In the fast-paced environment of the pediatric emergency department, physicians in training are constantly identifying opportunities to enhance their practice.
By recognizing spaces where new technologies can be applied, fellows are able to bridge communication gaps and clarify medical procedures for young patients and their anxious parents.
Throughout her training, pediatric emergency medicine fellow Niharika Goparaju, M.D., MPH, has encountered numerous communication barriers in pediatric care, particularly as young patients grapple with fear and anxiety around medical procedures. In one memorable case, she treated a 3-year-old with suspected appendicitis. The child needed an abdominal ultrasound but was visibly nervous and had difficulty sitting still. Goparaju turned to the artificial intelligence text-to-image model DALL-E to create a vibrant, child-friendly visual of the procedure. The result? A calmer, more cooperative patient.
Goparaju uses text-to-image model DALL-E to create custom images for her pediatric patients. The prompt for this image was to “create an illustration of a friendly physician performing an ultrasound on the abdomen of a 3-year-old girl.”
This is the future of patient care envisioned by Goparaju — embracing new technologies like generative AI to address persistent communication gaps due to patient age, language barriers and health literacy. Goparaju has already crafted a variety of images in her own work at the Dell Children’s Medical Center emergency department, from the general use of a CT scan machine to how albuterol is administered during an hourlong nebulizer treatment. And in a perspective recently published in the Annals of Emergency Medicine, she outlined future opportunities for best adapting these tools.
“Visuals serve a range of purposes and have proven especially helpful for patients and parents with low health literacy or language barriers, often in collaboration with an interpreter,” Goparaju says. “The images not only helped me explain medical procedures for families, but also have helped calm anxious children, making them more comfortable and willing to participate in their care.”
A Useful Tool … With Room for Improvement
AI-generated images can offer a unique advantage: customization. Unlike generic images from legacy medical image libraries, these visuals can be tailored to specific patients and scenarios making explanations more relevant and understandable. They are also quickly generated and cost-effective, providing a significant edge in both patient care and medical training.
The AI-generated images often require a number of prompt iterations to create the image needed. Goparaju used the following prompts to create this image of a physician putting a cast on a child. Original prompt: Show an image of a cast being placed on a child in a hospital. Adjusted prompt: Illustrate a physician casting a 5-year-old’s fractured left arm. The child is sitting on an examination table, looking bravely at his arm. The physician is focused on wrapping a lime green fiberglass cast around the child’s arm.
But before that can happen in real time at the patient bedside, more data input and involvement are needed from medical experts to ensure that images are consistently accurate in their depictions of unique medical scenarios. For now, Goparaju is developing a library of accurately generated images for cases she commonly encounters or images she anticipates will be useful.
“As the quality of the tools improve, I can see them being used extensively in medical education, from simulations and teaching aids to exam preparation, and to accurately depict conditions, procedures and medications with minimal prompting,” Goparaju says. “If the provider can spend less time ensuring appropriateness and medical accuracy, we can spend more time personalizing the images and highlighting the areas we want to focus on for the encounter.”
Sujit Iyer, M.D., pediatric emergency medicine fellowship program director, predicts that these tools will be especially helpful across both pediatric and adult emergency medicine, where physicians are routinely helping patients whose symptoms don’t display a clear or singular diagnosis, and therefore require a confusing gauntlet of tests, imaging and expert advice to determine the best path forward.
“I can’t tell you how many times I’ve drawn a picture of the colon or even the brain on a paper towel trying to explain what appendicitis is or a concussion,” Iyer says. “These tools help improve the standardization of how we communicate plans, procedures and even pathophysiology, and allow providers to have more meaningful human connections with these tools as adjuncts in the communication toolbox.
Drawing on Cross-Team Expertise
Beyond improving communication and patient care in the pediatric emergency department, Goparaju foresees the use of text-to-image generated images as a valuable mentorship tool, teaching her peer trainees and faculty how to enhance their interactions with patients, particularly in complex pediatric cases.
She hopes to upload the generated image library to easily accessible devices so pediatric emergency physicians can explain procedures or conditions to families in a clear and interactive way — and crucially, she’ll look to collaborate with health professionals like child life specialists to refine the content, drawing on their experience and best practices in child development and psychology.
“The expertise of child life specialists in calming and preparing children for procedures, along with the insights of my colleagues, could be invaluable in refining the visual content and ensuring that the images are effective and age-appropriate,” Goparaju says. “By working together, I hope to empower others to adopt this technology, streamlining communication and enhancing patient care across the board.”
The Value of the Trainee Perspective
Goparaju’s approach to integrating AI into medicine exemplifies how residents and fellows are not passive learners, but vital contributors to the enhancement of patient care. In the pediatric emergency room alone, Dell Med fellows have improved translation services for discharging patients, established web-based resources for patient education and created AI-generated videos to help explain complicated procedures.
“It is a privilege to see the health care system through the new eyes and skillsets of residents and fellows,” Iyer says. “Undoubtedly, our trainees are more facile and raised in a world where communication and information are always at their fingertips — however, the health care system still has work to do to filter, explain and help patients digest what information is important and relevant.
“Training fellows like Niharika to understand the nuances of pediatric emergency care, identify subtle findings from the doorway and then leveraging her enthusiasm to improve the patient experience is why most of love working in medical education.”
Graduate medical education, or GME, refers to the period of education in a particular specialty or subspecialty following completion of medical school. This continuation of training through residency and fellowship programs provides the clinical and educational experience needed for physicians to achieve autonomy, deliver high-quality patient care, and prepare for challenges in an evolving health care landscape.
Dell Med, in partnership with Ascension Seton, is home to 480 resident and fellow physicians and sponsors more than 40 residency and fellowship programs ranging from family medicine and neurology to pediatric emergency medicine and cardiovascular disease.