Identifying Ascites in Patients with Cirrhosis Using Administrative Codes and Diuretic Use: A Multicenter Study

Area of Science:

• Hepatology
• Clinical Informatics
• Medical Record Analysis

Background:

• Ascites significantly increases patient morbidity, mortality, and healthcare costs.
• Accurate patient identification is crucial for large-scale ascites research.
• Current ICD-10 codes for ascites have poor performance in patient identification.

Purpose of the Study:

• To evaluate the effectiveness of ascites-related ICD-10 codes.
• To assess the utility of diuretic use and cirrhosis codes for identifying ascites patients.
• To propose an improved method for ascites identification in large datasets.

Main Methods:

• Retrospective analysis of three independent patient cohorts (n=315).
• Evaluation of performance characteristics for ascites ICD-10 codes (K70.31, K70.11, K71.51, R18.8).
• Assessment of loop and/or potassium-sparing diuretics combined with cirrhosis codes.

Main Results:

• Algorithms using any ascites code showed improved sensitivity over R18.8 alone.
• Diuretic records combined with cirrhosis codes demonstrated high sensitivity and specificity (e.g., 82.3% sensitivity, 89.1-93.5% specificity in cohort 2).
• Ascites codes alone were found to be insensitive.

Conclusions:

• Diuretic records combined with cirrhosis codes enhance ascites identification.
• This combined approach offers a more reliable method for ascites detection in large studies.
• The proposed method should be adopted for future large-scale ascites research.