Interactive versus static decision support tools for COVID-19 - Trial DRKS00028136

Timeline & Enrollment

Summary

In early and mid-2020 numerous online tools appeared advising laypersons on whether their medical symptoms indicate that they are at risk of having acquired a COVID-19 infection. These publicly availabe decision support systems were designed to aid laypersons in self-triaging, and deciding whether self-isolation or quarantine was required of them. Content-wise most tools were very similar basing their recommendations on the official guidelines by the World Health Organisation and/or the national public health agency, i.e. for the United States the Center for Disease Control. However, the appearance of these tools differed substantially with some being designed as simple flowcharts while others were programmed as more interactive conversational agents. Our study investigated whether these different modes of presentation impact their (perceived) capability to support layperson decision-making. In November 2020, we recruited 196 US residents via prolific.co to participate online. We asked them to appraise seven fictitious descriptions of patients (clinical vignettes) with or without symptoms indicative of COVID-19. We asked the study participants to decide on whether these fictitious patients should be advised to seek healthcare and how they should behave, for example whether self-isolation or quarantine is necessary. Participants were randomly assigned to three groups: in one group participants were supported by a chatbot-like decision support tool mimicking ones publicly available at that time. The second group received advice from static decision support tool (a flowchart), again mimicking ones publicly available at that time. Participants in the control group did not receive any decision support. To assess whether one kind of tool was more helpful than the other, we analysed participants decision accuracy and surveyed their mental effort in deciding, their decisional uncertainty, their self-reported trust in the tools, the perceived usefulness and ease of use of the tools, and lastly whether they would be inclined to use these mock tools again.

ICD-10 Classifications