＊日本語の記事はこちら
hosted by Cyber Civilization Research Center (CCRC), Keio University
co-hosted by the Medical Inclusion working group, CCRC
Life Intelligence System Consortium, SFC Research Institute
Value Society Platform Laboratory, SFC Research Institute
WIDE Project
Event Information
Date: Monday, May 22, 2023, 12:30-17:50 (Event reception opens at 12:00)
Place: G-lab, 6F, East Building, Mita Campus, Keio University
Click here to view the program
↓Seminar recording video first half: from the organizer’s greeting to before the panel discussion
Opening
Greeting from Keio University
Masayuki Amagai (Vice President of Keio University)
The COVID-19 pandemic has recently been reclassified as a class five disease alongside common infectious diseases such as seasonal influenza. As we approach the conclusion of the Corona era, it becomes evident that digitization and data sharing have emerged as crucial pillars of transformative change at an unprecedented pace. Notably, the rapid development of Chat GPT within a short timeframe serves as a testament to the advancements in this realm. While the advent of the Internet brought about substantial transformations in the 1990s, Chat GPT has surpassed it in terms of its rapid proliferation, leveraging the foundation laid by the Internet.
The medical field is also undergoing significant shifts. Historically, medicine operated within a closed system, where patients sought diagnosis and treatment at hospitals, departing once their symptoms improved. The responsibilities of hospitals were confined mainly within their physical walls. However, propelled by the digital revolution, hospitals are extending their reach beyond these boundaries, seeking to engage with individuals in their daily lives and real-world conditions. Standardization, establishing rules, and developing robust systems are central to this transformative process, all of which assume paramount importance.
Organizer’s Remarks
Jun Murai (Professor and Co-Director, Cyber Civilization Research Center, Keio University)
Although the Internet predated the coronavirus, it is undeniable that the period of digital transformation (DX) amid the pandemic has provided an unprecedented opportunity for all citizens to embrace the advantages of a digital society. Notably, remote work has become commonplace, digital services are conveniently accessed through smartphones, and wearable devices empower us to monitor our daily health conditions.
In the past, specialized knowledge that was once confined to experts is now being widely disseminated to the general public through the Internet. There will be a growing convergence between universally popularized healthcare and specialized medical care, fostering increased collaboration. Creating avenues for diverse healthcare and medical professionals to collaborate is imperative, and this has become a vital mission for universities. The Cyber Civilization Research Center, a dedicated institution studying the societal impacts of digital technology, aims to facilitate networking opportunities among experts in advanced medical care, healthcare services, and technology, supporting this collaborative endeavor.
Introduction
Hitomi Sano (Project Researcher, Graduate School of Media and Governance, Keio University)
Conventional medical care operates within a centralized structure where doctors and medical institutions have served as focal points for various medical resources. However, the advent of the Internet: the fundamental technology for cyber civilization and the democratization of information, has facilitated the free exchange of digital information, leading to a new user model centered around individuals in medicine and health. Moreover, as an interconnected network, the Internet integrates diverse networks, bridging different domains, including medical and healthcare, and establishing a novel social infrastructure. The transformation of this user model and the evolution towards an expanded public health domain will determine the growth trajectory in medical DX.
While Internet technology has been actively integrated into the competitive landscape of industries, fostering globalization and the development of a digital society, there is a growing need to consider ethics and the public dimension that have yet to be fully addressed within this competitive backdrop. This seminar seeks to position medical and health DX as a flagship driver of overall global DX. It calls for new connections and collaborations to foster co-creation to address these unprecedented challenges.
Opening Remarks
Sumio Matsumoto (Director Emeritus, National Hospital Organization, Tokyo Medical Center)
After extensive deliberations, the standardization of electronic medical records (EMR) would be focused on the circulation of information based on the “three documents, six pieces records” framework. And there is an ongoing debate regarding sharing this information, which originates from the medical field, through EMR information exchange services without obtaining the explicit consent of the individuals involved. However, the potential implications of such a practice are a subject of concern, as it may attract public criticism.
Furthermore, when obtaining individual consent for utilizing medical information in conjunction with the My Number Card, it is worth considering whether the current categorization, limited to surgical procedures and other medical data, is adequate. Handling sensitive information such as genomic data and mental illnesses poses significant ethical considerations that must be addressed. Engaging in comprehensive discussions surrounding managing diverse medical information encompassing Personal Healthcare Records (PHR) is imperative.
Moreover, physicians entrusted with the responsibility of handling patients’ medical information encounter a variety of use cases, including emergencies, instances with ample time for evaluation, and scenarios devoid of complicating factors.
Considering these multifaceted perspectives, deciding the appropriate governance responsible for integrating the diverse information streams currently managed by various government administrative bodies becomes crucial.
Today, I hope thoughtful discussions are aimed at extending healthy life expectancy, enhancing the accessibility and convenience of medical care and nursing services, and mitigating the burdens healthcare professionals face.
Keynote Speech
“Data utilization in the health and medical fields”
Yoshihide Ezaki (Director， Institute for Social Policy Issues)
What lessons can we derive from the COVID-19 pandemic? One significant lesson is the importance of “appropriate fear.” While it is useless not to be completely fearless, fear without careful consideration can also be problematic. Health should be viewed not as the “ultimate goal” but as a “condition” and a “result” of accomplishing particular objectives. Unthinking fear and refusing to leave our homes can make us vulnerable, weakening our immune systems. The decline in immune function has been closely associated with increased mortality among older adults during the COVID-19 pandemic. While basic measures like handwashing, disinfection, masks, and vaccination are crucial in combating infectious diseases, it is equally important to take actions that strengthen our immune system.
When contemplating the future of healthcare, it is vital to consider the true significance of the era of a 100-year life, which entails understanding the authentic nature of human beings. Humans are said to have a natural lifespan of 120 years. In the present era, many individuals can live out this 120-year lifespan, and medical and healthcare services exist to support this achievement.
At what age does an individual become a senior citizen? Who decided that growing older equates to becoming weaker? Suppose older individuals can lead fulfilling lives and contribute meaningfully to society. In that case, their immune systems will strengthen, enabling them to live with vitality, thereby increasing their healthy life expectancy—the default assumptions about a person’s lifetime change in an aging society. However, AI trained in an environment that fails to capture these changes may continue to produce inaccurate analysis results.
The nature of diseases is also evolving. While the existing healthcare system has primarily focused on exogenous diseases, such as infectious diseases, the future will prioritize endogenous conditions that necessitate personalized treatment approaches, such as ailments related to aging and lifestyle. Expanding the framework of patient self-health management through DX is well-suited for addressing these diseases.
Japan has a population exceeding 100 million and a vast repository of high-quality data. Although technical issues and privacy concerns exist, Japan’s healthcare system can lead the world by establishing clear treatment goals and leveraging accumulated data.
“Data sharing for improving medicine: A lesson from COVID-19”
Makoto Suematsu (Director, Central Institute for Experimental Animals、Professor Emeritus, Keio University)
For many years, I have conducted extensive research on the significance of data collaboration, focusing on wide-area collaboration and distributed integration.
A further internal division exists within the medical field, often leading to specific researchers or specialists collecting data on particular diseases at specific hospitals. Various challenges arise when sharing data due to stakeholders’ differing political and power structures. There is a tendency to confine data within closed research communities, making it impossible to address medical issues from a global perspective. Consequently, it became necessary to implement conditions such as “No share, No budget” when distributing research funds in specific fields.
To illustrate the structure of the medical data problem, we can identify the following scenarios:
・Alternate Attendance Type: Resembling a tax with little return during the Shogun Era. For instance, data from cancer genome research is collected and analyzed in one location. In contrast, the data acquired by cancer genome centers from core hospitals nationwide are never returned, yielding no discernible patient benefits.
・Black Hole Type: Absorbing everything with no opportunity for utilization, for example, National Database (NDB).
・Universal Availability in a Decentralized Network: A management approach where everyone mutually handles data will be necessary for the future.
One example of a successful medical database is the Initiative on Rare and Undiagnosed Diseases (IRUD). It encompasses information on phenotypes, genomic data, and contact details of primary physicians treating children with rare and intractable diseases nationwide. Although there are restrictions on commercial use, progress in international collaboration has expanded the scope of case matching, resulting in rapid advancements in diagnosing previously undiagnosed conditions. Achieving global data sharing necessitates software development, such as universal plugs that can connect to the database from each country, as well as the establishment of trustworthy relationships between individuals. An illustrative case is the reliance of the Republic of Lithuania on Japan for matching undiagnosed diseases, as the former encountered difficulties exporting domestic genome data to neighboring countries due to diplomatic tensions. This example highlights the interconnectedness between geopolitics and medicine and successfully demonstrates global medical data science.
Additionally, the Global Initiative on Sharing All Influenza Data (GISAID), initially established as a database for bird flu, has proven invaluable for the new coronavirus and has become a worldwide common database. In infectious diseases, the information contributed to a database called ProMED-mail, maintained primarily by volunteers on a limited budget, marked the beginning of understanding the infection status.
Finally, scientific autonomy must be independent of politics from various countries and specific communities to facilitate the real-time sharing of essential medical information, such as infectious disease data, through an international database.
“Implementation of Generative AI in Healthcare: Challenges and Potential”
Kazuhiro Sakurada (Professor, Department of Extended Intelligence for Medicine, The Ishii-Ishibashi Laboratory, Keio University School of Medicine)
Until now, numerous industries have undergone automation, but the medical and health field presents a unique challenge due to the inherent difficulty in expressing its complexities through algorithms. Traditional science has tended to identify causal mechanisms and explain phenomena by viewing the world as a machine. However, understanding non-equilibrium and nonlinear open systems, which cannot be easily explained through causal mechanisms, has posed a significant challenge in scientific endeavors. Nonetheless, generative AI advancements are bridging this gap and driving further progress in science and academia.
The application of AI in medicine involves integrating reliable medical knowledge, such as standard treatments. There is a growing demand for general-purpose medical AI systems that can incorporate various existing AI diagnoses and utilize them in communication-based settings. Additionally, AI is used in basic medical research to facilitate discoveries. One practical approach involves leveraging AI machine learning techniques to break down complex problems into their fundamental components and identify causal relationships. However, it is crucial to ensure the high reproducibility of the data used for analysis; otherwise, the research may lack practicality.
While the widespread adoption of foundation models and generative AI, such as ChatGPT, is remarkable, the relationship between AI models and data is evolving. In the past, possessing vast amounts of data granted an oligopolistic advantage. However, once a high-quality model is generated, everyone desires to analyze their data using it. The more the model is utilized, the more accurate it becomes, emphasizing the increasing need for widespread adoption. Since commercial models developed by companies often do not disclose their parameters, it is essential to have openly accessible models within the academic community. Therefore, shifting our perspective from solely on data integration to establishing frameworks for creating new open models is pressing.
Speech from Universities
“Expectations for the digital transformation in surgery using Web3-based telesurgical network in the era of robotic surgery”
Koichi Suda (Professor, School of Medicine Department of Gastroenterological Surgery, Fujita Health University)
(Reference information) Please refer to the previous lectures by Dr. Suda through the link provided below.
https://www.ccrc.keio.ac.jp/firsthalf_report_2ndmedicalseminar_en/
At Fujita Health University, we have established a collaborative environment between industry and academia to address surgical challenges through initiatives such as joint research and development of the Japanese-made surgical robot ‘Hinotori.’
Last year, telesurgery guidelines were published, guiding definitions of telesurgery, including telementoring, telesurgical support, and full telesurgery. From now on, it is crucial to establish a remote surgical support environment that enables a supervising doctor to assist a team of local doctors directly through remote control, allowing for joint surgeries by seamlessly transferring the operation authority between local and remote doctors.
The first challenge we must tackle is network latency and fluctuations. In an experiment involving suture ligation using Hinotori, the surgeon’s movements were affected, and operation time was extended when the latency ranged between 100 milliseconds and 125 milliseconds. Surgeons have expressed discomfort if the delay exceeds 50 milliseconds. In a recent demonstration experiment conducted over a distance of 300 km between Tokyo and Nagoya, a delay of 31 milliseconds was experienced during a complete remote gastrectomy performed on a live organism (pig) with human-like anatomy. The operation was completed without any issues. We aim to develop the global Hinotori telesurgery platform by establishing communication technology that maintains a delay of 50 milliseconds, even over distances of 10,000 km.
Effective communication between local and remote locations is essential to improve telesurgery safety and facilitate remote operations. We must explore the optimal operator’s cockpit environment for remotely controlling the robot and investigate efficient ways to share various patient information, such as panoramic views, voice, and biological data from the operating room.
Leading companies in the robotic surgery industry have been actively participating and accumulating digital data related to surgery, known as ‘Surgical Intelligence.’ This surgical intelligence primarily consists of log information generated by surgical assistance robots and holds significant potential for commercialization through AI analysis, making it valuable intellectual property.
We aim to promote and disseminate Japan’s unique medical DX by addressing key issues, including:
・Enhancing the environment for surgical education using telesurgery platforms and surgical intelligence.
・Ensuring compatibility with new economic models, such as Web3 that support the medical DX environment.
・Establishing an advanced medical information network that collects diverse medical information centered around Surgical Intelligence and utilizing AI to analyze data within the medical network.
“Expectations for the digital transformation in medicine other than surgery –From Image Recognition to Natural Language Processing-”
Masahiro Jinzaki (Professor, School of Medicine Department of Radiology and Deputy Director, Keio University Hospital)
(Reference information) Please refer to the previous lectures by Dr. Jinzaki through the link provided below.
https://www.ccrc.keio.ac.jp/firsthalf_report_2ndmedicalseminar_en/
DX is thriving in the industrial sector, and AI research is also advancing in the medical field. However, the effective utilization of AI implementations in practical settings still needs to be improved. AI development and implementation are distinct challenges.
Just as the internet has become a widely adopted technology among the general public, AI should also become a popular technology accessible to a broader audience through DX. To achieve this, it is crucial to establish a safe and inclusive process that facilitates the utilization of AI by individuals with varying skill levels. The objective of Keio University Hospital’s AI hospital is not only to implement AI in specialized medical care but also to endorse the low-skilled utilization of IT and AI technologies, such as digital signage and guide robots, which many hospitals can quickly adapt.
While deep learning excels in image recognition tasks, particularly in streamlining specific operations or enhancing image quality, its efficacy in actual multitasking diagnoses, such as dealing with multiple organs and detecting unexpected lesions, is limited. However, with the emergence of technologies like ChatGPT, which harnesses natural language processing, there is a growing trend towards low-skilled AI that extends beyond image recognition and is expected to contribute significantly to developing comprehensive AI hospitals.
Our focus will be accumulating consistent data throughout the patient journey, from pre-illness to hospitalization to post-illness, and leveraging AI to enhance the accuracy of patient condition prediction. The key direction for future medical DX is not solely the implementation of existing AI in hospitals but rather envisioning the next-generation hospital within the digital society and conducting research and development that aligns with this concept in collaboration with venture companies and other stakeholders.
Speech from Industries
“Shaping the Future of Nursing Care with Data ～” Nursing care Real Data Platform” promoted by SOMPO Care～.”
Takahiro Iwamoto (Sompo Care Inc. Exective Director Chief Digital Officer General Manager,egaku Business)
The demand for nursing care is increasing daily, while Japan’s working-age population is declining, resulting in a shortage of nursing care workers. This shortage and high recruitment costs are challenging for the nursing care industry.
The Japanese government has introduced the ‘LIFE’ (Long-term care Information system For Evidence) initiative to promote the utilization of nursing data for scientific care. However, assessing how effectively suppliers in the nursing care industry can implement this digital transformation (DX) initiative is essential.
At Sompo Care, we are dedicated to enhancing operational efficiency and sustainability in the nursing care industry. We have developed ‘egaku,’ a comprehensive system that integrates and analyzes data from all nursing care services. By leveraging DX technologies such as sleep sensors and recording systems, ‘egaku’ has consolidated approximately 500 pieces of fragmented data, enabling a visual representation of the nursing care service. ‘egaku’ empowers us to provide an optimal nursing care process tailored to each customer, replacing the reliance on care workers’ intuition and experience with standardized awareness and know-how.
Furthermore, we have developed a ‘predictive care model’ that assesses the user’s health condition based on the level of care required and converts state transitions into big data. This model suggests potential changes in the user’s need and offers proactive countermeasures through alerts. By leveraging ‘egaku,’ we can deliver optimal nursing care services at an early stage, significantly improving the users’ quality of life (QOL).
As we expand the implementation of ‘egaku’ to numerous nursing care providers, we aim to drive transformation across the entire nursing care industry. We are also committed to promoting data collaboration with various vendors involved in nursing care, implementing Data Free Flow with Trust (DFFT) principles, and integrating with the LIFE system.
“Self-monitoring, data utilization, and issues of dissemination from the perspective of PHR”
Hiroyuki Kazuma (Advisory Specialist on Healthcare Alliance, OMRON Healthcare Corporation)
We are committed to harnessing data in preventive medicine to lead the global effort towards better health and achieve the ambitious objective of ‘zero events,’ eliminating sudden seizures and other medical emergencies.
Within the cardiovascular field, our initiatives primarily revolve around delivering innovative devices, telemedicine services, and developing AI systems that leverage measurement data collected at home to support diagnosis and treatment. Our portfolio of innovative devices includes wearable sphygmomanometers aimed at miniaturization, brachial sphygmomanometers with electrocardiographs that incorporate indicators beyond blood pressure, and portable electrocardiographs. Furthermore, our PHR app, OMRON Connect, which seamlessly integrates data from healthcare equipment, has already garnered over 1.4 million downloads.
The COVID-19 pandemic has heightened health awareness and underscored the significance of monitoring vital signs at home. However, stringent advertising regulations imposed by the Pharmaceuticals and Medical Devices Law present challenges in promoting medical devices. Consequently, non-medical devices, which may offer lower accuracy, have gained more popularity and market presence.
There are several crucial considerations surrounding Personal Health Records (PHR). We must diligently adhere to the Personal Information Protection Law, Pharmaceuticals and Medical Devices Law, Medical Practitioners Law, and other pertinent regulations.
To foster widespread adoption of PHR, relaxing advertising regulations for home medical devices is imperative, thereby enhancing the accuracy and reliability of collected data. Moreover, if we aim to collaborate with public medical care services, discussing medical fees is essential while considering the burden on medical providers.
In this manner, we must engage in comprehensive discussions encompassing various aspects of PHR, ranging from the accuracy of lifelog data to collaboration opportunities with medical institutions, ensuring a well-rounded understanding of the subject matter.
“Future of Health Care with PHR x AI”
Takayuki Hamada (Executive Officer, JMDC inc.)
We are actively engaged in the health big data business, providing all stakeholders with an extensive range of healthcare-related data. Our primary data sources encompass receipts and medical examination data obtained from health insurance associations, and we proudly possess the largest epidemiological database in Japan, comprising an impressive collection of over 809 million receipts. The data curated by JMDC grants us invaluable insights into patient conditions spanning numerous hospitals and pharmacies.
One of our flagship offerings, ‘Pep Up,’ is one of Japan’s largest Personal Health Record (PHR) platform, boasting an impressive user base of 1.2 million individuals. This platform provides essential PHR functions to monitor daily health conditions, empowering users to track their well-being. It facilitates comprehensive analysis by integrating receipt and medical examination data. This analytical capacity is particularly valuable in the aftermath of manufacturing and marketing new coronavirus vaccines.
By amalgamating longitudinal data, including examination results and disease outbreaks, we possess the capability to develop diverse predictive AI models that establish meaningful links between daily behavioral factors and potential consequences. Leveraging our extensive data resources, we construct an optimal intervention prediction AI model, enabling the delivery of personalized recommendations. This proactive approach ultimately strives to curtail medical costs throughout Japan. We aspire to become a national PHR service that seamlessly connects with 20 million users, revolutionizing healthcare outcomes significantly.
“Solving problems through data utilization and DX.”
Hiroki Miyamoto (Managing Executive Officer / Group Director, Healthcare Business Group, MTI Ltd.)
We are developing a comprehensive healthcare business centered on “supporting health management from infancy to old age: from minus one year old to 100 years old.”
Since its inception in 2000, our women’s health information service, ‘Luna Luna,’ has amassed an impressive user base, exceeding 19 million downloads. Through this platform, we offer home-based fertility support, leveraging the power of the Luna Luna algorithm. This algorithm, crafted using the extensive data accumulated within the app, has demonstrated a remarkable 136% improvement in pregnancy probability compared to conventional methods (based on in-house research). In 2021 alone, the Lunaluna app recorded approximately 280,000 pregnancies, constituting one-third of Japan’s annual births. Moreover, it is one of the use cases for Personal Health Records (PHR), fostering collaboration with medical institutions.
Another unique offering within our portfolio is the ‘Boshimo’ app, an electronic maternal and child handbook, which over 540 local governments have successfully adopted. This invaluable tool supports parenting for approximately one-third of newborn children, receiving high praise for its user-friendly management of children’s vaccination schedules, which can often prove burdensome for parents. Our overarching objective is to alleviate parental anxiety and child-rearing burdens by leveraging digital transformation. By collaborating closely with residents, governmental entities, and medical institutions, we aim to facilitate administrative applications and streamline procedures of child-rearing complexities.
Click here for the second half of the seminar.
(written by Hitomi Sano , photo by Shinichi Yamazaki)