Validation of Claims‐Based Algorithms for Intussusception Identification in Japanese Hospitals: A Multicenter Analysis From the CLEAR Study

PURPOSE: Intussusception is a serious condition and a rare adverse event of rotavirus vaccination. While health insurance claims data are increasingly used for epidemiological research, claims-based algorithms for identifying intussusception have not been validated in Japan. This study aimed to validate multiple claims-based algorithms for intussusception identification against ultrasound findings as the reference standard. METHODS: This validation study used claims data and ultrasound findings from three Japanese hospitals between 2011 and 2024. The study population comprised all patients who received care during each hospital's data period. We developed 10 algorithms combining an intussusception diagnostic code (ICD-10: K56.1), hospitalization status, ultrasound imaging codes, and surgery codes. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each algorithm. Age- and sex-stratified analyses were also performed. RESULTS: Among 701 263 patients (mean age: 41.8 years; 52.0% female), 79 true intussusception cases (mean age: 8.3 years; 41.8% female) were identified (prevalence: 0.0113%). While sensitivity (54.4%-77.2%) and PPV (8.9%-28.1%) varied according to definition stringency, all algorithms achieved near-perfect specificity and NPV (≥ 99.9%). The algorithm that only used the diagnostic code yielded the highest sensitivity, while the algorithm that incorporated all four criteria achieved the highest PPV. Algorithm validity varied by age, with higher sensitivity and PPV for children aged ≤ 5 years. CONCLUSION: The algorithms demonstrated a trade-off between sensitivity and PPV, and researchers should select algorithms according to their study objectives. Our findings provide a robust foundation to guide algorithm selection for intussusception research using Japanese claims data.