Self-diagnosis through chatbot-based symptom checkers: user experiences and design considerations
Open Access
Author:
You, Yue
Graduate Program:
Informatics
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
March 20, 2020
Committee Members:
Xinning Gui, Thesis Advisor/Co-Advisor Mary Beth Rosson, Thesis Advisor/Co-Advisor Saeed M Abdullah, Committee Member Mary Beth Rosson, Program Head/Chair
Keywords:
Symptom checker AI algorithms chatbot app review
Abstract:
In recent years, there has been a growing interest in developing Artificial Intelligence (AI)-based chatbots in the healthcare market, which use human-like conversations to interact with users. One popular type of AI-based chatbots is an AI-based symptom checker (AISC) app, which provides potential diagnoses for users and assists them with self-triaging. Despite the popularity of such AISC apps and their high ratings in major app stores, little research has been undertaken to investigate users’ perception, accountability, transparency, and data policies of AISC apps. To investigate AISC apps and explore how users evaluate and perceive the effectiveness of AISC apps, we conducted a feature review, a review analysis, and an interview study. We found that existing AISC apps lack credentials and transparency to verify their credibility and safety, which brings challenges to users’ trust and privacy protection. We also found that users evaluated AISC apps by comparing their experiences of using AISC apps with offline consulting experiences. Users perceived existing AISC apps lacked support for diverse diseases and user groups, flexible symptom inputs, user-friendly conversations, and comprehensive health history. Based on the results, we derived discussions and implications for AI policymaking, AI algorithms design, and conversational design of healthcare chatbots.
