Beyond Accuracy: Defining the Role of Artificial Intelligence in Pediatric Electrocardiography

Q: What is the clinical evidence from this study?

Study design: Editorial. Population: Pediatric arrhythmias and heart disease (n=178). Intervention: Large Language Models (ChatGPT, Copilot, Gemini) vs. Pediatrician. Primary outcome: Diagnostic accuracy.

Sakshi Sachdeva; Navdeep Ahuja

doi:10.1007/s13312-026-00423-3

August 27, 2026Indian PediatricsOpen Access

Beyond Accuracy: Defining the Role of Artificial Intelligence in Pediatric Electrocardiography

View Full Paper

Ask AI

Bookmark

Share

Key result

A pediatrician outperformed large language models in pediatric ECG interpretation, achieving 78.7% accuracy compared to 71.9% for ChatGPT, 66.9% for Copilot, and 50.0% for Gemini.

Why the study?

Do large language models (ChatGPT, Copilot, Gemini) improve diagnostic accuracy in pediatric ECGs compared to a pediatrician?

Key Points

To examine the diagnostic performance, limitations, and practical clinical role of multimodal large language models in pediatric electrocardiogram interpretation.
Analyzed findings from a comparative study evaluating three public large language models (ChatGPT, Copilot, and Gemini) against a pediatrician on 178 pediatric ECGs under zero-shot conditions without clinical history.
The pediatrician achieved 78.7% overall diagnostic accuracy, followed by ChatGPT at 71.9%, Copilot at 66.9%, and Gemini at 50.0%, with moderate agreement for the pediatrician (κ = 0.566) and ChatGPT (κ = 0.410).
Copilot produced a false-positive rate of 24.7%, whereas Gemini missed nearly two out of every five pathological rhythms with a false-negative rate of 39.9%.
In the emergency-intervention subgroup, diagnostic sensitivity was 69.2% for the clinician but only 46–54% across the three artificial intelligence models, despite overall accuracy exceeding 86% for every evaluator due to low condition prevalence.

Structured PICO

Do large language models (ChatGPT, Copilot, Gemini) improve diagnostic accuracy in pediatric ECGs compared to a pediatrician?

P

Population

178 pediatric ECGs evaluated to compare the diagnostic accuracy of large language models versus a pediatrician.

I

Intervention

Three widely available large language models (ChatGPT, Copilot, Gemini) under standardized, zero-shot conditions

C

Comparator

A pediatrician

O

Outcome

Diagnostic accuracy of ECG interpretation

Main Result

Absolute Event Rate: 71.9% vs 78.7%

Limitations

General-purpose LLMs are built for language rather than specialized ECG interpretation.
The study withheld clinical context, supplying only the scanned tracing with age and sex.
AI-ECG models developed in high-income countries require external validation in LMIC populations before adoption.
General-purpose LLMs are built for language, with multimodal image handling added subsequently, rather than being optimized for a single task like ECG interpretation
Clinical context was deliberately withheld, supplying only the scanned tracing with age and sex
Most AI-ECG models have been developed on datasets from high-income countries, demanding external validation in LMIC populations

Implication

While large language models show potential for pediatric ECG interpretation, they currently underperform trained clinicians and lack the reliability needed for independent diagnostic use, highlighting the need for context-aware, specialized AI tools.

Abstract

No abstract yet. The journal hasn't published one we can show.

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Sachdeva et al. (2026) conducted an editorial in Pediatric arrhythmias and heart disease (n=178). Large Language Models (ChatGPT, Copilot, Gemini) vs. Pediatrician was evaluated on Diagnostic accuracy. A pediatrician outperformed large language models in pediatric ECG interpretation, achieving 78.7% accuracy compared to 71.9% for ChatGPT, 66.9% for Copilot, and 50.0% for Gemini.

synapsesocial.com/papers/6a92bf6e8e5d7d1fc0c12401 https://doi.org/10.1007/s13312-026-00423-3

View Full Paper

Ask AI

Bookmark

Share

Beyond Accuracy: Defining the Role of Artificial Intelligence in Pediatric Electrocardiography

Key Points

Structured PICO

Main Result

Limitations

Implication

Abstract

Discussion

Cite This Study

Beyond Accuracy: Defining the Role of Artificial Intelligence in Pediatric Electrocardiography

Key Points

Structured PICO

Main Result

Limitations

Implication

Abstract

Discussion

Cite This Study

Keep reading

More from Indian Pediatrics

Keep reading

More from Indian Pediatrics