Methods for constructing patient medication instructions using generative artificial intelligence
Objective To establish a basic framework for patient medication instructions,study the method of construct-ing patient medication instructions based on generative artificial intelligence,generate medication instructions that are easier for patients to understand,and promote rational medication use.Method Compared the similarities and differences in patient medication instructions across five countries,We incorporate core elements mentioned in the medication instructions of four or more countries,to form a basic framework for patient medication instructions.Based on this framework and using generative artificial intelligence and prompt,We provide examples of patient medication instructions.Results The patient medication in-structions from the USA,EU,Japan,and Australia were compared.A framework for patient medication instructions was estab-lished,incorporating information such as drug name,indications,dosage and administration,precautions,adverse reactions,storage methods,drug information,and manufacturer details.Prompt included directive words,background descriptions,out-put requirements,and related materials.Using the QWEN 2.0 model,an example of patient medication instructions to pregab-alin was generated.Compared to the original medication instructions,the length of the patient medication instructions was re-duced from 15 395 words to about 1052 words,showing a significant improvement in readability.The accuracy was 77.8%,and with the oversight of experienced pharmacists,the accuracy could reach the same level.Conclusion We constructed a framework for patient medication instructions.Generative artificial intelligence can efficiently produce patient medication in-structions with improved readability,which were then reviewed and approved by pharmacists to ensure accuracy.
NETL
NSTL
万方数据
