Validity of Migraine Diagnoses in Korean National Health Insurance Claims Data

Purpose: Accurate case identification using administrative datasets relies on diagnostic coding, yet these codes’ accuracy for migraine remains uncertain. This study aimed to validate the diagnostic accuracy of International Statistical Classification of Diseases and Related Health Problems 10th Revision (International Classification of Diseases, ICD-10) codes for migraine, migraine without aura (MOA), and migraine with aura (MA) in the Korean National Health Insurance Service database.Methods: We retrospectively reviewed the electronic medical records of 500 patients (migraine [G43.X], 200; MOA [G43.0], 200; MA [G43.1], 100) from secondary and tertiary hospitals between January 2019 and December 2024. Diagnoses confirmed by headache specialists using the International Classification of Headache Disorders, third edition served as the gold standard. Validation metrics included the positive predictive value (PPV), negative predictive value, sensitivity, specificity, and the kappa statistic. Diagnostic accuracy was assessed based on ICD-10 claim frequency and improved by combining diagnostic codes with prescriptions for migraine medications.Results: A single ICD-10 claim had a PPV of 74.00%. Accuracy improved significantly with increased claim frequency (≥3 claims: PPV, 81.14%; sensitivity, 98.61%; specificity, 28.26%), particularly when combined with medication prescriptions (≥3 claims with medication: PPV, 94.96%; sensitivity, 91.87%; specificity, 85.37%). MOA (≥3 claims with medication: PPV, 95.20%) and MA (≥3 claims with medication: PPV, 93.65%) showed similar trends. Excellent inter-rater reliability was observed (kappa, 0.806–0.951), with no significant accuracy differences between hospitals.Conclusion: Employing multiple claims and prescriptions improved the accuracy of migraine diagnoses using ICD-10 codes, supporting the use of this method in epidemiological studies and health policy decisions.