University of the Ryukyus Hospital Boosts Clinical Research Efficiency by 2.33x with 'Ubie Generative AI,' Slashes Initial Admission Record Creation for Emergency Transfers from 2 Hours to 5 Minutes

Ubie, Inc. (Headquarters: Chuo-ku, Tokyo; Co-CEOs: Yoshinori Abe, Kota Kubo; hereafter 'Ubie'), with its mission to 'guide people to appropriate medical care with technology,' has implemented 'Ubie Generative AI' at the National University Corporation University of the Ryukyus Hospital (Location: Ginowan City, Okinawa; Hospital Director: Mikio Suzuki; hereafter 'University of the Ryukyus Hospital'). The processing efficiency of registry registration tasks by CRCs (Clinical Research Coordinators) has improved 2.33-fold, from 1.85 cases per hour to 4.31. Furthermore, the creation of initial admission records based on faxes during emergency transfers, which typically took about 2 hours, has been shortened to as little as 5 minutes, advancing operational transformation across multiple departments from clinical research support to daily operations.

You can view more details about this case study here: https://intro.dr-ubie.com/vWIOd1Xr2/hYojLtq-

Background and Objectives of the Initiative Two years after the 'Physician Work Style Reform' was enacted in April 2024, while task-shifting from physicians to other professions is progressing in the medical field, improving the overall working environment for all healthcare professionals remains an urgent issue. Furthermore, as the 2026 medical fee revision further strengthens the evaluation of 'Medical DX and Data Utilization,' fundamental operational efficiency improvements are required to build a sustainable healthcare delivery system.

The University of the Ryukyus Hospital is the only Specified Functional Hospital in Okinawa Prefecture and is the southernmost and westernmost university hospital in Japan. It will relocate to the Nishi-Futenma campus in Ginowan City in January 2025 and serves as the core of Okinawa's health and medical hub under the philosophy of 'providing high-quality medical care from the patient's perspective and nurturing excellent medical professionals who contribute to the region and society.' The burden of document creation for medical staff, including doctors and nurses, is a common issue in many medical institutions. At this hospital as well, securing sufficient time for creating medical records such as discharge summaries was difficult, and a mechanism was needed to more reliably ensure the quality of medical care. Against this background, the hospital decided to introduce 'Ubie Generative AI,' which offers a high degree of freedom in customizing prompts and can be flexibly used by everyone from beginners to experts.

Specific Use Cases and Results

● CRC Registry Registration Task: Processing efficiency improved by 2.33x CRCs, who support clinical research, are responsible for collecting dozens of data items from electronic health records and registering them in a database. Previously, this required manually inputting data while reviewing vast medical records, limiting processing to 1.85 cases per hour. After introducing 'Ubie Generative AI,' the workflow was changed to first organize the necessary information from the records and then confirm and input it. The number of cases processed per hour increased from 1.85 to 4.31 (a 2.33-fold improvement), reducing the workload on CRCs.

● Initial Admission Record Creation for Emergency Transfers: Shortened from approx. 2 hours to as little as 5 minutes It is also being used for creating records from paper-based medical information. In one case of an emergency transfer, an initial admission record was created from a 23-page fax, completing a task that usually took about 2 hours in as little as 5 minutes.

● Transcription of Informed Consent (IC) Records With appropriate privacy protection and after fully explaining to and obtaining consent from the patient, the AI transcribes the content of the doctor's informed consent explanation from audio for use in records. By explaining to patients that the purpose of the transcription is accurate record-keeping by AI, the hospital aims to establish the new technology while gaining their understanding. Since this significantly eliminates the effort of transcribing spoken explanations into text, doctors who are already using it have commented, 'It makes it remarkably easier, so I definitely want to use AI for IC records,' and its use is expanding to multiple departments and professions.

The University of the Ryukyus Hospital aims to expand the use of 'Ubie Generative AI' to all clinical departments in the future. In addition to active lectures and case sharing among staff, the hospital will work to automate tasks and improve accuracy by further deepening its integration with the electronic health record system. This will help secure more time for staff to interact with patients while actively working to improve the quality of medical care.

(Image: QA session on advanced utilization methods held regularly)

Comment from Mr. Tetsuo Hirata, Vice Hospital Director / Director of Medical Information Management Center / Director of Patient Support Center / Head of Medical Information Department / Head of International Medical Support Office / Head of Cancer Registry Office, University of the Ryukyus Hospital

With doctors and nurses spending an enormous amount of time on documentation, there were cases where discharge summaries were inadequately written due to lack of time, and I felt there were issues in ensuring medical quality and safety. I expect that by using generative AI to create high-quality documents in a short amount of time, it will not only reduce the workload but also greatly contribute to the improvement of medical quality and safety.

Deputy Head of Medical Information Department / Medical Information, University of the Ryukyus Hospital