| アブストラクト | BACKGROUND: Researchers widely use the Diagnosis Procedure Combination (DPC) database for studies on acute-care hospitalization in Japan. However, the validity of acute kidney injury (AKI) diagnostic codes among patients who underwent surgery under general anesthesia is unclear. METHODS: Under the Japanese Next-Generation Healthcare Infrastructure Act, we obtained DPC data linked with the laboratory data of adult surgical patients with baseline and follow-up serum creatinine measurements from three hospitals (2021- 2024). Recorded diagnoses of AKI were identified using the International Classification of Diseases-10(th) Revision codes (N17.x). The reference standard was the diagnosis of AKI per the creatinine-based Kidney Disease: Improving Global Outcomes (KDIGO) criteria. The sensitivity, specificity, likelihood ratios, and predictive values were calculated. We evaluated the coding validity for identifying AKI stages 2 and 3. Subgroup analyses were conducted according to baseline estimated glomerular filtration rates (eGFR) (>/=60.0, 30.0-59.9, 15.0-29.9 mL/min/1.73 m(2)), type of admission (planned/emergent), and surgery type (cardiovascular/non-cardiovascular). RESULTS: Among 7,374 eligible patients, 663 (9.0%) met the KDIGO-defined AKI criteria, while 35 (0.5%) patients had the AKI codes. The sensitivity was 4.1% (95% CI, 2.7-5.9%), and specificity was 99.9% (95% CI, 99.8-99.9%). The sensitivity increased for patients with stages 2 and 3 AKI, lower eGFR, emergent admission, or cardiovascular surgery. CONCLUSIONS: Among surgical patients, AKI coding in the DPC exhibited high specificity and positive predictive value, albeit low sensitivity. Incorporating laboratory data is important for more accurate identification of patients with AKI for use in clinical studies. |
|---|
| 組織名 | Department of Clinical Epidemiology and Health Economics, Graduate School of;Medicine, The University of Tokyo.;Department of Prevention of Diabetes and Lifestyle-Related Diseases, Graduate;School of Medicine, The University of Tokyo.;Department of Real-World Evidence, Graduate School of Medicine, The University of;Tokyo.;Data Science Center, Jichi Medical University. |
|---|
