Validation of coding algorithms for identifying people with viral hepatitis using claims data according to different...

Area of Science:

• Health Informatics
• Epidemiology
• Medical Coding

Background:

• Claims data analysis is crucial for public health surveillance.
• Accurate identification of hepatitis B virus (HBV) and hepatitis C virus (HCV) is essential for patient management and disease control.
• Performance of coding algorithms varies significantly based on reference standards and study periods.

Purpose of the Study:

• To assess the performance of various coding algorithms for identifying HBV and HCV infections.
• To compare algorithm performance across different reference standards (RSs) and study periods (SPs).
• To determine the optimal coding algorithm for maximizing both positive predictive value (PPV) and sensitivity (Sen).

Main Methods:

• A proportional random sample of 10,000 patients aged ≥20 years was analyzed.
• Three hierarchical reference standards (RS1, RS2, RS3) and three study periods (4, 8, 12 years) were utilized.
• Six coding algorithms were evaluated using ICD-10-CM codes in Taiwan National Health Insurance claims data (2016-2019).

Main Results:

• The number of confirmed HBV and HCV cases increased with more inclusive reference standards and longer study periods.
• Positive predictive value (PPV) was lowest with RS1 and highest with RS3, increasing with study period length.
• Sensitivity (Sen) was highest with RS2 and a 4-year study period; the algorithm '≥2 outpatient or ≥1 inpatient codes' showed the best overall performance.

Conclusions:

• Different reference standards and study periods significantly impact the estimated PPV and Sen for HBV and HCV identification.
• The coding algorithm '≥2 outpatient or ≥1 inpatient codes' is recommended for optimal identification of individuals with HBV or HCV.
• This finding aids in improving the accuracy of disease surveillance using health insurance claims data.