Search across all content
projectThis system features an extensive idiomatic expression database and upon entering a primary symptom, utilizes AI-based neural network technology to present relevant follow-up questions regarding associated symptoms.
This information was sourced from "SIMTOMI: AI-Powered Multilingual Medical Support" by Observatory of Public Sector Innovation (OECD), https://oecd-opsi.org/innovations/simtomi-ai-powered-multilingual-medical-support/. Licensed under Creative Commons Attribution-ShareAlike 3.0 IGO (CC BY-SA 3.0 IGO).
Incheon City, with the Overseas Koreans Agency and Incheon International Airport, has South Korea's highest foreigner proportion among metropolitan cities at 4.9%, housing over 140,000 foreign residents. Surveys indicate that the main challenge for these residents is the language barrier, especially in medical contexts. Limited Korean proficiency often leads to reliance on inaccurate translations or gestures, risking inadequate medical care. Medical professionals face similar challenges with online translators when treating foreign patients.
To improve healthcare for multi-national residents and boost the local AI industry, Incheon City has partnered with Mediark Inc. They support the development of the AI-based multilingual medical support service SIMTOMI, addressing medical benefit asymmetries across languages and nations. This service innovates in four key areas.
Multilingual AI Symptom Checker: This system features an extensive idiomatic expression database and upon entering a primary symptom, utilizes AI-based neural network technology to present relevant follow-up questions regarding associated symptoms. Patients can select the symptoms that best match their condition from these questions. Additionally, the system includes a visual human body model to aid in accurately and intuitively indicating the location of pain, allowing users to express their symptoms more precisely.
As of January 2024, the service offers English, Chinese, Japanese Vietnamese, and Russian.
The goal is to assist foreigners and tourists in Korea to effectively communicate symptoms and receive proper medical care, overcoming language barriers. SIMTOMI also aids medical professionals in understanding patients' symptoms more efficiently, enhancing overall medical service efficiency. The service targets foreigners residing in Korea, international students, and tourists facing language challenges in healthcare settings. It's also useful for Koreans abroad needing to convey symptoms or purchase over-the-counter medication in foreign pharmacies and hospitals.
The future vision includes expanding the service area and adding more languages to reach a broader audience. The potential for global use, with cross-language compatibility for symptom occurrences, is being explored. The aim is to build a digital, safe, and healthy city by partnering with private companies to connect apps with medical systems.
Medical services for non-native speakers are often overlooked in many local jurisdictions. Previously, the other local government once tried pharmacy kiosks where users reported symptoms by choosing from a few symptom images. Conventional translation tools are not tailored for medical terminology, leading to a risk of mistranslation. Recently, although numerous symptoms check services exist, the majority are not powered by AI, and none of them offer the feature of summarizing symptoms.
This innovative service utilizes a search system capable of processing thousands of symptom expressions from various languages and employs an AI-driven questioning algorithm using advanced neural networks. It also incorporates technology for deriving a list of related diseases based on symptom similarity. Additionally, it provides appropriate language-specific expression matching in summaries, significantly lowering the risk of mistranslation and offering real-time translation.
Incheon Metropolitan City is continuously advancing projects to develop and demonstrate AI-based services, aiming to nurture local AI-based IT companies and enhance the convenience of medical services for foreigners. This service has attracted interest from some other regional city governments in south korea as part of their efforts to build a healthy city.
In collaboration with the Pharmaceutical Society, "Foreigner-Friendly Pharmacy" stickers are being distributed to affiliated 1600 pharmacies and to all pharmacies in Incheon. Non-native Korean speakers located in Korea can immediately find out the location of pharmacies open at night and on holidays, as well as hospitals for various medical specialties, through the SIMTOMI app.
The technology used in this service has achieved the goal of proof of concept with ISO 13485 certification for medical devices and several patent applications.
The Incheon City selected Mediark Inc., a medical IT company based in Incheon, for the implementation of this service. Mediark Inc. was deemed suitable due to its numerous patent applications and ownership of AI-based medical questioning technology. To ensure the smooth establishment of the service, collaborations were formed with the Incheon Medical Association, the Pharmaceutical Association, multicultural family centers, and Incheon Global Campus Foundation to execute this project.
Foreign citizens can feel medically secure within the city by using SIMTOMI. They'll get disease lists related to their symptoms, specialty guidance, and location services to find suitable medical facilities. Medical staff benefit from quick symptom comprehension and enhanced efficiency in medical record-keeping. Pharmacists receive support in accurately recommending over-the-counter medications, while government officials can contribute to creating a medically safer international city.
The evaluation of the service was conducted via surveys among doctors, pharmacists, and patients. It focused on satisfaction, reuse rates, and referral proportions. Satisfaction and intention to reuse targets were set at over 70 points and 80% respectively. The results achieved were 84.4 points for satisfaction and a 95.0% intention to reuse, meeting the pilot project's objectives. Our ultimate goal is to statistically prove maximized clinical efficiency by reducing consultation times and enhancing collaboration efficiency with various hospitals. We aim to increase clinical efficiency by over 20%, focusing on reducing medical consultation and charting times.
Symptoms that require extensive physical examination, such as those often seen in dermatology, had limitations when assessed only through symptom questioning by text. To address this, we added a photo uploading function before the final summary, allowing doctors to refer to visual data for a better understanding. We also Furthermore, Pharmacies requested a more streamlined service to address the issue of lengthy time consumption in symptom checks that involve detailed medical questioning. Therefore, we developed a database for over-the-counter medications, enabling direct communication of desired medications and associated symptoms for improved service delivery.
To guarantee the success of this innovation, considering the diversity of foreign nationalities and expanding language support is key. We aim to support 10-15 languages for a more comprehensive service. Additionally, a two-way medical communication feature is essential, one that translates doctors' diagnoses and test explanations directly into the patient's native language, thereby minimizing the use of traditional translator. Furthermore, To enhance the use of multilingual medical support services, active collaboration and promotion with medical institutions like hospitals and pharmacies are necessary. Continuous improvement of functionalities for better service provision and ensuring adequate financial support are also crucial.
The field of medicine, including medical questions and diagnoses about symptoms, follows globally shared and agreed-upon guidelines, making it suitable for globalization. The only difference lies in the language of expression. This aspect provides our innovation with significant global scalability. Initially developed for the unique demographics of Incheon, it is versatile enough for any area with diverse populations. Other regions can also adopt SIMTOMI to improve medical access. Especially strong in multilingual environments, it ensures consistent analysis of physical symptoms, enhancing healthcare quality. This system's adaptability makes it valuable for improving medical services in varied linguistic and cultural contexts globally.
The usage rates of the service may vary depending on the healthcare insurance and system of each country, as well as the proportion of non-native speakers. In South Korea, the treatment of foreigners can pose challenges for doctors due to the longer duration of work and consultations, which occur without an increase in medical fees. This creates an ongoing obstacle in promoting care for foreign patients. If systems were implemented to benefit medical providers, it is likely that more private companies would offer related services. Continuous improvement, informed by feedback from both medical staff and patients, is also crucial.
The symptom summary reports generated by SIMTOMI are anonymized and do not allow for the identification of individual patients. Currently, the transmission of symptom information is not conducted online, but rather through face-to-face visits to hospitals or pharmacies, where the information is shared via QR codes. Additionally, the list of related diseases derived from symptoms is explicitly stated as not being a diagnosis. Depending on the case, doctors and pharmacists must be aware of the possibility of misinterpretation of this information and treat patients accordingly. Therefore, this service cannot replace actual medical consultations conducted by doctors and is only intended to be used as an aid in the treatment process.
Budget: Not disclosed
Connect with 500,000+ public servants solving your hardest challenges.
Connect with 500,000+ public servants solving your hardest challenges.
Help public servants worldwide learn from your work, what worked, what flopped and what you'd do differently
Share your project
Log in or sign up to continue the conversation