What is it about?
Chronic liver disease related to fat accumulation in the liver is very common, but it often goes unnoticed until serious damage has already developed. Detecting liver fibrosis early is difficult because the most accurate tests are invasive, expensive, or not routinely available in primary care. In this study, we developed an artificial intelligence–based clinical support tool that helps general practitioners identify patients at higher risk of liver fibrosis using only routine blood tests and basic clinical information. By analyzing real-world medical records collected over many years, the system can highlight patients who may benefit from further evaluation or specialist referral. This approach supports earlier detection, reduces unnecessary testing, and makes liver disease screening more accessible in everyday clinical practice.
Featured Image
Photo by Steve Johnson on Unsplash
Why is it important?
What makes this work unique is its strong focus on real-world primary care and on practical feasibility. Instead of relying on specialized imaging, invasive procedures, or complex data, the proposed approach uses only routine laboratory tests and clinical information that are already available to general practitioners. This makes the system immediately applicable in everyday practice, without changing existing workflows. The work is also timely because liver disease related to metabolic disorders is rapidly increasing worldwide, while primary care services are under growing pressure. By enabling earlier and more cost-effective risk stratification, this approach can help prioritize patients who truly need further investigations or specialist referral. Ultimately, it may contribute to earlier diagnosis, better use of healthcare resources, and improved patient outcomes at a population level.
Perspectives
This work reflects our interest in translating machine learning into tools that can be realistically adopted in primary care. Rather than pursuing increasingly complex models, we focused on understanding how routine clinical data can be used to support earlier and more practical risk assessment. Working with real-world electronic health records reinforced the importance of robustness, interpretability, and alignment with everyday clinical workflows, which I see as essential directions for future research.
Michele Bernardini
Universita degli Studi eCampus
Read the Original
This page is a summary of: Machine Learning-Based Clinical Decision Support System for Hepatic Fibrosis Risk Prediction in General Practice, ACM Transactions on Computing for Healthcare, January 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3788673.
You can read the full text:
Contributors
The following have contributed to this page
