The Role of AI in Advance Care Planning

What is Advance Care Planning (ACP) and Why is it Important?

Advance Care Planning (ACP) is a structured process that enables individuals to articulate their healthcare preferences and values in advance, particularly in cases where they may lose the capacity to communicate their wishes. It plays a crucial role in end-of-life care, ensuring that patients' desires are respected and guiding healthcare providers in making appropriate decisions in critical situations. ACP fosters patient autonomy and helps mitigate family stress during difficult moments. Studies have shown that ACP leads to better patient and family satisfaction, reduces unnecessary hospital admissions, and aligns medical care with patients' personal values (1).

Artificial intelligence in ACP

The Current Use of AI in Healthcare Conversations

Artificial intelligence (AI) is increasingly being integrated into healthcare, and one of the key areas of exploration is its role in patient-provider conversations. AI-driven chatbots are currently used to support routine tasks, such as providing information, scheduling appointments, and managing chronic diseases (2). These chatbots are often powered by natural language processing (NLP) algorithms, which allow them to interpret and respond to patient queries in real time. For example, chatbots like Woebot have shown success in mental health interventions, using conversational AI to deliver cognitive behavioral therapy to patients with anxiety and depression (3).

In recent years, AI has begun to show promise in healthcare conversations, offering tools that can help bridge gaps in communication, reduce the burden on healthcare professionals, and improve access to timely information. But its role in more emotionally charged and complex conversations—like those surrounding ACP—is still under scrutiny.

Pain Points of ACP Facilitation

ACP discussions, while essential, are often fraught with emotional and logistical challenges. Many patients and families struggle with the emotional weight of planning for future incapacity or death. Healthcare professionals often face time constraints, leaving insufficient time for meaningful ACP conversations. Additionally, the variability in facilitators’ approaches can lead to inconsistencies in the quality of discussions. In some cases, patients are unable to express their preferences clearly, resulting in care that may not align with their values.

Furthermore, disparities in patient education, language barriers, and health literacy can make ACP discussions even more difficult, as patients may not fully understand the options available to them or the implications of their decisions.

How AI Can Address These Pain Points

AI has the potential to alleviate some of the challenges associated with ACP discussions by offering consistent, accessible, and non-judgmental facilitation. AI-powered chatbots can provide patients with 24/7 access to information, allowing them to reflect on their healthcare preferences at their own pace before engaging in face-to-face discussions with a healthcare provider. This can help reduce the pressure of having to make decisions in a short time frame and provide more comprehensive, personalized guidance.

Additionally, AI can standardize ACP conversations, ensuring that all important topics are covered and facilitating more structured discussions. By offering prompts and clarifications based on patients' responses, chatbots can help patients understand complex medical concepts in a way that aligns with their health literacy levels . AI can also provide translations or cultural adaptations to improve accessibility for non-native speakers.

Strengths and Weaknesses of Using AI in ACP Discussions

Strengths:

• Consistency and Standardization: AI can ensure that key elements of ACP are consistently discussed, reducing the variability seen in human-facilitated conversations.

• Accessibility: AI chatbots are accessible anytime and can provide support in multiple languages, ensuring that ACP discussions are available to a wider population.

• Reduced Pressure: Patients can engage with the chatbot at their own pace, without the time constraints that might accompany in-person conversations.

• Improved Decision-Making: By presenting patients with personalized, scenario-based prompts, AI can help patients think through the implications of their choices more comprehensively.

Weaknesses:

• Lack of Emotional Intelligence: While AI can provide structured information, it cannot replicate the empathy, nuance, and sensitivity that human facilitators bring to ACP discussions, which is crucial when dealing with such emotionally charged topics.

• Trust Issues: Many patients express concerns about data privacy and trust when interacting with AI systems, particularly in sensitive areas like end-of-life care (4).

• Limited Comprehensiveness: AI systems may still miss nuances or fail to address specific patient concerns that a human facilitator might naturally pick up on during a conversation (5).

• 
  

What Needs to Be Done to Make AI in ACP Possible

For AI to play a meaningful role in ACP discussions, several improvements and safeguards are needed. First, AI chatbots must be enhanced to detect emotional cues and respond in a way that is perceived as empathetic, or at the very least, supportive. Integrating more advanced NLP algorithms and machine learning techniques can help AI systems adapt better to the emotional and contextual nuances of ACP.

Second, healthcare providers must establish clear data protection protocols to ensure patients' information is secure. Given the personal nature of ACP discussions, addressing concerns about privacy and data misuse is critical for gaining patient trust in AI-driven solutions.

Finally, AI should be integrated into a hybrid facilitation model, where it complements human facilitators rather than replacing them. AI could handle preliminary discussions or routine information-sharing, leaving human facilitators to focus on emotional support and decision-making verification.

Ethical Considerations in Using AI for ACP

The use of AI in ACP raises several ethical concerns, most notably around data security, informed consent, and patient autonomy. Given the sensitive nature of ACP, patients must feel confident that their personal data is protected and that AI systems will not misuse or share this information without their explicit consent (6).

Moreover, there is the question of how much autonomy should be delegated to AI systems. While AI can guide patients through their options, it must not overstep its role by making assumptions about what is best for the patient. Ensuring that patients remain fully in control of their decisions is essential.

Another key ethical issue is equity of access. AI systems must be designed to be inclusive, taking into account cultural, linguistic, and socio-economic factors that might affect a patient's ability to engage with the technology.

Conclusion

AI holds significant promise for improving the accessibility and consistency of Advance Care Planning, offering patients and healthcare providers valuable tools to navigate difficult conversations. However, its role must be carefully considered and designed to complement—rather than replace—the human element in ACP discussions. By addressing current limitations, ensuring privacy and empathy, and integrating AI into hybrid care models, we can leverage AI’s potential to support more personalized and effective ACP conversations.

Author(s)

The Article was written by Dr Laurence Tan, a Palliative Physician working in Singapore.

Acknowledgements

The main text of this article includes contents generated by ChatGPT, a language model developed by OpenAI in San Francisco, CA, USA. Additionally, ChatGPT provided assistance in sentence editing.

References

1. Detering KM, Hancock AD, Reade MC, Silvester W. The impact of advance care planning on end of life care in elderly patients: randomised controlled trial. BMJ. 2010 Mar 23;340

.

2. Laranjo L, Dunn AG, Tong HL, Kocaballi AB, Chen J, Bashir R, et al. Conversational agents in healthcare: A systematic review. J Am Med Inform Assoc. 2018;25(9):1248-58.

3. Fitzpatrick KK, Darcy A, Vierhile M. Delivering cognitive behavior therapy to young adults with symptoms of depression and anxiety using a fully automated conversational agent (Woebot): A randomized controlled trial. JMIR Ment Health. 2017;4(2)

4. Kopecky K, Krijestorac A, Finlay M. Cybersecurity and AI in healthcare: Building trust in intelligent systems. Healthc Inform Res. 2021 Jan;27(1):32-41.

5. Tulsky JA, Beach MC, Butow PN, Hickman SE, Mack JW, Morrison RS, et al. A research agenda for communication at the end of life. J Palliat Med. 2017 Jul 1;20(7):729-38.

6. Nebeker C, Torous J, Bartlett Ellis RJ. Building the case for actionable ethics in digital health research supported by artificial intelligence. BMC Med. 2019;17(1):137.