Validation of Liver Cirrhosis Network Criteria for Compensated Cirrhosis Through Natural Language Processing for Predicting Liver-Related Events.
retrospective_cohort · Level III
Where this comes from
- Record sourced from PubMed, PMID 42549810.
- Also identified by DOI 10.14309/ajg.0000000000004179.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
The liver cirrhosis network (LCN) developed a pragmatic clinical consensus definition for cirrhosis that we aimed to validate with natural language processing in 2 large VA centers. Of the 5465 patients identified with ICD codes for cirrhosis, 42.7% met the LCN definition. Meeting the LCN definition increased the probability of a liver-related event (LRE; new ascites, SBP, hepatic encephalopathy, HCC or variceal bleed) by 44% in multivariable analysis despite adjustment versus patients who did not meet these criteria. Each additional component of the LCN criteria increased the risk of LREs by 43%, which was led by FIB-4 and platelet counts.