AI Chatbot Development for Answering High-Alert Drug Injection Compatibility Questions in Drug Information Services

Abstract

The objective of this research to develop an AI chatbot and to study its effectiveness in answering questions related to the compatibility of high-alert drug injection in drug information services at Queen SavangVadhana Memorial Hospital. It is a design and development-based research, dividing into 2 phases. Phase 1 is the development of an AI chatbot based on Adapted Waterfall system which includes collecting and studying data on questions and answers. Its purpose is to determine the question format for creating of the AI chatbot. Phase 2 is the process of evaluating the effectiveness of the AI chatbot in three parts: 1) the ability to answer questions, 2) the efficacy of AI, and 3) other recommendations. The evaluation process involves seven pharmacists who participated in in-depth interviews after using the developed AI chatbot for one month. The AI chatbot was able to correctly answer all high-alert drug injection-related questions, with a 100% accuracy rate. Evaluation results on the performance indicated strongly agreement on the immediate response of the AI chatbot (Mean score 6.93±0.27 of 7). High satisfaction level on continuously use of AI chatbot and strongly recommend to others was found (Mean score 6.81±0.06 of 7). Other results also suggested that the participants paid attention on the importance of accurate and precise information on the AI chatbot response. Reliable evidence, thus, should be included in the response. The developed AI chatbot is effective for answering questions in Drug information services, as it provides a correct and immediate response. High satisfaction level on continuous use of AI chatbot was found in all AI users.