Ha-Kyung Kong, Ruijia Zhu, Audrey Zhixin Lu, Mingao He, Weichao Yuwen
Abstract: Chatbots are increasingly used in healthcare as an accessible and scalable means to communicate information. While visualization can provide summaries of health data, their impact on the comprehension, perceived usefulness, and trustworthiness of chatbots is yet unknown. We examined the effects of visualization (intervention group) of health data compared to text-only representation (control group) among 96 family caregivers in a randomized controlled trial. The results showed that visualizing the relationship between symptom levels (e.g. stress) and health solutions had a significant impact on the comprehension of the health trends and the perceived effectiveness of the solution. While there was an increased trust level in the visualization group, the difference was not statistically significant. The qualitative feedback aligned with the quantitative results. We discussed future directions in leveraging the synergistic effects of visualization and chatbots to increase health literacy and encourage technology adoption.