Validity of diagnostic codes and liver-related laboratory abnormalities to identify hepatic decompensation events in the Veterans Aging Cohort Study.

PURPOSE: The absence of validated methods to identify hepatic decompensation in cohort studies has prevented a full understanding of the natural history of chronic liver diseases and impact of medications on this outcome. We determined the ability of diagnostic codes and liver-related laboratory abnormalities to identify hepatic decompensation events within the Veterans Aging Cohort Study (VACS). METHODS: Medical records of patients with hepatic decompensation codes and/or laboratory abnormalities of liver dysfunction (total bilirubin ≥ 5.0 g/dL, albumin ≤ 2.0 g/dL, INR ≥ 1.7) recorded 1 year before through 6 months after VACS entry were reviewed to identify decompensation events (i.e., ascites, spontaneous bacterial peritonitis, variceal hemorrhage, hepatic encephalopathy, hepatocellular carcinoma) at VACS enrollment. Positive predictive values (PPVs) of diagnostic codes, laboratory abnormalities, and their combinations for confirmed outcomes were determined. RESULTS: Among 137 patients with a hepatic decompensation code and 197 with a laboratory abnormality, the diagnosis was confirmed in 57 (PPV, 42%; 95%CI, 33%-50%) and 56 (PPV, 28%; 95%CI, 22%-35%) patients, respectively. The combination of any code plus laboratory abnormality increased PPV (64%; 95%CI, 47%-79%). One inpatient or ≥2 outpatient diagnostic codes for ascites, spontaneous bacterial peritonitis, or variceal hemorrhage had high PPV (91%; 95%CI, 77%-98%) for confirmed hepatic decompensation events. CONCLUSION: An algorithm of 1 inpatient or ≥ 2 outpatient codes for ascites, peritonitis, or variceal hemorrhage has sufficiently high PPV for hepatic decompensation to enable its use for epidemiologic research in VACS. This algorithm may be applicable to other cohorts.

The Yale cTAKES extensions for document classification: architecture and application.

BACKGROUND: Open-source clinical natural-language-processing (NLP) systems have lowered the barrier to the development of effective clinical document classification systems. Clinical natural-language-processing systems annotate the syntax and semantics of clinical text; however, feature extraction and representation for document classification pose technical challenges. METHODS: The authors developed extensions to the clinical Text Analysis and Knowledge Extraction System (cTAKES) that simplify feature extraction, experimentation with various feature representations, and the development of both rule and machine-learning based document classifiers. The authors describe and evaluate their system, the Yale cTAKES Extensions (YTEX), on the classification of radiology reports that contain findings suggestive of hepatic decompensation. RESULTS AND DISCUSSION: The F(1)-Score of the system for the retrieval of abdominal radiology reports was 96%, and was 79%, 91%, and 95% for the presence of liver masses, ascites, and varices, respectively. The authors released YTEX as open source, available at http://code.google.com/p/ytex.