Exploring the Effect of Explanation Content and Format on User Comprehension and Trust in Healthcare

TL;DR

This study investigates how explanation content and format influence user understanding and trust in AI healthcare tools, finding that explanation format often has a greater impact than content on user perceptions.

Contribution

It provides empirical evidence on the effects of explanation content and format on user trust and comprehension in healthcare AI, highlighting the importance of presentation style.

Findings

Occlusion-1 explanations increased trust and comprehension more than SHAP.

Text explanations were preferred over chart formats when controlling for content.

Explanation format significantly influences user perception more than explanation content.

Abstract

AI-driven tools for healthcare are widely acknowledged as potentially beneficial to health practitioners and patients, e.g. the QCancer regression tool for cancer risk prediction. However, for these tools to be trusted, they need to be supplemented with explanations. We examine how explanations' content and format affect user comprehension and trust when explaining QCancer's predictions. Regarding content, we deploy the SHAP and Occlusion-1 explanation methods. Regarding format, we present SHAP explanations, conventionally, as charts (SC) and Occlusion-1 explanations as charts (OC) as well as text (OT), to which their simpler nature lends itself. We conduct experiments with two sets of stakeholders: the general public (representing patients) and medical students (representing healthcare practitioners). Our experiments showed higher subjective comprehension and trust for Occlusion-1 over SHAP explanations based on content. However, when controlling for format, only OT outperformed SC, suggesting this trend is driven by preferences for text. Other findings corroborated that explanation format, rather than content, is often the critical factor.