Recent studies reveal that artificial intelligence (AI) is showing promise in enhancing diagnostic accuracy in pediatric healthcare, especially in identifying rare diseases. A study published in Pediatric Investigation found that advanced AI models outperformed pediatric clinicians in diagnostic accuracy, with the combined efforts of AI and human expertise achieving the best results. This underscores AI’s potential as a valuable tool for improving diagnostic precision and patient outcomes in pediatric care.
The ability to accurately diagnose pediatric conditions is often complicated by the subtle or overlapping symptoms of rare diseases. Delays in diagnosis can lead to treatment postponement and increased risk of complications. While AI has shown potential in healthcare, previous studies often relied on simplified cases, leaving a gap in understanding its performance in real-world settings. Addressing this, Dr. Cristian Launes and his team at Hospital Sant Joan de Déu in Barcelona conducted a study using real pediatric clinical cases, comparing AI models with 78 pediatric clinicians on 50 cases, including both common conditions and rare diseases.
The study, published on March 25, 2026, used patient summaries from the first 72 hours of presentation to reflect real clinical practice. The evaluations focused on diagnostic accuracy and consistency, determining whether the correct diagnosis appeared as the top prediction or within the top five suggestions. Advanced AI models demonstrated higher overall diagnostic accuracy than clinicians, particularly in rare disease cases. However, clinicians showed strengths in complex scenarios, indicating distinct approaches between humans and AI in diagnostic reasoning.
Although the study did not test a real-time “human-plus-AI” workflow, it estimated potential complementarity using a “union” approach. This approach asked whether the correct diagnosis appeared in the Top-5 list of either clinicians or AI models, achieving a 94.3% Top-5 union accuracy. Dr. Launes emphasized that AI should serve as a clinician-supervised second opinion, broadening differential diagnoses and reducing missed diagnoses, provided there is critical interpretation and oversight.
Additional clinical information such as laboratory or imaging results improved diagnostic performance for both clinicians and AI systems. This emphasizes the importance of continuous clinical assessment and suggests that AI systems are most effective when integrated into information-rich workflows. The study highlights the promise of AI-assisted tools in supporting more accurate diagnoses, particularly for rare diseases, and suggests that integrating AI into clinical workflows could foster collaborative, data-driven decision-making among clinicians, engineers, and policymakers.