ICD-11 pilot in West China: feasibility evaluation and implications for future implementation Authors: Xingyu Ming, Han Wei, Die Zhou, Jinyuan Tao, Bangfu Qian, Hongying Li, Jinliang Hu Published: 2026 ABSTRACT AbstractThis study was designed to comprehensively evaluate the feasibility of implementing ICD-11 as a replacement for ICD-10 in West China in terms of morbidity and mortality coding, summarizing practical experiences and providing valuable insights for the transition to ICD-11 in other regions and countries. Coding data for ICD-11 and ICD-10 were collected during a two-month period from the pilot hospital. Additionally, we also conducted a questionnaire survey and focused interviews among 20 coders. After implementing ICD-11 coding, 92.30% of the primary disease codes remained unchanged in terms of chapter allocation. 41.10% of the codes involved postcoordination, with a maximum of eight codes in postcoordination. Topology Scale Value, Anatomy and topography, and Histopathology were the most frequently used extension code dimensions during postcoordination. Most coders believed that ICD-11 more accurately expressed clinical diagnoses in comparison to ICD-10, the pre-assigned postcoordination provided was user-friendly, but a large number of cases required manual postcoordination, the quality of medical record writing did not meet the requirements for implementing ICD-11, and conducting coding audits was challenging. The evolution from ICD-10 to ICD-11 was smoother compared to previous version upgrades. Coders acknowledged the improvements of ICD-11, but overall the coding process was time-consuming and complex, with certain coding obstacles. The implementation of ICD-11 to replace ICD-10 was considered feasible if solid groundwork could be laid for improving the quality of medical records, with proper training and the release of official ICD-11 textbooks of China.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-56119-y.Keywords: ICD-11, ICD-10, Implementation, Evaluation, Obstacle, FacilitatorsSubject terms: Standards, Outcomes research Background The International Classification of Diseases (ICD) is a standardized classification system for diseases and health-related conditions. The release of International Classification of Diseases-10th Revision (ICD-10), it has significantly transformed health statistics and helped make health policy more precise. ICD-10 has been widely applied in various fields, such as hospital management, medical insurance payment, medical quality management, international accreditation of healthcare organizations, and professional capacity assessment1,2. However, long-term use of ICD-10 has shown some limitations, for example, a lack of flexibility and lower specificity in certain categories. Many countries continuously update their national revisions to better adapt to new health trends, but this results in inconsistencies between national revisions and the international version3,4. To ensure internationally comparable morbidity data, a classification system with a large number of synonyms and details is required to accommodate the constantly expanding range of ICD-10 revisions and medical terminology scopes in different countries. The latest version of the International Classification of Diseases, the 11th Revision (ICD-11), incorporates advancements in medical technology and scientific research, aiming to address existing challenges and continuously evolving to meet new ones emerging within the realm of medical science5. ICD-11 has been significantly improved on the basis of ICD-10, manifesting not only in the addition of new chapters such as Sleep-wake disorders, Traditional Medicine, and the expansion of category capacity (code capacity), but also postcoordination, which is the most remarkable new function. It represents a powerful and effective means of expanding the coverage, expressive capability, and granularity of terminology, enabling users to specifically encode clinical concepts recorded by healthcare practitioners through code combinations (known as “cluster”)6–8. Designed based on modern information technology and devices, ICD-11 is sufficiently flexible for future modifications and utilization along with other classifications and terminologies. ICD-11 may more accurately capture emerging and novel health conditions and allow diagnostic and therapeutic information to be summarized according to a variety of user objectives, providing opportunities for disease coding while emphasizing the need for accurate diagnosis and treatment supported by reliable data9–11. Furthermore, ICD-11 introduces new “extension codes” that encompass additional critical clinical details such as severity, anatomy, histopathology, temporality, and dimensions of injury, among others. These extension codes could improve communication among healthcare providers, facilitate research, and enhance global health monitoring12. Several studies have compared ICD-11 with ICD-10, and most of them focused on the structural changes in disease classification, coding accuracy and training13–22. These studies have generally affirmed that ICD-11 can reflect the latest advancements in the field, with improvements in logical chapter structure, clearer expression, and increased clinical usability. Due to the lack of understanding of the new version and training efficiency, the coding accuracy of ICD-11 is lower than ICD-10. Nevertheless, research based on pilot experiences remains limited in the world. To bridge such a gap, the current study was the first pilot study on the utilization of the ICD-11 for Mortality and Morbidity Statistics (ICD-11 MMS) for coding in West China. As a vast, culturally diverse country with a complex disease spectrum, China faces great challenges and more detailed issues in fully implementing ICD-11. The purpose of this study was to analyze the coding utilization, obstacles, and facilitating factors of ICD-11 in a pilot project of a large tertiary public hospital in Sichuan Province, China, and comprehensively assess the feasibility of implementing morbidity and mortality coding in China with ICD-11 replacing ICD-10, so as to provide valuable insights and references for the transition to ICD-11 in other regions and countries. Methods The pilot project was organized and implemented by the Healthcare Administration and Medical Management Bureau of National Health Commission of China (NHC), which is mainly responsible for formulating and supervising industry policy norms, standards for medical technology applications, medical quality and safety, and medical services. This two-month ICD pilot project aimed to collect information, accumulate experience, gradually expand its implementation nationwide, and align China’s health data standards with international practices. Sichuan, the largest province in the west China, with a large population (83.7 million individuals), and vast geographical area (486,000 square kilometers)23. includes three autonomous prefectures for ethnic minorities. The results of the Sichuan pilot are highly representative due to its broad disease spectrum coverage, including the use of traditional medicine chapter codes24. Sichuan Academy of Medical Sciences·Sichuan Provincial People’s Hospital, as a leading comprehensive hospital in Sichuan province, offers clinical departments covering almost all specialties. Its medical record statistics department was the provincial medical record quality management and control centre designated by the Health Commission of Sichuan Province, making it a ideal choice for a pilot hospital. The coders of this department all have medical education backgrounds, with the majority holding master’s degrees and having an average of 10 years of work experience. The hospital is located in Chengdu, the provincial capital of Sichuan province, a national central city with the seventh largest GDP in China and a total of 4,300 beds. In 2022, it served a total of 5,088,615 outpatient patients and 173,182 inpatients, with nearly 10% of inpatients coming from outside the province. The hospital’s ICD coding work relies on the Medical Records Management System (MRMS), which embedded the ICD-11 coding tool during the pilot period. Coders used this tool for ICD-11 coding. To ensure that it did not interfere with other data reporting and medical insurance settlements, the hospital implemented a dual-track coding system, simultaneously coding both ICD-10 and ICD-11. For ICD-10 coding, we employ the Chinese National Clinical Version 2.0. The pilot consists of six stages (Fig. 1): Pilot process. Organization and mobilization: The hospital leadership organized a initial meeting for the ICD-11 pilot project, with participation from the medical records department, information department, and some clinical physicians. A special task force led by the hospital president was formed, and the tasks, responsibilities, and timelines for the relevant departments were defined. Organization and mobilization: The hospital leadership organized a initial meeting for the ICD-11 pilot project, with participation from the medical records department, information department, and some clinical physicians. A special task force led by the hospital president was formed, and the tasks, responsibilities, and timelines for the relevant departments were defined. Training and examination: The medical records department organized all staff to participate in a two-day ICD-11 online coding training organized by NHC. The training course includes the introduction and basic knowledge of ICD-11, pilot project program, methods for using the Chinese ICD-11 website, main content of every chapter (scope, structure, new features, differences from ICD-10, practical exercises), and the guide for using coding tools. After the training, twenty coders took an online examination and all passed it. Training and examination: The medical records department organized all staff to participate in a two-day ICD-11 online coding training organized by NHC. The training course includes the introduction and basic knowledge of ICD-11, pilot project program, methods for using the Chinese ICD-11 website, main content of every chapter (scope, structure, new features, differences from ICD-10, practical exercises), and the guide for using coding tools. After the training, twenty coders took an online examination and all passed it. Embedding coding tool: The national pilot project uniformly used the ICD-11 disease diagnosis coding tool developed by NHC, a Chinese version similar to the ICD-11 coding tool on the official website of the World Health Organization. The coder could get a search result by searching the disease name or coding. The tool would automatically pop up the possible cluster and extension code, and it also provides personal coding collection function. It enables users to add frequently used ICD-11 codes (especially the complex codes generated by manual postcoordination) to the collection for quick retrieval. The information department collaborated with the medical records department to embed the ICD-11 disease diagnosis coding tool into the medical records management system. This enabled data retrieval, storage, and transmission of ICD-11 within the system and supported dual-track coding with both ICD-10 and ICD-11. Coders can access the ICD-11 coding tool interface by clicking on the blank space for disease coding after reading electronic medical records. Embedding coding tool: The national pilot project uniformly used the ICD-11 disease diagnosis coding tool developed by NHC, a Chinese version similar to the ICD-11 coding tool on the official website of the World Health Organization. The coder could get a search result by searching the disease name or coding. The tool would automatically pop up the possible cluster and extension code, and it also provides personal coding collection function. It enables users to add frequently used ICD-11 codes (especially the complex codes generated by manual postcoordination) to the collection for quick retrieval. The information department collaborated with the medical records department to embed the ICD-11 disease diagnosis coding tool into the medical records management system. This enabled data retrieval, storage, and transmission of ICD-11 within the system and supported dual-track coding with both ICD-10 and ICD-11. Coders can access the ICD-11 coding tool interface by clicking on the blank space for disease coding after reading electronic medical records. Implementation: The dual-track coding of ICD-10 and ICD-11 was completed for all discharged cases from February 1 to March 31, 2022. In terms of coding audits, we followed our standard three-level audit system for both versions of coding. Level one audit involved cross-auditing, level two audit focused on key cases (such as deaths, surgeries, rare diseases, etc.), and level three audit was a recheck of the previous audit results. The coded case data, including existing front page data, clinical diagnosis, and associated ICD-11 codes with descriptions, were submitted to NHC in required format. Implementation: The dual-track coding of ICD-10 and ICD-11 was completed for all discharged cases from February 1 to March 31, 2022. In terms of coding audits, we followed our standard three-level audit system for both versions of coding. Level one audit involved cross-auditing, level two audit focused on key cases (such as deaths, surgeries, rare diseases, etc.), and level three audit was a recheck of the previous audit results. The coded case data, including existing front page data, clinical diagnosis, and associated ICD-11 codes with descriptions, were submitted to NHC in required format. Questionnaire survey and focus interview: Questionnaire surveys and focused interviews with the 20 coders participating in the pilot project were conducted. Their experiences, coding obstacles encountered during the use of ICD-11, and perspectives on future implementation were collected. The quality control expert group of the medical records department then carefully reviewed the collected content. The selection criteria for experts were having a medical professional background, having more than 10 years of coding work experience, holding a senior professional title, and being familiar with various clinical professional knowledge. Questionnaire survey and focus interview: Questionnaire surveys and focused interviews with the 20 coders participating in the pilot project were conducted. Their experiences, coding obstacles encountered during the use of ICD-11, and perspectives on future implementation were collected. The quality control expert group of the medical records department then carefully reviewed the collected content. The selection criteria for experts were having a medical professional background, having more than 10 years of coding work experience, holding a senior professional title, and being familiar with various clinical professional knowledge. Analysis and summary: Due to the time-consuming nature of ICD-11 coding and a lack of unified coding audits rules, this study extracted and analyzed the dual coding of primary disease diagnoses used during the pilot period to ensure coding quality. The primary diagnosis coding was more accurate and reliable. For statistical analysis, we used Stata 17 to organize the database and conducted a comparative analysis of chapter distributions of primary diseases between the two coding versions using cross-tabulation tables (with a significance level of P < 0.05). Descriptive analysis was performed for the application of postcoordination and extension code dimensions of ICD-11, coding obstacles, coder experience, and facilitators of implementation. This comprehensive evaluation aimed to assess the feasibility of using ICD-11 as a replacement for ICD-10 coding and summarized lessons learned during implementation, aiming to provide implementation recommendations. Analysis and summary: Due to the time-consuming nature of ICD-11 coding and a lack of unified coding audits rules, this study extracted and analyzed the dual coding of primary disease diagnoses used during the pilot period to ensure coding quality. The primary diagnosis coding was more accurate and reliable. For statistical analysis, we used Stata 17 to organize the database and conducted a comparative analysis of chapter distributions of primary diseases between the two coding versions using cross-tabulation tables (with a significance level of P < 0.05). Descriptive analysis was performed for the application of postcoordination and extension code dimensions of ICD-11, coding obstacles, coder experience, and facilitators of implementation. This comprehensive evaluation aimed to assess the feasibility of using ICD-11 as a replacement for ICD-10 coding and summarized lessons learned during implementation, aiming to provide implementation recommendations. All procedures had been performed in accordance with the guidelines of the Declaration of Helsinki. Results Changes after using ICD-11 A total of 25,879 patients were discharged during the entire pilot period, and 20 coders participated in dual-track coding and completed 137,998 ICD-11 codes. From the perspective of using ICD-11 for primary disease coding, stem codes covered 22 out of 28 chapters, and 19 ICD-11 chapters can be directly mapped to the ICD-10 chapter (Table 1). Among the six chapters that were not covered, the chapters Sleep-wake disorders, Conditions related to sexual health, Supplementary section for functioning assessment (three unique chapters in ICD-11) and the chapter Codes for special purposes were not involved in the pilot due to a short duration because there were no inpatients with a primary diagnosis of the diseases described in these chapters. The chapter External causes of morbidity or mortality, which served as an additional code in ICD-10, retained a similar position in ICD-11 but cannot be used as a main disease code. The chapter Extension Codes, unique to ICD-11, primarily aimed to overcome the shortcomings of previous versions such as ambiguity in left-right, laterality, severity scale values, and positions. They can only be used in conjunction with stem code through postcoordination. Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism in ICD-10 were split into two chapters of Diseases of the blood or blood-forming organs and Diseases of the immune system in ICD-11. The Supplementary Chapter Traditional Medicine Conditions - Module I, an exclusive chapter in ICD-11, showed a usage frequency of 0.16%. Eight chapters, including Diseases of the blood or blood-forming organs, Diseases of the immune system, Endocrine, nutritional or metabolic diseases, Diseases of the nervous system, Diseases of the circulatory system, Diseases of the digestive system, Diseases of the musculoskeletal system or connective tissue, and Symptoms, signs, or clinical findings, not elsewhere classified, were significantly affected (P < 0.05) during the transition from ICD-10 to ICD-11. After using ICD-11 coding, the proportion of patients classified under Diseases of the nervous system was increased by 3.05%, while that of patients classified under Diseases of the circulatory system was decreased by 4.23%. The remaining thirteen chapters were not significantly affected, with a 0% rate of coding changes in four chapters, namely, Diseases of the respiratory system, Pregnancy, childbirth or the puerperium, Certain conditions originating in the perinatal period, and Factors influencing health status or contact with health services. Table 1Two versions of patients’ main disease codes, statistics by chapter.ChapterChapter name of ICD-11Patient Coverage(%)Chapter name of ICD-10Patient Coverage(%)χ P 1Certain infectious or parasitic diseases1.92Certain infectious and parasitic diseases2.010.5310.4662Neoplasms14.59Neoplasms14.580.0010.9703Diseases of the blood or blood-forming organs0.57Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism0.70147.057< 0.0014Diseases of the immune system1.345Endocrine, nutritional or metabolic diseases1.91Endocrine, nutritional and metabolic diseases2.184.7280.0306Mental, behavioural or neurodevelopmental disorders0.53Mental and behavioural disorders0.510.0600.8067Sleep-wake disorders0.008Diseases of the nervous system6.03Diseases of the nervous system2.98280.251< 0.0019Diseases of the visual system6.28Diseases of the eye and adnexa6.260.0080.92810Diseases of the ear or mastoid process0.93Diseases of the ear and mastoid process0.950.0750.78511Diseases of the circulatory system9.17Diseases of the circulatory system13.40230.933< 0.00112Diseases of the respiratory system6.28Diseases of the respiratory system6.28--13Diseases of the digestive system12.99Diseases of the digestive system12.394.1900.04114Diseases of the skin1.00Diseases of the skin and subcutaneous tissue1.131.8750.17115Diseases of the musculoskeletal system or connective tissue3.29Diseases of the musculoskeletal system and connective tissue4.5655.471< 0.00116Diseases of the genitourinary system6.98Diseases of the genitourinary system6.930.0500.82217Conditions related to sexual health0.0018Pregnancy, childbirth or the puerperium5.78Pregnancy, childbirth and the puerperium5.78--19Certain conditions originating in the perinatal period0.79Certain conditions originating in the perinatal period0.79--20Developmental anomalies1.22Congenital malformations, deformations and chromosomal abnormalities1.270.0000.98521Symptoms, signs or clinical findings, not elsewhere classified2.51Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified1.5658.188< 0.00122Injury, poisoning or certain other consequences of external causes2.77Injury, poisoning and certain other consequences of external causes2.790.0260.87323External causes of morbidity or mortality0.00External causes of morbidity and mortality0.00--24Factors influencing health status or contact with health services12.96Factors influencing health status and contact with health services12.96--25Codes for special purposes0.00Codes for special purposes0.00--26Supplementary Chapter Traditional Medicine Conditions - Module I0.16VSupplementary section for functioning assessment0.00XExtension Codes0.00 Two versions of patients’ main disease codes, statistics by chapter. When coders reviewed the medical record and found that the stem code could not fully express the clinical concept of the case, it was necessary to use postcoordination. The proportion of the codes used postcoordination was 41.10% and the maximum number of additional codes used for a single clinical expression reached eight. Among all the codes used postcoordination, the average number of postcoordination was 1.5. The three most frequently used types of postcoordination codes were XK9K, XK8G, and XK9J. 64.88% of the coding results used one postcoordination, 21.63% used two postcoordinations, and 11.75% used three postcoordinations. Notably, we observed 5.70% of postcoordination only used stem codes, 31.42% only uses extension codes, and the remaining 3.98% concurrently employed both types. Three chapters with the highest frequency of postcoordination usage were Chap. 9 Diseases of the visual system, Chap. 22 Injury, poisoning or certain other consequences of external causes, and Chap. 10 Diseases of the ear or mastoid process. Among them, 97% of ophthalmic diseases and 77% of ear and mastoid process diseases with clear medical record information used codes XK9J (bilateral), XK8G (left side), and XK9K (right side) to indicate laterality. In addition, when the external causes were clear, 90% of patients with injuries or poisoning used codes from the chapter External causes of morbidity or mortality as additional codes for supplementary explanations. The three most frequently used extension code dimensions in the “X” chapter were Topology Scale Value, Anatomy and topography, and Histopathology, accounting for 81% in total. This was because the tumor chapter required the use of Histopathology to complement its pathological type and differentiate between malignant and benign conditions. Three unused dimensions were Consciousness, Capacity or context, and Health Devices, Equipment and Supplies (Table 2; Fig. 2). Overall, after the adoption of ICD-11 coding, the overall classification structure was relatively stable, with 92.30% of the primary disease codes belonging to the same chapters as before. A few chapters were affected by the coding version update, 41.10% of coding used postcoordination to express completeness (Fig. 3). Table 2Postcoordination application in each chapter.ChapterNumber of codesProportion of postcoordination application(%)Postcoordination of stemCodes only (%)Postcoordination of extension codes only(%)The maximum number of postcoordination149918.841.4016.8332377669.760.0569.60531476.120.683.404434722.4814.997.491549461.7460.930.81261375.114.380.0038156136.583.0132.9349162597.110.0097.1121024076.670.0076.67211237442.888.6822.79812162530.460.3129.78413336128.446.6920.055142608.850.008.8531585166.635.9950.41716180643.7412.6830.95318149529.8315.3211.776192050.980.490.4912031621.520.0021.5222164915.250.3014.9532273192.0715.054.1082433540.810.000.81126260.001.4016.830total25,87941.105.7031.42- Postcoordination application in each chapter. Application of extension codes with different dimensions. ECPRSD: Extension codes of particular relevance to skin diseases; HDES: Health Devices, Equipment and Supplies. The usage of extension codes corresponds to the dimensions employed by ICD-11 to supplement critical diagnostic information in detail. Comparison of chapters of patients’ main diseases. Coding obstacles We requested each coder to document various coding obstacles encountered and organized experts to review them. The following is a summary of coding obstacles during the entire dual-track coding work (based on ICD-11 version 01/2025): Lack of sub-chapter code range identification: Compared to ICD-10, ICD-11 lacked clear identification of code ranges for sub-chapters. For example, the sub-chapter “Gastroenteritis or colitis of infectious origin” under Chap. 01 did not display the inclusive code range it contains in the same way as the sub-chapter “Intestinal infectious diseases (A00 - A09)” under Chapter I of ICD-10. Lack of sub-chapter code range identification: Compared to ICD-10, ICD-11 lacked clear identification of code ranges for sub-chapters. For example, the sub-chapter “Gastroenteritis or colitis of infectious origin” under Chap. 01 did not display the inclusive code range it contains in the same way as the sub-chapter “Intestinal infectious diseases (A00 - A09)” under Chapter I of ICD-10. Incomplete or missing preset postcoordination: This was the most common issue encountered. For instance: (1) Sub-codes under “GA07 Salpingitis and oophoritis” did not have unilateral or bilateral postcoordination. (2) In the code “ND72 Foreign body in respiratory tract,” only “ND72.5 Foreign body in bronchus” lacked postcoordination for anatomical location. (3) There was also a lack of sub-codes related to medical abortion under the code “JA00 Abortion”. Incomplete or missing preset postcoordination: This was the most common issue encountered. For instance: (1) Sub-codes under “GA07 Salpingitis and oophoritis” did not have unilateral or bilateral postcoordination. (2) In the code “ND72 Foreign body in respiratory tract,” only “ND72.5 Foreign body in bronchus” lacked postcoordination for anatomical location. (3) There was also a lack of sub-codes related to medical abortion under the code “JA00 Abortion”. Inaccurate retrieval results: “J35.0 Chronic tonsillitis” can only be classified as “CA0F.Y Other specified chronic diseases of tonsils or adenoids” in ICD-11, which was less accurate than the previous code. Inaccurate retrieval results: “J35.0 Chronic tonsillitis” can only be classified as “CA0F.Y Other specified chronic diseases of tonsils or adenoids” in ICD-11, which was less accurate than the previous code. Time-consuming coding process: There were three reasons for the increased time required: (1) Coders still used ICD-10 thinking mode for coding as they were unfamiliar with ICD-11; (2) User-unfriendly use of coding tools often required multiple clicks to complete a cluster; (3) Due to a finer granularity of ICD-11, many diagnoses required multiple postcoordination codes to fully express, necessitating the retrieval of more clinical record details for effective coding and frequent communication with clinical physicians. Time-consuming coding process: There were three reasons for the increased time required: (1) Coders still used ICD-10 thinking mode for coding as they were unfamiliar with ICD-11; (2) User-unfriendly use of coding tools often required multiple clicks to complete a cluster; (3) Due to a finer granularity of ICD-11, many diagnoses required multiple postcoordination codes to fully express, necessitating the retrieval of more clinical record details for effective coding and frequent communication with clinical physicians. Coder experience From the results of face-to-face questionnaires with 20 coders, 70% of them considered that ICD-11 was more accurately and could clearly express clinical diagnoses in comparison to ICD-10, and 40% strongly agreed with this statement. After becoming proficient in using ICD-11 and with comprehensive coding tools, 40% of the respondents believed that coding with ICD-11 would not be more user-friendly, while 30% thought they were the same in terms of operability. Regarding the choices for pre-assigned postcoordination, 60% found them user-friendly, while 10% did not. In terms of expressing complete diagnoses, 75% of the coders agree that a large number of medical records require manual postcoordination, while 25% express strong agreement. As for the quality of clinical record writing in the hospital, 45% of the respondents believed it was sufficient for the implementation of ICD-11, while the rest held opposite views. In terms of coding audits for ICD-11 coding, 65% of the coders found it challenging, 20% found it very challenging, and only 5% found it easy to implement (Fig. 4). Questionnaire survey results on coder experience. Facilitators of implementation One-on-one interviews with coders identified and summarized the following six facilitators of ICD-11 implementation: improving the quality of medical record writing, conducting specialized ICD-11 training, utilizing intelligent coding tools, releasing the official ICD-11 textbooks of china, establishing unified coding audits standards, and supporting health policy reforms. The detailed contents were summarized in the table, and will be further discussed and explained in the discussion sections (Table 3). Table 3Facilitators of ICD-11 implementation.FacilitatorsImplementation detailsImproving the quality of medical record writingλ Complete and accurate medical record informationλ Reduced reading time for codersλ Improved coding accuracy and efficiencyConducting specialized ICD-11 trainingλ Repetitive trainingλ Encouraging doctors to actively provide ICD-11-required medical record elementsλ Establishing a national and provincial teacher training databaseλ Creating a national-provincial-city-hospital training networkUtilizing intelligent coding toolsλ Improving response speedλ Simplified coding processλ Adding a personal coding collection functionλ Introducing AI learning to enhance retrieval efficiencyReleasing the official ICD-11 textbooks of chinaλ Breaking down knowledge barriers of ICD-11λ Professional, and effective self-studyEstablishing unified coding audits standardsλ Guaranteeing coding auditsλ Enhancing coding qualityλ Standardizing medical record writing and coding workRelevant medical policy reformsλ Disciplinary professional evaluationλ Performance assessmentλ Health insurance payment Facilitators of ICD-11 implementation. λ Complete and accurate medical record information λ Reduced reading time for coders λ Improved coding accuracy and efficiency λ Repetitive training λ Encouraging doctors to actively provide ICD-11-required medical record elements λ Establishing a national and provincial teacher training database λ Creating a national-provincial-city-hospital training network λ Improving response speed λ Simplified coding process λ Adding a personal coding collection function λ Introducing AI learning to enhance retrieval efficiency λ Breaking down knowledge barriers of ICD-11 λ Professional, and effective self-study λ Guaranteeing coding audits λ Enhancing coding quality λ Standardizing medical record writing and coding work λ Disciplinary professional evaluation λ Performance assessment λ Health insurance payment Discussion This paper evaluated the pilot implementation of ICD-11 coding based on real-world data and assessed the feasibility of implementing ICD-11 in China by analyzing changes in classification chapters after ICD-11 coding, coder experience, coding obstacle, and facilitators. The classification results of the 13 chapters were hardly affected by the version changes, while there are significant differences in the proportions of Diseases of the circulatory system and Diseases of the circulatory system after the application of ICD-11. Chapter movement was mainly a result of the shift of stroke, cerebral hemorrhage, and other cerebrovascular diseases from Diseases of the circulatory system to Diseases of the nervous system. The underlying reason was that the first clinical symptoms of stroke and cerebral hemorrhage were neurological symptoms, and long-term misclassification has led to a lag in the allocation of medical resources and training of neurological professionals, which fully demonstrated that ICD-11 was more suitable for medical development and technological progress in terms of classification structure17,25. Although most coders in the questionnaire survey considered the use of ICD-11 not user-friendly, it was also a major reason why ICD-11 was more progressive in comparison to ICD-10. One long cluster code can express rich diagnostic and therapeutic information that was not possible with ICD-10. Almost all chapters used Topology Scale Value and Anatomy and topography to express key information such as distribution, laterality, functional anatomy, and surface topography. Nearly 60% of the coders found the pre-assigned postcoordination efficient. However, when users needed to manually perform postcoordination on the code, it was often time-consuming and affected work efficiency. Overall, the transition of chapter classification structure from ICD-10 to ICD-11 was relatively stable. The usage and acceptance of postcoordination were positive, with 92.30% of coded chapters remaining unchanged. In terms of code expression in ICD-11, relevant studies have shown that 24.3% of codes can be fully expressed in the transition from ICD-9-CM codes to ICD-10-CM. In the transition from ICD-10-CM to ICD-11, 23.5% can be fully expressed without using postcoordination. After improving the postcoordination, this figure was further increased to 58.7%, which indicated that the transition from ICD-10 to ICD-11 coding may be more stable than previous version updates26,27. However, in terms of coding accuracy, research conducted in Iran and South Korea showed that the coding accuracy of ICD-11 was 74.2% and 71.6% respectively, both lower than that of ICD-1018,19. We also found some obstacles in using the ICD-11 coding process. This version lacked subdirectory coding range identifiers though the code capacity of ICD-11 has been significantly increased, and this was user-unfriendly for users, especially during the promoting stage. We suggest creating clearer and more explicit subdirectory range identifiers similar to ICD-10. As mentioned in the results section, we also encountered issues with incomplete or missing preset laterality coordination. Specifically, there were missing laterality coordination elements, which increased the possibility of incomplete description of the disease profile. Thus, it is hoped that the ICD-11 maintenance team will investigate and make improvements to accommodate the code requirements in different countries. Furthermore, due to the reclassification of diseases in ICD-11, the search results for certain codes were inaccurate, which demanded further analysis and comparison. However, the proficiency levels of coders varies across the nation, and incorrect search results may result in basic errors. Accurate search results can significantly improve the efficiency and accuracy of ICD-11 coding. Therefore, testing and improving the search functionality should be a priority for developers. In addition, this study also investigated major reasons accounting for increased time required by ICD-11 coding, which included insufficient familiarity with the coding system, less user-friendly coding tools, and finer granularity. Surveys and interviews were conducted to further understand the opinions of coders participating in the pilot program, and most coders reported that ICD-11 more accurately represented clinical diagnoses compared to ICD-10. They found the pre-assigned postcoordination provided in ICD-11 was user-friendly, but still many cases required manual postcoordination, which made coding audits challenging. Half of the coders did not think that ICD-11 was more user-friendly for coding compared to ICD-10, even with improved proficiency and coding tools. It was noted that the current quality of medical record writing in the hospital demanded further improvement to comply with the implementation of ICD-11. Based on these findings, we identified six facilitators to promote the implementation of ICD-11, including aspects such as the quality of medical record writing, training, intelligent coding, textbooks, auditing standards, and health policies. Implementing ICD-11 would set higher requirements for the quality of medical record content. Comprehensive and accurate medical record information was believed to be the foundation for efficient coding. The pilot program showed that incomplete and inaccurate medical records often led to an inability to use more comprehensive postcoordination to express the diagnosis and treatment process. Before transitioning to ICD-11, it would be necessary to establish a national-provincial-city-hospital training network and provide repeated training for coders and clinical physicians because they have been accustomed to coding using the ICD-10 for many years. This will not only improve the familiarity and practical skills of coders with ICD-11, but also encourage clinical physicians to pay more attention to higher-quality medical record writing for ICD-11 coding. Relevant studies conducted in Korea and Iran indicate that the coding accuracy in the early stages of ICD-11 application is lower than that of ICD-10, necessitating effective and comprehensive training for coders before nationwide implementation18,19. Although the current ICD-11 tools can support coding activities, they were not user-friendly. The responsiveness of the tools was dependent on the hospital’s information technology capabilities. Thus, we suggest adding a personal coding collection function to the tools, allowing coders to quickly find manually postcoordinated codes or more complex codes in the future. Additionally, the coding tools and workflow should be optimized from a user experience perspective to make them more user-friendly. In this case, AI could also be considered in tool updates because AI can accurately identify corresponding codes even with vaguely worded search terms through continuous learning. Moreover, releasing an official textbook could also support a nation-wide promotion of ICD-11, and in this way, we could break knowledge barriers and enable all coders and university students to acquire ICD-11 knowledge through self-study, even if they are unable to participate in professional training. Currently, ICD-11 lacked effective and unified coding audits rules. It is essential to promptly establish corresponding rules to standardize medical record writing and coding work and improve coding quality. Lastly, to ensure a smooth implementation of the new coding version, healthcare policy reforms related to performance assessment, health insurance payments, and discipline evaluation should be synchronously promoted by multiple relevant departments. At the same time, it is also necessary to make adjustments to the coding version used for data extraction during the implementation of these policies. Limitations and future work The data were collected from a large general hospital, and although there was a wide variety of patient population and types, the results may not be representative of all healthcare institutions in China. Moreover, the survey and interview results mainly reflect the subjective opinions of coders. Regional differences may limit the generalizability of our findings even we have carefully reviewed and analyzed the content with experts. However, these results can still serve as valuable references. In the future, with the strengthening of the quality management of medical record writing at the national level and the continuous improvement of the quality of medical records, coders will need to read medical records in more details and acquire comprehensive knowledge to balance information for presenting the diagnosis and treatment process, medical quality and safety management, healthcare payment, and performance evaluation. Coding granularity should be adjusted based on different purposes. The implementation of ICD-11 will undoubtedly pose a significant challenge for coders. Additionally, even with the introduction of more AI in coding, the selection of modified codes in ICD-11 may still be subjective, experiential, and professional that could limit its application. Conclusions Overall, the evolution from ICD-10 to ICD-11 was considered to be smoother in comparison to previous version upgrades. Apart from the time-consuming and complex coding process, some coding obstacles needed to be addressed, but still coders acknowledged the improvements of ICD-11. Implementing ICD-11 to replace ICD-10 was considered feasible with solid groundwork for improving the quality of medical records, professional training, and release of official guidelines. Supplementary Information Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Author contributions Jinliang Hu and Xingyu Ming conceived and designed the study. Xingyu Ming and Han wei performed article writing and statistical analysis. Hongying Li and Jinyuan Tao performed dual-track coding and quality review. Zhou Die and Bangfu qian performed transformation of the information system. All authors contributed substantially to its revision. Data availability The data that support the findings of this study are available from Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are, however, available from the authors upon reasonable request and with permission of Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China. Declarations Competing interests The authors declare no competing interests. Ethics approval and consent to participate All individuals and/or their parents provided informed consent to participate in this study and approval was provided by Medical Research Ethics Committee of Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors share first authorship: Xingyu Ming and Han Wei Contributor Information Hongying Li, Email: 2508107446@qq.com. Jinliang Hu, Email: hu_scsy@126.com. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Data Availability Statement The data that support the findings of this study are available from Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are, however, available from the authors upon reasonable request and with permission of Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China.