Determinants of Laypersons' Trust in Medical Decision Aids

Symptom checker apps are decision support systems that help lay people decide whether, where and how to seek medical help. To support decision-making, these systems must not only provide accurate recommendations, but also be trustworthy. To date, few studies have examined the extent to which laypeople trust symptom checker apps and which factors influence trust. In other decision support systems, for example, anthropomorphism (humanizing the systems) has been shown to increase trust. In practice, a some developers also use the visualization of an artificial intelligence. This study aimed to investigate various influences on laypersons' trust in symptom checker apps. Primarily, we investigated the influence of anthropomorphic framing and framing as AI on trust. We hypothesized that both types of framing would increase trust. Using an online experiment, we recruited 494 US residents. They were asked to indicate whether or not medical help should be sought for a case vignette. Afterwards, the result of a symptom checker app was shown (which always contradicted the participants' answer). Subsequently, participants were asked to appraise the case vignette again.