BACKGROUND
Following the separation of medical treatment and drug dispensing in Taiwan, patients are now required to bring their prescriptions to the pharmacy and request the pharmacist to dispense the medication. The prescription serves as the basis for the pharmacist during the medication dispensing process. Currently, prescriptions are presented in paper format. Patients present a prescription issued by a medical unit and visit the pharmacy to request the pharmacist to dispense it. Upon receiving the paper prescription, the pharmacist must manually enter the drug information into the computer. However, the excessive and complex information in the prescription leads to time wastage for the pharmacist and often results in information errors.
OBJECTIVE
This study aims to utilize AI optical recognition and AI neural network machine learning technology to accurately identify the details of prescription paper and drug information. This will enable the establishment of an optical recognition artificial intelligence system and pharmaceutical care database. By utilizing this system, pharmacists will be able to record accurate information in real-time while performing dispensing services, ensuring precise patient medication information and allowing for more accurate pharmaceutical care.
METHODS
In this study, we collaborated with the Pharmacists Association, which collected prescription pads from major medical institutions in Taiwan. The collected prescription pads were first de-identified. Then, the information and graphic orientation of the prescription pads were manually marked. Optical character recognition (OCR) and machine learning using neural networks were then performed to build a database for artificial intelligence recognition of prescription papers.
RESULTS
This study successfully establishes an artificial intelligence-based prescription identification system that allows users to directly upload images of paper prescriptions. The system utilizes artificial intelligence algorithms and accesses a specifically designed database for identification purposes to accurately identify prescription information. The identified information is then presented based on the relevant fields in the system interface.
CONCLUSIONS
The field of pharmaceutical care has witnessed advancements in artificial intelligence technology. Artificial intelligence recognition technology is being utilized to enhance pharmacists' drug judgment and facilitate collaboration and implementation of home care services. Furthermore, a nationally standardized medical integration database is being established to support the integration of medical resources in Taiwan and serve as the foundation for cross-border service cooperation.