteristics of automatic indexing errors in
PubMed. As demonstrated by this study, the automatic indexing algorithm can
experience errors when encountering words used in non-literal contexts. Until
automatic indexing errors can be resolved, librarians should be mindful of their
existence, and instruct researchers on how to potentially avoid future indexing errors in
their own manuscripts.

6



Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

3rd Place – A multi-modal evaluation: Harnessing generative AI to
understand the state-of-the-art of literature search automation.
Authors: Taneya Koonce, Deputy Director, Annette Williams, Dario Guise, Jing Su,
Mallory Blasingame, Poppy Krump, and Nunzia Giuse, Vanderbilt University Medical
Center.

Summary: In collaboration with informatics colleagues at our institution, this
investigation explores the feasibility of delegating tasks currently undertaken by our
library team to generative artificial intelligence (AI) tools. Through a meticulously
designed research endeavor, we sought to comprehensively assess the contemporary
capabilities of generative AI in the domain of PubMed search compilation. Objectives:
In collaboration with informatics colleagues at our institution, this investigation
explores the feasibility of delegating tasks currently undertaken by our library team to
generative artificial intelligence (AI) tools. Specifically, we explored the efficacy of AI
tools in formulating reliable search strategies; we will be reporting on the current
state-of-the-art of AI tools while engaged in the capacity of “searcher.” Through a
meticulously designed research endeavor, we sought to comprehensively assess the
contemporary capabilities of generative AI in the domain of PubMed search
compilation.
Methods: The project leverages an internally-developed “Quick Clinical View” feature
of our institutional medical record that aims to provide clinicians with rapid insights
into patient conditions at the population level by highlighting concepts with high
strengths of associations. It uses the database of population-level associations created
by an institutional natural language processing (NLP) system, which maps free-text
terms from all non-image clinical documents at our medical center to the Unified
Medical Language System (UMLS) terminology. Our team is exploring whether
generative AI has advanced enough to automatically integrate literature searches into
this system—a potentially transformative feature that could rapidly identify
unfamiliar/unexpected clinical associations. From the “Quick Clinical View” of ten
health conditions (e.g., small cell lung carcinoma, orthostatic hypotension), we selected
two concepts each (i.e., 20 condition/concept pairs) that occur rarely in the general
population. To assess the feasibility of automating the literature search integration, the
team created generative AI prompts designed to produce PubMed search strategies in
the PICO format. Our team evaluated search strategy results for the 20 selected
condition/concept pairs across three different large language models (LLMs):
ChatGPT, Google Gemini, and Microsoft Copilot by comparing them to the team’s
gold-standard search output.
Results: The LLM tools generated collectively 150 terms/phrases tagged as MeSH, of
which, 58 (39%) were accurate. ChatGPT generated accurately labeled MeSH terms
45% (41/92) of the time and Copilot 31% (17/56). Gemini labeled two terms as MeSH;
neither were accurate. The 64 relevant citations identified by our team through their
gold-standard searches were furthermore compared to the search results of each LLM.
The LLM tools’ search results included 47% (30/64) of these citations overall;
ChatGPT included 17% (11/64), Gemini included 22% (14/64), and Copilot included
8% (5/64), with instances yielding zero search results.
Conclusions: The generative AI-created search strategies often failed to produce
PubMed results, were frequently inaccurate in labeling search terms/phrases as MeSH,
and for the searches that did yield results, the gold-standard selected citations were

7



Editorial Reviewed Article Hypothesis , Vol. 37, No. 2, 2025

present less than half of the time. Moving forward, it will be important to evaluate the
entire set of citations retrieved by the LLM tools as they may offer value-added to the
citations identified by the librarians. This study, once more, underlines the critical
importance of having knowledge workers with the necessary domain knowledge involved
in assessing and evaluating the performance of generative AI tools.

8