AI support for ethical decision-making around resuscitation: proceed with care.

Area of Science:

• Medical Ethics
• Health Informatics
• Clinical Decision Support

Background:

• Artificial intelligence (AI) demonstrates high performance in healthcare, primarily in diagnosis and outcome prediction.
• The role of AI in supporting complex clinical decisions, especially those involving patient preferences, remains less defined.
• Ethical considerations for AI in healthcare are crucial, particularly for life-sustaining treatment decisions.

Purpose of the Study:

• To explore ethical questions surrounding AI design, development, and deployment for cardiopulmonary resuscitation (CPR) and Do Not Attempt Resuscitate (DNAR) decision-making.
• To understand current practices and challenges in resuscitation decision-making, exacerbated by situations like the COVID-19 pandemic.
• To assess clinician openness to AI-based decision support for code status determinations.

Main Methods:

• Qualitative study involving interviews with healthcare professionals at a university hospital.
• Exploration of the current status quo in resuscitation decision-making processes.
• Investigation into the potential role of AI systems in supporting DNAR decisions.

Main Results:

• Current resuscitation decision-making faces significant challenges, including lack of patient preference knowledge, time constraints, and clinician bias.
• Healthcare professionals expressed considerable openness to utilizing AI-based decision support tools for code status.
• A need for ethically informed AI development and implementation in critical care settings was identified.

Conclusions:

• AI holds potential to enhance decision-making processes for CPR and DNAR orders.
• Further development and implementation of AI in this domain require careful consideration of ethical, methodological, and procedural preconditions.
• Addressing challenges in current practices can be improved through thoughtful integration of AI decision support.