Artificial Intelligence applications in Advance Care Planning

TL;DR

This paper reviews how artificial intelligence can improve advance care planning by enhancing documentation, decision-making, and access, while also highlighting challenges like bias and ethical concerns.

Contribution

The paper provides a novel scoping review of AI applications in advance care planning, emphasizing equity, communication-centered approaches, and the need for ethical AI implementation.

Findings

AI tools like predictive models and chatbots improve ACP by identifying high-risk patients and increasing literacy.

Natural Language Processing significantly reduces ACP documentation review time.

Only 25% of AI tools address caregiver burden or surrogate decision-maker needs.

Abstract

This scoping review examines artificial intelligence (AI) applications in advance care planning (ACP), synthesizing evidence from databases like PubMed, Web of Science, and Scopus (2019-2024), ten articles, using PRISMA guidelines. The review focused on AI tools for ACP documentation, decision-making, and workflow optimization, excluding studies lacking clinical validation or patient-centered outcomes. Findings reveal promising AI applications in ACP: predictive models, like Stanford’s EHR-based mortality predictor, improved high-risk patient identification, triggering earlier ACP discussions. Natural Language Processing reduced ACP documentation review time significantly. AI-powered chatbots increased ACP literacy among older adults by 40%, addressing health disparities. Challenges like algorithmic bias from non-diverse datasets exacerbates inequities in ACP access persist. Clinician resistance stems from unclear liability frameworks and poor EHR interoperability. Only 25% of tools address caregiver burden or surrogate decision-maker needs. Research gaps include limited evidence on AI-driven e-planning tools’ cost-effectiveness, lack of standardized metrics for evaluating AI’s impact on patient autonomy and understudied ethical risks of over-reliance on AI for sensitive end-of-life conversations. Future directions focus on equity-driven design, the integration of e-planning into electronic medical records, and the assessment of AI’s capacity to alleviate caregiver burden and enhance preparedness of surrogate decision-makers. This research emphasizes the necessity to transform ACP from a documentation-oriented process to a communication-centered continuum, utilizing AI to democratize access and tackle systemic inequalities. It highlights the potential of AI to transform ACP, while emphasizing the importance of ethical implementation and patient-centered outcomes.