A Korean field trial of ICD-11 classification under practical clinical coding rules to clarify the reasons for inconsistencies.

Background: The World Health Organization (WHO) announced the release of the 11th edition of the International Classification of Diseases (ICD) in May 2019. Although Statistics Korea has been involved in the ongoing research on ICD-11 since 2017, we have been unable to achieve agreement on the gold standards for case scenario clinical coding in previous studies due to high levels of variance in the coding results of participants. Objective: The purpose of this study was to enhance clinical coding accuracy and consistency in ICD-11 by identifying and clarifying the reasons for these inconsistencies through the use of clear clinical coding rules. Method: A pre-experimental design was applied. Two clinical coding field trials (FTs) were conducted in 'ICD-11 for Mortality and Morbidity Statistics (2022 Mar)' targeting diagnostic terms and case scenarios. In the first FT, clinical coding rules were derived by analysing the results, while the second FT was performed under the clinical coding rules set by the first FT. Results: Across the two FTs, accuracy rates for diagnostic terms (75.8% and 71.8%, respectively) were higher than for case scenarios (62.5% and 71.9%). The main reason for the low accuracy levels was post-coordination. Conclusion: For case scenario clinical coding, low accuracy could be explained by variance in clustering methods between participants. This suggests that the accuracy of ICD-11 clinical coding could be increased if the variance between clustering methods can be reduced through the use of a clear coding guide. A guide for various ambiguous cases in each institution and the provision of a proper post-coordination list in the stem code could also be effective.