#135: GPT: Great Pediatric Tools! Harnessing the Power of AI in Pediatric Medicine

March 12, 2025 | By Sam Masur

Audio

Summary:

In this episode of The Cribsiders, we dive into the fascinating world of artificial intelligence (AI), large language models (LLM’s), and its implications for pediatric medicine. Our informatics expert and pediatric hospitalist, Dr. Tristan Nichols, discusses the current applications of AI in clinical practice on the ward, ethical considerations, and what the future may hold as AI continues to expand. Tune in to learn how AI is shaping the way we diagnose, treat, and care for our youngest patients.

AI Pearls

1. Embrace AI as a Learning and Clinical Tool, but with Caution
   AI tools  are already being used by learners and clinicians to enhance medical education, expand differential diagnoses, and improve documentation efficiency. However, it’s essential to remain aware of potential biases, inaccuracies, and ethical concerns when using these tools in patient care.
   
   
2. Develop AI Literacy and “Prompt Engineering” Skills
   Knowing how to effectively interact with AI models—through precise and structured queries—can significantly improve the accuracy and relevance of AI-generated responses. Medical professionals should learn how to craft prompts carefully to optimize AI assistance while ensuring patient safety and minimizing misinformation.
   
   
3. AI in Medicine Raises Ethical and Equity Concerns
   AI models can perpetuate biases present in training data, leading to disparities in healthcare recommendations. Additionally, issues like patient privacy, data security, and accessibility of AI tools must be considered. Clinicians should advocate for responsible AI implementation that prioritizes equitable and ethical patient care.
   
   

AI and LLM Notes 

Understanding LLM’s

• Large language models process and generate text based on massive datasets, allowing them to assist in medical reasoning but also posing risks of misinformation.

Examples of Clinical Applications of AI: 

• Differential Diagnosis Expansion: AI is assisting learners in identifying rare conditions by suggesting broader differentials.
• Discharge Instructions: AI tools are generating patient-friendly discharge summaries, improving understanding and adherence.
• AI Scribes: AI-powered transcription tools are streamlining documentation, allowing providers to focus more on patient care.
• Patient Education: AI-generated explanations tailored to different literacy levels help communicate complex medical conditions to families.
• Medical Literature Summarization: AI is being used to summarize journal articles, providing quick insights for busy clinicians.
• Clinical Decision Support: AI-assisted tools are helping providers analyze patient data and recommend potential next steps in care.
• AI-Powered OSCEs: AI chatbots are simulating standardized patient interactions for medical learners to practice history-taking and clinical reasoning.

Challenges and Limitations: 

• Bias and Misinformation: AI models are prone to bias, as they learn from historical data that may already contain inequities, potentially leading to inaccurate or harmful recommendations.
• Hallucinations: AI can generate incorrect or misleading information, necessitating clinician oversight to verify accuracy before applying AI-generated recommendations.
• Data Privacy Concerns: Ensuring patient data remains secure is a major issue, especially with AI tools that rely on cloud-based processing.
• Regulatory and Legal Uncertainty: The evolving landscape of AI regulation in medicine leaves questions about liability, responsibility, and best practices for integration.
• Dependence on AI: Over-reliance on AI tools may weaken clinicians’ critical thinking and diagnostic skills over time if not balanced with traditional clinical reasoning.
• Resource Accessibility and Equity: AI tools may not be available to all healthcare settings equally, exacerbating disparities in access to high-quality medical care.
• Environmental Impact: The computational power required to train and run AI models leads to increased energy and water consumption, raising concerns about sustainability in the long term.
• Future Considerations: AI will likely play an increasing role in decision support and patient care, but human oversight remains critical to ensure safe and ethical implementation. Ongoing research and regulation will be key in shaping the responsible use of AI in pediatrics.

Citations

Reis, M., Reis, F. & Kunde, W. Influence of believed AI involvement on the perception of digital medical advice. Nat Med 30, 3098–3100 (2024). https://doi.org/10.1038/s41591-024-03180-7

Witkowski, K., Okhai, R. & Neely, S.R. Public perceptions of artificial intelligence in healthcare: ethical concerns and opportunities for patient-centered care. BMC Med Ethics 25, 74 (2024). https://doi.org/10.1186/s12910-024-01066-

Learning Objectives

After listening to this episode listeners will…  

1. Describe the fundamental principles of artificial intelligence and machine learning in medicine.
2. Identify current AI applications in pediatric healthcare, including diagnostic tools and decision support systems.
3. Recognize ethical and bias-related concerns in AI development and implementation.
4. Discuss potential future advancements in AI and their impact on pediatric practice.

Disclosures

The Cribsiders have no financial disclosures. 

Citation

Nguyen H, Nichols T, Berk J, Chiu C, Masur S.. “#135: GPT: Great Pediatric Tools! Harnessing the Power of AI in Pediatric Medicine”. The Cribsiders Pediatric Podcast. March 12, 2025.