PreCare: Designing AI Assistants for Advance Care Planning (ACP) to Enhance Personal Value Exploration, Patient Knowledge, and Decisional Confidence

TL;DR

This work tackles the gap between online Advance Care Planning (ACP) and clinical ACP consultations by designing PreCare, an AI-assisted ACP platform. It derives design insights from two formative studies with ACP professionals and online-ACP users, then validates three AI assistants—values introspection, real-time knowledge Q&A with curated FAQs, and personalized impact analysis—through usability and comparative UX studies. Results show excellent usability and significant improvements in exploration of personal values, ACP knowledge, and decisional confidence, with 92% of participants preferring AI-assisted PreCare. The study demonstrates the feasibility and value of integrating domain-knowledge, safety, and personalized AI reasoning into patient-facing ACP tools, with broad implications for decision-making systems, SDM, and future thanato-technology.

Abstract

Advance Care Planning (ACP) allows individuals to specify their preferred end-of-life life-sustaining treatments before they become incapacitated by injury or terminal illness (e.g., coma, cancer, dementia). While online ACP offers high accessibility, it lacks key benefits of clinical consultations, including personalized value exploration, immediate clarification of decision consequences. To bridge this gap, we conducted two formative studies: 1) shadowed and interviewed 3 ACP teams consisting of physicians, nurses, and social workers (18 patients total), and 2) interviewed 14 users of ACP websites. Building on these insights, we designed PreCare in collaboration with 6 ACP professionals. PreCare is a website with 3 AI-driven assistants designed to guide users through exploring personal values, gaining ACP knowledge, and supporting informed decision-making. A usability study (n=12) showed that PreCare achieved a System Usability Scale (SUS) rating of excellent. A comparative evaluation (n=12) showed that PreCare's AI assistants significantly improved exploration of personal values, knowledge, and decisional confidence, and was preferred by 92% of participants.

Figures (5)

• Figure 1: Comparison between (1) Current online ACP and (2) PreCare Design in terms of (A) Exploration of Personal Values, (B) ACP Knowledge, and (C) Making decisions. In the Current online ACP: (A1) Participants use a text area to answer questions for exploration of personal values. However, the input is not used for further discussion, and users may simply skip important questions that are crucial for deeper reflection; (A2) If participants have questions, they need to search the internet, but much of the information found online is not verified by professionals. Additionally, some users may be unaware of the knowledge gaps they have or may not know what or how to ask; (A3) Participants miss some aspects before making a decision, and the website does not provide some crucial information before making decisions. In PreCare Design: (B1) We utilized interactive conversations, designed with domain knowledge of social workers, to facilitate introspection of personal values. Compared to current online ACP platforms, which are limited by a lack of personalized questioning, our assistant asks more follow-up questions to encourage a deeper exploration of personal values; (B2) We integrated physicians' domain knowledge into a prioritized list of frequently asked questions and enabled real-time Q&A for participants; (B3) By providing comprehensive information and facilitating interactive conversations, with input from physicians, nurses, and social workers, PreCare ensures that participants consider all important aspects before making their decisions.
• Figure 2: Screenshots of the 3 PreCare AI Assistants: (A) two-way conversation to ensure thorough introspection of users' personal values. (B1) Each key medical knowledge topic includes prioritized, top 3 FAQs and a text field for real-time Q&A. (B2) Shows an actual Q&A example from the study (P10), the Q&A was appended to the FAQ to enable users to continue asking additional questions. (C) conversational assistant reviews important aspects before making final decisions.
• Figure 3: UI and Assistant Design Iteration: (A) Initial vs. (B) Final designs. For Assistant 1, personal value exploration, (A1) testing of the initial design revealed limited follow-up capability; (B1) we added a response evaluation system to enhance follow-up capability. For Assistant 2, ACP knowledge, (A2) the initial design featured a standard chatbot interface for Q&A; (B2) based on feedback, we integrated Q&A into the FAQ to improve usability.
• Figure 4: User experience study participants' script analysis and feedback on the AI assistants: (A) the word count in PreCare with AI assistants was significantly higher(p<.001) than without AI assistants; and (B)Participants’ ratings regarding the perceived quality of the AI assistant according to metrics by Abd-Alrazaq et al. AbdAlrazaq2020Technicalmetricsusedto
• Figure 5: Preference ratings for Advance Care Planning were measured on a 10-point scale, comparing PreCare with versus without AI assistants. Participants significantly preferred PreCare with AI assistants for all aspects, including Exploration of Personal Values(p<.01), Knowledge Completeness(p<.01), Decisional Confidence(p<.05), and Overall(p<.05).