【Introduction】In recent years, the number of foreign patients visiting medical clinics in depopulated areas as well as in urban areas has been increasing in Japan. This can be attributed to factors such as the increased acceptance of foreign technical trainees to compensate for the labor shortage, as well as a rise in foreign tourists visiting Japan. On the other hand, the information obtained through medical interviews is crucial for the diagnosis and treatment of diseases. In response to this situation, the Ministry of Health, Labour and Welfare created a manual for medical institutions. And medical institutions nationwide have been working to improve their system for accepting foreign patients. In fact, the Ministry has published multilingual explanatory materials for foreigners on its website. However, there are limitations. For example, in the headache questionnaire, it does not include crucial information for differentiating headache classifications, such as the location, characteristics, impact on daily life, prodromes or auras. With the aim of addressing these limitations, we proposed the creation of a multilingual headache questionnaire using the translation function of generative AI and discussed the possible features and challenges.
【Materials and Method】We examined the trends in the resident foreign population in our region and the foreign patients at our hospital from 2015 to 2022. We utilized a generative AI to translate the headache questionnaire that our department had previously created. The 10 selected languages for translation were based on the nationalities of foreign patients who frequently visited our hospital.
【Results】 The trends of the resident foreign population and the foreign patients showed an increase from around 2015 and continued to rise until the COVID-19 pandemic in 2020, with a strong correlation between the two. The nationalities of foreign patients spanned 40 countries over the entire period, highlighting the diversity and multilingual nature of the patient population. We edited the questionnaire by listing the translated results and the original Japanese for each question item. As a result, it became possible to understand the results of the medical interview even if the foreign patient’s language was completely incomprehensible.
【Discussion】This approach can accommodate all languages, including rare ones, and proves valuable for native languages. It allows for a more precise assessment of the interview content, increasing its uniqueness. It can also be applied to various other questionnaires, such as headache diaries, making it highly versatile. It can quickly adapt to unexpected visits, and is efficient. AI-generated translation has been noted that human oversight is necessary due to issues such as literal translation errors, cultural nuances being missed and mistranslation of specialized terminology. However, since the language used in medical questionnaires is at a conversational level, it is believed that significant issues can be avoided.
【Conclusion】 This approach is expected to enhance diagnostic efficiency and reduce the risk of information gaps or shortages among foreign patients, which is expected to be increasingly valuable in the diagnosis and treatment of foreign patients.
View full abstract