Critique of ICD Code Accuracy in Identifying Cardiovascular Events in CKD and Kidney Failure Populations
Year: 2025
Type: article
Abstract: We commend Zemke et al.1 for their important work published in JASN evaluating the accuracy of International Classification of Diseases (ICD) codes in identifying cardiovascular events compared with physician adjudication in CKD and kidney failure populations. Although the study provides valuable insights, we believe the low sensitivity (15%–48%) across all outcomes warrants further discussion, particularly for atrial fibrillation, which had a notably low sensitivity of 15%. This could be due to underreporting of chronic or stable cases in ICD codes. It would be beneficial to include additional diagnostic methods like regular electrocardiogram screenings to better capture these cases.2 In addition, the study found that ICD-Tenth Revision codes performed significantly worse than ICD-Ninth Revision codes, especially for myocardial infarction and heart failure, showing a decrease in positive predictive value and sensitivity. Given the importance of ICD-10 for contemporary clinical practice, further investigation into the potential reasons behind this discrepancy is crucial. It may reflect either coding practices or clinical challenges such as elevated troponin levels in patients with CKD, which complicate myocardial infarction diagnosis.1,3 Although the high negative predictive values and specificity are useful for excluding false positives, the low sensitivity raises concerns about using ICD codes for identifying true-positive events in clinical trials. ICD codes may be more useful in epidemiologic studies where excluding false positives is the goal but less so in clinical trials that require precise identification of events.4 Finally, the study's exclusive use of primary ICD codes may limit its applicability. The inclusion of secondary diagnosis codes could enhance the accuracy of event identification, as seen in the secondary analysis, where positive predictive value decreased when only primary codes were used. This finding underscores the complexity of cardiovascular event coding in patients with multimorbidities like CKD.4 In conclusion, although the study contributes significantly to understanding ICD code accuracy, future research should explore strategies to improve sensitivity, especially for atrial fibrillation,2 and examine ICD-10 performance in CKD populations. In addition, incorporating both primary and secondary ICD codes will be crucial for enhancing event identification in clinical research.
Authors: Yingyu Wang, Chenyu Lu, Jiale Chen, Yabin Zhou
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7125189897