Articles

A method for characterizing disease progression from acute kidney injury to chronic kidney disease.

Fang, Yilu; Nestor, Jordan; Ta, Casey; Kneifati-Hayek, Jerard; Weng, Chunhua

This study looks at patients who develop acute kidney injury (AKI) and tries to understand who is most likely to go on to develop chronic kidney disease (CKD). Using electronic health records, we tracked how patients’ health changed over time and grouped them into different clinical patterns based on lab results and medical history. We found that patients follow different recovery paths after AKI, and each path carries a different risk of developing CKD. While some known risk factors like diabetes and heart disease were important, we also identified new patterns that may signal higher risk. This approach could help doctors identify high-risk patients earlier and develop tools to guide closer monitoring and treatment to prevent long-term kidney damage.

Keywords: Acute kidney injury; Chronic kidney disease; Multi-state modeling; Natural language processing; Survival analysis

Files

  • thumbnail for A-method-for-characterizing-disease-progression-from-_2025_Journal-of-Biomed.pdf A-method-for-characterizing-disease-progression-from-_2025_Journal-of-Biomed.pdf application/pdf 1.27 MB Download File

Also Published In

Title
Journal of Biomedical Informatics
DOI
https://doi.org/10.1016/j.jbi.2025.104956

More About This Work

Academic Units
Biomedical Informatics
Medicine
Nephrology
Published Here
May 6, 2026