Validity of International Classification of Diseases, Tenth Revision, Codes for Cirrhosis

Area of Science:

• Medical Informatics
• Hepatology
• Epidemiology

Background:

• Accurate patient identification is crucial for research utilizing administrative databases.
• Existing research on the accuracy of ICD-10 codes for cirrhosis is limited.
• Cirrhosis diagnosis accuracy impacts the validity of health research.

Purpose of the Study:

• To evaluate the diagnostic accuracy of major ICD-10 codes for identifying cirrhosis.
• To assess the performance of individual and combined ICD-10 codes in diverse patient cohorts.
• To provide recommendations for improving cirrhosis patient identification in administrative data.

Main Methods:

• Calculated sensitivity, specificity, PPV, NPV, and AUROC for major ICD-10 cirrhosis codes.
• Validated code accuracy in a chronic liver disease cohort (n=3,396) and a general patient cohort (n=1,560).
• Cirrhosis diagnosis was confirmed by individual chart review.

Main Results:

• Individual ICD-10 codes demonstrated high specificity (98.08-100%) but low sensitivity (0.27-55.70%).
• Positive predictive values (PPVs) were high in the chronic liver disease cohort (88.41-100%) but lower in the general population (55.53-66.76%).
• The AUROC for at least one code (0.79) surpassed that of any single code (0.50-0.65).

Conclusions:

• Individual ICD-10 codes are insufficient for accurately identifying cirrhosis patients in the general population.
• Combining diagnostic codes or using them with a chronic liver disease diagnosis enhances identification performance.
• Recommendations are provided to improve the accuracy of cirrhosis patient selection in health research.