ICD-11 Field trial in Japan

Following the guideline and the platform for ICD-11 FT (ICD-11 Fit) offered by WHO, an ICD-11 Field Trial (FT) in Japan was performed from Aug to Sept 2017 targeting information managers who have a crucial role in coding in hospitals and clinics in Japan. Three hundred seventy-eight health information managers participated as raters with the cooperation of the Japan Health Information Manager Association and the Japan Hospital Association. In the ICD-11 FT, two types of evaluations were performed:  line coding (LC - coding of diagnostic terms) and case coding (coding of case scenarios). Here we report an evaluation of LC.

In LC, 298 diagnostic terms from 22 chapters were provided. Table 1 shows a chapter breakdown of diagnostic terms at the time of performing ICD-11 FT. Raters coded each term with both ICD-10 and ICD-11, then evaluated the difficulty of coding, the level of detail of the code, and the level of ambiguity of the code as an individual evaluation. Coding times were measured automatically. The raters who completed all diagnostic terms went on an overall evaluation. We compared the results of individual evaluations among chapters, then summarized overall evaluations. The ICD-11 field trial was performed in English, but we prepared some translation documents for platform and coding guidance.  22.0% of participants completed coding of all diagnostic terms, 56.9% did not finish all of the coding, and unfortunately about 20% did not start. The median number of coded diagnostic terms of participants who did not finish all the coding was 32.0 diagnostic terms.

Comparing between ICD-10 and ICD-11 among chapters (Table 2)
The percentage of respondents who had difficulties in ICD-11 coding was higher in 19 chapters including chapter 1 (infectious disease), 2 (neoplasms), and 3 (blood disease) than those of in ICD-10 coding. There were no differences in chapter 4 (immune disease), 9 (visual disease), and 20 (developmental anomalies) between ICD-10 and ICD-11. In additional analysis, we focused on the differences of the percentage between ICD-11 and ICD-10 by calculating the odds ratio. We found that there were big differences between ICD-10 and ICD-11, which showed more than ten odds, in chapter 3 (blood disease), 8 (nervous system), 16 (genitourinary system), 18 (pregnancy), 19 (perinatal period), and 22 (injury).

Even in comparison with the level of detail of the code, and the level of ambiguity of the code, the evaluation of ICD-10 was better as a whole. However, the evaluation of detail and ambiguity of ICD-11 in chapter 4 (immune system) and 7 (sleep disorder) was better than that of ICD-10.

75 raters gave an overall evaluation. 45.3% said the rate of the coverage of ICD-11 was good or very good, and 73.3% said the level of detail in ICD-11 was just right. Free comments showed the level of detail was different among chapters. 46.7% said that they had difficulty using ICD-11. The main reason for this was that the FT was performed in English and they were not given adequate education about ICD-11 and its tools. 37.3% said the electronic tools were good, however, they needed more accurate fuzzy string searching.

We believe that characteristics of chapters are crucial information from the point of ICD-11 introduction and education in clinical field. The difficulty, detail and ambiguity evaluation of ICD-11 showed differences in each chapter. More detailed analysis among chapters will be needed.

The biggest bottleneck in this Filed Trial was the lack of English language skills and understanding of the coding system of ICD-11 due to it. In our additional questionnaire, about 70% of participatts said they did not have confidence in their English skills. Securing translation quality as well as support and education for understanding the coding system will be the key to success when introducing ICD-11 in Japan.

This study was funded from Ministry of Health, Labour and Welfare of Japan. We wish to thank the timely help given by health information managers, the Japan Health Information Manager Association, and the Japan Hospital Association.

Table 1: A chapter breakdown of 298 diagnostic terms in a line coding

No Name of chapter n (%)

For information:
Yoko Sato, National Defense Medical College Japan/National Institute of Public Health Japan, Japan WHO-FIC Collaborating Centre, email: ysato@ndmc.ac.jp 
Hiroshi Mizushima, National Institute of Public Health Japan, Japan WHO-FIC Collaborating
Hiromitsu Ogata, National Institute of Public Health Japan, Japan WHO-FIC Collaborating/Kagawa Nutrition University, Japan
Kei Mori, Ministry of Health, Labour and Welfare of Japan