A “hat trick” is a slang term for the achievement of a positive feat “three times” in a sports match. While the term describes an achievement of “three times” in a variety of sports, it is most frequently associated with an ice hockey player scoring three goals in a single game.

Quite different than its use in a sports setting, three researchers associated with the Center for Pediatric Data Science and Analytic Methodology in the Johns Hopkins All Children’s Institute for Clinical and Translational Research scored a recent “hat trick” when each received an award at three different professional medical conferences, advancing the hospital’s vision to become the premier pediatric academic health system in the Southeastern United States.

Velopharyngeal Dysfunction — ASPS Best Podium Presentation Award

In 2025, Molly MacIsaac, B.S., a fourth-year medical student at The Renaissance School of Medicine, a graduate medical school of Stony Brook University, took time off between her third and fourth years of study to work at Johns Hopkins All Children’s Hospital in St. Petersburg, Florida, as a research assistant. In October 2025, she presented a research project at the American Society for Plastic Surgeons (ASPS) meeting in New Orleans titled “Development of a Clinical Workflow and Machine Learning Algorithm for Hypernasality Diagnosis Using Cleft Audit Protocol for Speech-Augmented-Americleft Modification (CAPS-A-AM) Rated Speech Samples in Pediatric Patients with Velopharyngeal Dysfunction.” She received an award for having the conference’s Best Podium Presentation.

“When I was a child, my family sponsored a girl in Sri Lanka who had a cleft lip and palate,” recalls MacIsaac. “We wrote each other letters for years and, during this time, her cleft lip/palate was repaired. This was my initial inspiration to pursue medicine and is one of the many reasons I’m interested in craniofacial surgery.

“Our project focused on improving access to diagnosis for children with velopharyngeal dysfunction, a speech condition often seen after cleft palate repair,” MacIsaac continues. “We developed a clinical workflow to collect high-quality speech samples and score them using CAPS-A-AM, the gold standard for perceptual speech assessment.”

In this work mentored by Alex Rottgers, M.D., and Jordan Halsey, M.D., the research team, “trained” a machine learning (ML) algorithm using Mel spectrograms of sustained vowels, which allowed them to distinguish — with high accuracy — hypernasal speech from normal speech.

“Our goal is to create a portable, artificial intelligence (AI)-based diagnostic tool that can support speech-language pathologists and improve care in both specialty centers and low-resource settings.”

“TransfusaScope”: AABB Outstanding Abstract Award

At the national meeting of the Association for the Advancement of Blood & Biotherapies (AABB), held in San Diego in October 2025, Emilee Greager, M.D., a pediatric resident at Johns Hopkins All Children’s Hospital, presented “TransfusaScope: Visualizing the Physiologic Response Before, During and After Transfusion.” She subsequently won the AABB conference’s Outstanding Abstract Award.

At the conference she presented the innovative technology named “TransfusaScope,” a tool designed to capture and visualize cross-sectional physiologic data surrounding transfusion events. It integrates vital signs, laboratory trends, provides patient-specific responses in “real time” and enables precision transfusion decisions tailored to individual physiology.

Greager shared that neonatal and pediatric populations are particularly vulnerable to transfusion-related complications, and that clinical decision-making in pediatric transfusion can be difficult due to the reliance on extrapolation of evidence from adult transfusion data.

“TransfusaScope’s validation was performed with retrospective data on 30 pediatric inpatients who underwent red blood cell (RBC) transfusion,” she explains. “Simultaneously, the tool extracts and integrates data from the electronic medical record (EMR), provides continuous vital sign monitoring systems and laboratory information systems. Key variables, such as transfusion timing, product type, lab values and vital signs, were extracted and displayed on an interactive dashboard.”

TransfusaScope was designed with synchronized event timelines, color-coded trend overlays, interactive filtering tools for stratification, and real-time anomaly detection algorithms.

“Data accuracy and integrity were validated through manual comparison against the EMR, and clinician feedback guided the refinement of both the user interface and backend logic to ensure usability and clinical relevance,” Greager says. “TransfusaScope can advance pediatric transfusion analytics and improve clinical outcomes.”

When AI Works Quicker than, but Just as Accurately as a Human: 2^nd Annual JHM Child Health Research Symposium Best Abstract Award

For his presentation titled “Comparing the Performance of Large Language Models against Human Reviewers across a Multi-Stage Scoping Review Workflow,”Frederick Kuo, M.D., an assistant professor in the Department of Anesthesia at Johns Hopkins All Children's Hospital, won the Best Abstract Award at the second annual Johns Hopkins Medicine Child Health Research Symposium/14th Annual Research Symposium held in January 2026.

Kuo’s team built a step-by-step AI “pipeline” using large language models to complete a literature review that took only a week to finish as compared to an earlier “manual” human review that took a year to complete.

“From title and abstract screening through full-text checks, the models closely matched human decisions and accuracy,” Kuo says. “Traditionally, literature reviews involve a person sifting through every relevant article, deciding what to include, extracting data and summarizing the evidence — work that can take months to years and quickly becomes outdated, yet still guides daily scientific and clinical decisions.”

Kuo told symposium attendees that when extracting, organizing and summarizing evidence, the models’ outputs stayed aligned with human results and preserved the same overall conclusions.

“These preliminary findings suggest that AI could ease workload and keep reviews current without sacrificing reliability,” he concludes.

 Technological Advancements to Provide Better Patient Outcomes

“We are very proud of our researchers’ accomplishments and of our contributions to advancing Johns Hopkins All Children’s as a national leader in pediatric research,” says Luis M. Ahumada, Ph.D., executive director of Advanced Technology and Data Science at Johns Hopkins All Children’s and assistant professor at Johns Hopkins University School of Medicine. “We are constantly testing and piloting new technologies, but there is always human activity to validate the accuracy.”

Ahumada explains that one AI technology they are working with is “OpenAI Whisper,” a “voice-to-text” translator tool that suppresses background noise and deals effectively with spoken accents.

“We have been supporting and collaborating with large teams of physicians and scientists,” he says.

The technology discussed in the presentation on velopharyngeal dysfunction is just one example. Ahumada says that because everyone might not have access to a speech pathologist, they are developing an app that can be downloaded onto a phone and accessed by speech pathologists.

Regarding “TransfusaScope,” its development will allow for more accurate observation of pediatric transfusions. Ahumada reiterates the reality that pediatric patients are quite different than adult patients and this technology will allow for pediatric patient data to be monitored in “real time.”

Finally, the ability to employ large language models to extract and process immense amounts of data in electronic medical records, as Kuo and his team are working on, will save significant time and resources, reducing processing time from weeks, months and years to hours.

“These are not just academic exercises,” he says. “Our center works to ensure that these tools and technologies will have a direct impact on how medicine is practiced and will help to improve patient outcomes.”