Developing chatbot simulations of patient interviews for history taking education

Abstract

Aim or Purpose: To develop a prototype chatbot which simulates a patient to allow dental students to practice history taking skills as part of the patient interview for treatment planning. Materials and Methods: An existing patient case used in the first year of the Doctor of Dental Medicine program at the University of Sydney was adapted as the chatbot patient on the Google Dialogflow platform. Chatbot design targeted a specific learning outcome – to familiarize students with the Sydney Dental School patient interview protocol via interacting with the chatbot patient, Pia Patel. The prototype was tested by a range of academics, both clinicians (general dental practitioners) and non-clinical academics, from which their documented feedback informed further development throughout a period of six months. Results: The chatbot prototype of Pia Patel was successfully created, tested by academics and updated over a period of six months. A significant issue faced was programming the chatbot to respond consistently to the multitude of ways a single question can be asked. Overall feedback was generally positive, in terms of platform usability and authenticity of chatbot interactions. Conclusion/Clinical Significance: The use of chatbot patients in dental schools is an opportunity to improve foundational clinical skills through repeated exposure. The next steps include student trials of the chatbot to assess qualitative (feedback) and quantitative (student performance) data, the addition of more cases and the use of alternative platforms such as ChatGPT. Our results indicate the potential of artificial intelligence, particularly the usage of chatbots to improve student learning outcomes and subsequently patient care.