Palliative Care Physicians' Perceptions about Using Artificial Intelligence for Prognostication

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Palliative Care Research

Background:

• Statistical and AI methods have long informed clinical prognostication.
• Machine learning models are increasingly integrated into electronic health records.
• Physician perceptions of AI prognostic tools in palliative care are understudied.

Purpose of the Study:

• To assess palliative care physician perspectives on the clinical use and implementation of AI-based prognostic tools.
• To understand current prognostic practices and AI knowledge among these physicians.

Main Methods:

• A national survey of 2,500 Hospice and Palliative Medicine physicians in the US was conducted.
• The survey assessed current prognostic practices, AI knowledge, and perceived benefits/risks of AI prognostication.
• 537 completed surveys were analyzed.

Main Results:

• Most respondents were early technology adopters but had low knowledge of AI and AI-based prognostication.
• While 72.9% provide prognoses routinely, only 24.0% have used AI-generated prognoses.
• Physicians generally hold favorable attitudes, believing AI can facilitate earlier palliative care, improve end-of-life planning, and reduce team disagreements, though concerns about liability and patient hope exist.

Conclusions:

• Palliative care physicians report limited current use of AI prognostic tools.
• Attitudes towards AI-based prognostication are generally favorable, with current users expressing more positive beliefs.
• Further research and education may be needed to address concerns and promote appropriate AI adoption.