Linguistically Communicating Uncertainty in Patient-Facing Risk Prediction Models

TL;DR

This paper explores how to effectively communicate uncertainty in AI risk prediction models to patients by addressing natural language communication challenges, with a focus on IVF outcome prediction.

Contribution

It introduces a novel design for communicating model uncertainty in natural language tailored for patient-facing healthcare applications.

Findings

Identifies key challenges in natural language communication of uncertainty.

Proposes a design addressing communication and understandability issues.

Focuses on IVF outcome prediction as a case study.

Abstract

This paper addresses the unique challenges associated with uncertainty quantification in AI models when applied to patient-facing contexts within healthcare. Unlike traditional eXplainable Artificial Intelligence (XAI) methods tailored for model developers or domain experts, additional considerations of communicating in natural language, its presentation and evaluating understandability are necessary. We identify the challenges in communication model performance, confidence, reasoning and unknown knowns using natural language in the context of risk prediction. We propose a design aimed at addressing these challenges, focusing on the specific application of in-vitro fertilisation outcome prediction.