Brain Sleep Co., Ltd. (Headquarters: Chiyoda-ku, Tokyo; Representative Director: Kosuke Michibata; hereinafter "Brain Sleep"), Nippon Telegraph and Telephone East Corporation (Headquarters: Shinjuku-ku, Tokyo; Representative Director and President: Fukuzo Inoue; hereinafter "NTT East"), and Ota Memorial Institute Sleep Center (Location: Kawasaki City, Kanagawa Prefecture; Center Director: Shintaro Chiba; hereinafter "Ota Sleep Center") have successfully developed a machine learning model to reduce the burden of interviews used in diagnosing obstructive sleep apnea*1, achieving a reduction in the number of interview questions while maintaining diagnostic accuracy. The reduction in interview questions using a machine learning model developed based on diagnostic data held by the Ota Sleep Center, which boasts one of the highest numbers of obstructive sleep apnea cases nationwide, is an unprecedented achievement. Going forward, the aim is to utilize this model to minimize the burden of interviews during diagnosis. These research findings were presented at the 46th Annual Meeting of the Japanese Society of Sleep Research (held in Fukuoka in September 2021) under the title "Examination of Obstructive Sleep Apnea Diagnosis using Questionnaires and Machine Learning Models" (presenters: Kenichi Yanagihara, Shintaro Chiba), garnering significant anticipation from the sleep research community. *1 Sleep apnea syndrome is, as the name suggests, a condition where "apnea" repeatedly occurs during "sleep," disrupting sleep. It is abbreviated as "SAS" from its English name, Sleep Apnea Syndrome. In the latest international classification of sleep disorders, the official disease name is obstructive sleep apnea (OSA). Research Background Japan is known to have the shortest sleep duration among OECD member countries, and this trend is decreasing year by year. Among these, sleep apnea syndrome, a particularly problematic sleep disorder, currently affects an estimated 400,000 to 500,000 patients receiving treatment in Japan, and this number is on the rise. Furthermore, with growing concerns about the risk of diseases and reduced labor productivity associated with sleep disorders, we believe that early detection and treatment of sleep disorders and reducing the burden on patients for diagnosis are crucial initiatives to address Japan's sleep challenges. Therefore, we conducted research utilizing machine learning, a cutting-edge technology. Research Method and Results In this study, the explanatory variables and objective variables were selected as follows: ◆Explanatory Variables*2 (Total of 117 items) 1. Questionnaire items: 111 questions - Original medical interview questionnaire: 79 questions - Japanese version of the Epworth Sleepiness Scale *3 (JESS): 8 items - Pittsburgh Sleep Quality Index *4 (PSQI): 24 items 2. Easily measurable items from overnight polysomnography *5 (PSG test): 6 items ◆Objective Variable*6 Diagnosis result of Obstructive Sleep Apnea (OSA) We created a machine learning model to determine whether the diagnosis result was "severe OSA" or not, and then confirmed the accuracy of its predictions. Using data from a total of 3,160 subjects, machine learning was performed, and the accuracy of the prediction results was evaluated. High diagnostic accuracy was confirmed across five evaluation models: accuracy, precision, recall (sensitivity), F-measure (balance between precision and recall), and AUC (learning performance). *2 Variables that influence the objective variable in machine learning. *3 Interview items for assessing daytime sleepiness. *4 A self-administered questionnaire for evaluating sleep quality. *5 A test that measures biological responses during the night, such as brain waves and eye movements. *6 The variable to be determined in machine learning. Roles of Each Company Brain Sleep: Provision of specialized knowledge regarding sleep disorders. NTT East: Implementation of machine learning and creation of machine learning models. Ota Sleep Center: Provision of data from 3,160 subjects and specialized knowledge regarding sleep disorders. Future Applications Moving forward, we are considering the deployment of services so that the algorithm developed through this research can be utilized by more people. We will strive to implement these research findings into society to facilitate early detection of sleep disorders and reduce the burden on patients. Company Overviews ■Brain Sleep Co., Ltd. (https://brain-sleep.com/) Brain Sleep was established in 2019 by Seiji Nishino, Director of the Stanford University Sleep & Circadian Neurobiology Lab, and Kosuke Michibata, with the mission of "The best sleep for the best life." The company primarily develops original products based on cutting-edge sleep research, provides sleep consulting to companies specializing in sleep, offers IT-based services, and disseminates medically-backed sleep information. Established: May 2019 Representative: Representative Director Kosuke Michibata Location: JP Tower 26F, 2-7-2 Marunouchi, Chiyoda-ku, Tokyo Seiji Nishino Professor, Department of Psychiatry, Stanford University School of Medicine; Director, Stanford University Sleep & Circadian Neurobiology Lab; Founder & Chief Research Advisor, Brain Sleep Co., Ltd.; Physician, Doctor of Medical Science; Certified Psychiatrist, Certified Sleep Specialist, Industrial Physician. ■Nippon Telegraph and Telephone East Corporation (https://www.ntt-east.co.jp/) Established: July 1999 Representative: Representative Director and President Fukuzo Inoue Headquarters: 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo Business Activities: Regional telecommunications services*8 and related services, objective achievement services, and utilization services in the East Japan region*7. *7 Hokkaido, Aomori Prefecture, Iwate Prefecture, Miyagi Prefecture, Akita Prefecture, Yamagata Prefecture, Fukushima Prefecture, Ibaraki Prefecture, Tochigi Prefecture, Gunma Prefecture, Saitama Prefecture, Chiba Prefecture, Tokyo Metropolis, Kanagawa Prefecture, Niigata Prefecture, Yamanashi Prefecture, and Nagano Prefecture.*8 Telecommunication services related to intra-prefectural calls, including telephone, dedicated lines, and integrated digital communications. ■Ota Sleep Center (https://www.ota-sleep.com/) Established in April 2004 as an institution specializing in the diagnosis and treatment of sleep disorders. At a time when the number of people suffering from sleep disorders is rapidly increasing and appropriate medical care is needed, it is one of the leading sleep disorder treatment centers that provides comprehensive diagnosis and treatment that not only addresses medical issues but also psychological and social backgrounds. Established: April 2004 Representative: Center Director Shintaro Chiba Location: 1-50 Nisshin-cho, Kawasaki-ku, Kawasaki City, Kanagawa Prefecture
Log in
