S83 Validation of ICD-10-CM Diagnosis Codes for Primary Sclerosing Cholangitis

Introduction: The International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) introduced a specific diagnostic code for primary sclerosing cholangitis (PSC) in 2018. The objective of our study was to assess the validity of ICD-10-CM code K83.01 for PSC, which could serve as a foundation for future PSC research using large datasets. Methods: We conducted a retrospective cohort study using patient charts of all adult service members from the Military Health System to identify all patient encounters from October 2018 to October 2024 that included the ICD-10-CM code K83.01 for PSC. Chart reviews were conducted to confirm the diagnosis of PSC, as evidenced by magnetic resonance cholangiopancreatography, endoscopic retrograde cholangiopancreatography, or liver biopsy showing PSC, or by documentation of the diagnosis from a gastroenterologist or hepatologist. The positive predictive value (PPV) was calculated to assess the accuracy of this code in identifying true PSC cases. We also stratified the PPV based on the number of PSC-coded encounters (1, 2, or 3 or more) to assess whether encounter frequency impacted diagnostic accuracy. Results: After adjudicating 170 charts containing the ICD-10-CM code K83.01 for primary sclerosing cholangitis (PSC), 121 cases were confirmed as true positives, 36 as false positives, and 13 were classified as indeterminate due to limited documentation. The overall positive PPV for confirmed PSC cases was 77.1%. In a secondary analysis stratified by the number of PSC-coded encounters, 29 patients had only 1 encounter, with a PPV of 65.4%. Seventeen patients had 2 encounters, with a PPV of 56.3%. Among the 124 patients with 3 or more encounters, the PPV was 82.6%. Indeterminate cases in each group were excluded from the PPV calculations. Conclusion: Our study demonstrates that the ICD-10-CM code for PSC achieves a high positive predictive value of 82.6% when patients have 3 or more PSC-coded encounters, indicating strong accuracy in identifying true PSC cases in this subgroup. While the overall accuracy is moderate, these finding highlights that requiring multiple coded encounters can substantially improve case identification. Nevertheless, a notable proportion of cases remain misclassified, suggesting that further refinement or additional diagnostic criteria may be necessary to enhance the code’s accuracy for broader clinical use.