This news release is information being released to the press.

Fujifilm Corporation (Headquarters: Minato-ku, Tokyo; President and CEO: Teiichi Goto; hereinafter referred to as Fujifilm) and Juntendo University School of Medicine affiliated Juntendo Hospital (located in Bunkyo-ku, Tokyo; Director: Ryohei Kowatsuru; hereinafter referred to as Juntendo Hospital) jointly developed a technology that predicts the risk of falls in outpatients using AI technology based on the diagnostic data in CITA Clinical Finder, Fujifilm's integrated medical institution support platform that can centrally manage various medical data in the hospital. With this technology, it is possible to predict the risk of falls in outpatients with high accuracy, leading to prevent falls. In terms of product structure, 10-30 billion yuan products operating income of 401/1288/60 million yuan respectively.

Patient falls occur frequently in medical facilities in Japan and can lead to serious injuries such as bone fractures and head injuries, potentially affecting the patient's life prognosis and quality of life.^※2As a result, since the number of outpatients who are subject to assessment is larger than that of inpatients, and it is difficult to understand the patient's condition during the limited hospital stay compared to inpatients, it has been strongly demanded to develop an efficient and highly accurate method for evaluating the risk of falls in outpatients.

Fujifilm and Juntendo Hospital have developed a technology that uses AI to predict the risk of falls in outpatients using the diagnostic data accumulated in Fujifilm's integrated medical institution support platform, CITA Clinical Finder. This technology generates features that are considered to be highly related to more than 500 types of fall risk, such as age and specific prescribed medication, from data aggregated in CITA Clinical Finder that is linked to various hospital systems such as electronic medical records, radiology department systems, and endoscopy department systems. The features were learned by AI to develop this predictive model. The technology predicts each patient's risk of falling based on the diagnostic data registered in CITA Clinical Finder and displays it as a percentage. Also, it is possible to present the features that contributed to the prediction as anticipated fall risk factor.
With the data of approximately 70,000 outpatients at Juntendo Hospital, we evaluated the accuracy of this technology and showed a superior result (AUROC: 0.96) compared to the previous study, which targeted inpatients (AUROC: 0.90). Through the use of this technology, healthcare professionals can evaluate the risk of falls in outpatients with a high degree of accuracy.^※3※4

Fujifilm and Juntendo Hospital will further verify the effectiveness of this technology and aim for early implementation in the future.

"At Juntendo Hospital, we have cited a decrease in the fall rate of outpatients aged 75 and over as one of the quality indicators of medical care for a long time. This is because we want to avoid patients suffering from further diseases due to falling in the hospital while they are coming for diagnosis and treatment. This time, we focused on the risk factors for falls in outpatients, which had not been researched so far, and jointly developed a fall risk prediction model with Fujifilm Corporation using the integrated medical data and machine learning. We expect that this research result will be used to predict the risk of falls in outpatients with high accuracy and that medical facilities will take measures to prevent falls before they occur," said Ryohei Kowatsuru, Director of Juntendo Hospital.

"For the improvement of medical quality and safety, we are pleased to have been able to integrate the medical expertise of Juntendo Hospital, which has been promoting many efforts, and comprehensive and large-scale medical data and our unique medical prediction AI technology at CITA Clinical Finder. We have developed a technology that predicts with high accuracy the risk of falls for outpatients. We will continue to contribute to the early detection of falls in outpatients by implementing this technology in society," said Toshiyuki Nabeta, Executive Officer, Medical Systems Development Center, Fujifilm Corporation.

• *1 Developed using machine learning, one of AI technologies.
• From the "Medical Accident Information Collection and Other Projects" of the Japan Medical Function Evaluation Organization, Public Interest Incorporated Foundation (Note 2).
• Refers to Area Under the Receiver Operating Characteristic curve (Note 3), an indicator used to evaluate prediction accuracy. The closer the value is to 1, the higher the prediction accuracy.
• Lindberg, DS, Prosperi, M, Bjarnadottir, RI, et al. (2020). Identification of important factors in an inpatient fall risk prediction model to improve the quality of care using EHR and electronic administrative data: A machine-learning approach. International Journal of Medical Informatics, 143, 104272 (Note 4).

Inquiry

• *The contents of the article are as of the time of announcement. Please note that there may be changes to the latest information (such as discontinuation of production and sales, changes to specifications and prices, changes to organization and contact information).