Sensitivity and positive predictive value of diagnosis codes for acute kidney injury in Denmark

丹麦急性肾损伤诊断代码的敏感性和阳性预测值

阅读:2

Abstract

BACKGROUND: Acute kidney injury (AKI) is associated with increased morbidity and mortality but is likely underrecorded in health registers. This study examined the sensitivity and positive predictive value (PPV) of AKI diagnoses compared with laboratory-identified AKI. METHODS: In this observational study we analysed data from the Danish National Patient Register and laboratory databases from January 2007 through November 2023. Diagnoses of AKI according to the International Classification of Diseases, 10th Revision (ICD-10) were compared with laboratory-identified AKI episodes defined by the Kidney Disease: Improving Global Outcomes (KDIGO) creatinine criteria. Sensitivity was defined as the proportion of laboratory-identified AKI episodes captured by ICD-10 codes within 30 days before or after the episode's index date and PPV was the proportion of ICD-10-coded AKI episodes confirmed by the KDIGO criteria within a ±30-day window. Analyses were stratified by sex, age, AKI stage, setting, comorbidity and short-term mortality. RESULTS: A total of 947 209 laboratory-identified AKI episodes and 80 649 ICD-10-coded AKI episodes were included. Overall, sensitivity was 7.5% [95% confidence interval (CI) 7.4-7.5], varying by stage (4.0% for stage 1 versus 21.7% for stage 3) and setting (6.0% for hospital acquired versus 8.6% for community acquired). The overall PPV was 90.6% (95% CI 90.4-90.9), with little variation across subgroups. CONCLUSION: ICD-10 codes of AKI demonstrate a high PPV, ensuring accuracy in identifying true AKI episodes. However, the low sensitivity highlights a risk of underestimating AKI occurrence. Laboratory data should be prioritized for comprehensive AKI identification and potential biases addressed when relying on diagnosis codes in research.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。