PharmaBulaBot: A Chatbot to Answer Drug Questions Based on Information from Pharmaceutical Package Inserts.

Context: The medicine inserts or pharmaceutical inserts are documents accompanying the medicines in their packaging with information and guidance to the patient. Problem: Reading and understanding the text of package inserts is not easy for most medication users. In addition, the amount of information in the medicine inserts and the font size could be more attractive to the user who wants accurate and efficient information. Solution: This work aims to present PharmaBulaBot, a chatbot capable of extracting knowledge from package inserts using Natural Language Processing techniques. The user, through questions, has access to the information on the package inserts interacting with an intelligent interface that intends to respond with a quality close to that of a human being in a quick, interactive, and efficient way. SI theory: This work was developed under the aegis of Soft Systems Theory. Method: This research has a prescriptive character, and its evaluation was carried out through the execution of experiments. The analysis of the results was carried out with a quantitative and qualitative approach. Summary of Results: From a total of 3,147 questions generated by human beings, PharmaBulaBot was able to answer 2,959 questions (94.03% of total) using sentence similarity. A local grammar was created and used in the experiments, improving the quality of responses to patients’ questions. Contributions and Impact in the IS area: The main contribution of this work was the creation of a chatbot capable of presenting answers to questions from patients with a quality close to that of physicians’ answers. This chatbot can integrate and interact with Health Information Systems in public pharmacy units.