ExaWizards and Kyoto University Jointly Developed AI to Evaluate the Text-based Reports Data on Safety of Pharmaceutical Products
—Helping to increase efficiency in the reports evaluation and contributing to promote safer use of pharmaceutical products
ExaWizards Inc. (headquartered in Minato-ku, Tokyo; Ko Ishiyama, Representative Director & President; hereafter, “ExaWizards”) has developed AI to evaluate the text-based reports data on the safety of pharmaceutical products jointly with the National University Corporation Kyoto University (hereafter, “Kyoto University”) based on the result of evaluation performed by Pharmaceuticals and Medical Devices Agency (hereafter, “PMDA”) on the incidents included in the Project to Collect and Analyze Medical “Near-Miss” Incidents in Pharmacies (hereafter, the “Near-miss Incidents in Pharmacies”) provided by the Japan Council for Quality Health Care (hereafter, “JQ”). The development has been implemented as the “Development of AI for evaluation of safety management measures against the Near-miss Incidents in Pharmacies” after being selected for the MHLW grants for clinical research on the ICT infrastructure construction/ artificial intelligence implementation projects in FY2020 (*1).
☑︎ Extracting the incidents that require measures in Scale 1 or 2 at Recall 96%, which is the index for “minimizing oversights.” Trying to further increase the accuracy, aiming for putting it to practical use in the next few years
About 145,000 Near Miss incidents on dispensing and prescription question occurred in pharmacies were reported in FY2019. PMDA extracted incidents from those involving “mix-up of dose/dosage form,” “drug mix-up,” “others,” and “prescription question,” and evaluated the necessity of the safety management measures against the physical factors of the pharmaceutical products for such incidents (*2). The safety management measures against the physical factors of the pharmaceutical products are implemented based on the evaluation result and according to whether or not the measures by the manufacturers of the pharmaceutical products are needed.
ExaWizards and Kyoto University jointly developed an AI system that learned the results of the past evaluations by PMDA to evaluate whether safety measures are required. We then had the AI system evaluate the necessity of safety measures on the five-level scale（*3）provided by PMDA according to the details of the incidents and formulation properties, and verified that AI’s classification of the incidents (which fall in Scale 1 and Scale 2 in (*3) for which PMDA determined that the measures were necessary has achieved 96% of Recall, which is the index for “minimizing oversights”. By reflecting the information including the drug efficacy and regulatory classification on the drug database of package inserts, we also used the method to extract and evaluate the formulation information that may involve severe health problems contained in the incidents in this verification.
By making the process to extract an incident that requires measures more efficient with AI, it is expected to eventually improve efficiency of the safety management measures operations in PMDA. We will further improve the system, aiming for putting it to practical use in the next few years.
We conducted the verification with the data in the previous form, and will work on the verification using the data in the latest form, adopted in March 2020, this year with the aim to achieve an accuracy level equivalent to or higher than this time. In the future, we will further advance development of the AI system that can be used in the actual operations to enable more efficient, automated operations so as to evaluate Near-miss Incidents in Pharmacies. More efficient and automated operations are expected to contribute not only to the streamlining of PMDA operations, but also to the promotion of safer use of pharmaceutical products through the secure and speedy implementation of safety measures.
Based on this development, ExaWizards will continuously offer the AI solutions that can help in solving various problems associated with medical care by further leveraging the natural language processing technology in the medical and pharmaceutical fields.
(*1) The progress in FY2020 was also reported at the 7th Annual Meeting of Japan Society of Clinical Safety (on-demand streaming is available from March 29 to June 7, 2021)
(*2) Source: The 23rd Report (January–March, 2020) from Project to Collect and Analyze Medical “Near-Miss” Incidents from Pharmacies. According to the 2nd conference for reviewing safe use of pharmaceutical products in FY2020 by PMDA, a total of 2,416 incidents were extracted from the 97,707 incidents covered in Project to Collect and Analyze Medical “Near-Miss” Incidents from Pharmacies that were reported between May 1, 2019 and December 31, 2019 to be evaluated by PMDA, and 6.7 % of them were identified that they might involve severe health problems from the aspect of the directions for use and names/packages of such pharmaceutical products (falling in Scale 1 or Scale in (*3).
Scale 1: Incidents for which measures by the manufacturers for safe use of pharmaceutical products were considered to be necessary or possible
Scale 2: Incidents for which certain measures have already been taken or under consideration by the manufacturers
Scale 3: Incidents for which physical measures by the manufacturers were considered to be difficult (human errors, human factors)
Scale 4: Incidents for which physical measures by the manufacturers were considered to be difficult (side effects, a lack of information)
Scale 5: Others (errors in transcription of insurer numbers from prescriptions, miscalculation of dispensing fees)
[About ExaWizards Inc.]
With the mission of “Solving social issues through Artificial Intelligence for future generations,” we are developing and commercializing AI products in various fields such as nursing care, medical care, HR, robotics, finance, and cameras, in order to solve industry and society-wide issues identified from individual company issues, while working on solving issues in each department and company-wide use of AI. Our members include AI engineers, software and hardware engineers, strategy consultants, UI/UX designers, domain experts in nursing care and other fields, researchers, policy experts, and other cross-disciplinary personnel. In Japan’s super-aging society, we are developing our business with a thorough understanding of the needs and issues in each field.
Corporate site: https://exawizards.com/
<Inquiries about this release>
Public Relations Department, ExaWizards Inc.