What is it about?
Acute appendicitis can be extremely hard to diagnose, especially in children. Delays or misdiagnoses can cause serious complications, while overdiagnosis may lead to unnecessary scans or surgeries. Dharma is an AI-powered framework that combines clinical knowledge with machine learning to help doctors detect appendicitis, assess its severity, and make evidence-based treatment decisions quickly. By enabling fast and accurate care, Dharma aims to improve outcomes for children and provide high-quality decision support even in resource-limited settings.
Featured Image
Photo by Katarzyna Zygnerska on Unsplash
Why is it important?
Diagnosing pediatric appendicitis remains one of the most common and consequential challenges in acute care, with high rates of delayed diagnosis, unnecessary imaging, and avoidable surgery, particularly in children and in resource-limited settings. Many existing clinical scores and AI tools are siloed, opaque, focused on binary diagnosis, or confined to studies and publications, limiting their utility in real-world, high-pressure clinical environments. Dharma is important because it integrates clinical reasoning with machine learning into a single, interpretable decision-support system that simultaneously addresses diagnosis and severity stratification, thereby enabling informed, evidence-based management decisions.It is designed to perform reliably even when the appendix is not visualized, laboratory data are incomplete, or advanced imaging is unavailable, conditions that frequently constrain routine care. By providing highly sensitive prognostic risk estimates for progression to complicated appendicitis and remaining usable in low-resource, high-volume settings, Dharma demonstrates how clinically grounded AI can improve decision-making, reduce both under- and over-treatment, and extend high-quality pediatric surgical care to settings where specialist expertise is scarce. More broadly, Dharma’s medicine-informed design offers a generalizable framework for developing trustworthy, usable clinical AI systems, informing solutions to other healthcare challenges and helping realize the full potential of data-driven AI in advancing 21st-century medicine.
Perspectives
Developing Dharma shaped how I think about clinical AI more broadly. Rather than treating machine learning models as standalone solutions, this work reinforced that effective clinical AI must be guided by medical knowledge and integrated into real clinical workflows. Models should function as cogs in the decision-making process—supporting, informing, and refining clinical judgment, rather than positioning themselves as the centerpiece or endpoint of care.
Dr. Anup Thapa Kshetri
Dharma-AI.org
Read the Original
This page is a summary of: Dharma: A novel, clinically grounded machine learning framework for pediatric appendicitis—Diagnosis, severity assessment and evidence-based clinical decision support, PLOS Digital Health, January 2026, PLOS,
DOI: 10.1371/journal.pdig.0000908.
You can read the full text:
Resources
Contributors
The following have contributed to this page
