Clinical obstacles to machine-learning POCUS adoption and system-wide AI implementation (The COMPASS-AI survey)

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Point-of-Care Ultrasound

Background:

• Point-of-care ultrasound (POCUS) is essential across medical fields.
• Artificial intelligence (AI) and machine learning (ML) offer potential to advance POCUS.
• Healthcare professional perspectives on AI integration in POCUS are not well understood.

Purpose of the Study:

• To assess global perceptions of AI in POCUS.
• To evaluate familiarity and adoption rates of AI within POCUS.
• To identify barriers to AI implementation in POCUS.

Main Methods:

• International web-based survey of POCUS-involved healthcare professionals.
• Survey covered demographics, AI familiarity, perceived utility, and barriers (technological, training, trust, workflow, legal/ethical).
• Data analyzed using descriptive statistics and comparative tests (chi-square, t-test).

Main Results:

• 1154 professionals surveyed on AI in POCUS barriers.
• 81.1% expressed enthusiasm for AI in POCUS.
• Top barriers: Training & Education (27.1%) and Clinical Validation & Evidence (17.5%).
• Perceptions varied by region, specialty, and experience.

Conclusions:

• Global survey offers insights into AI in POCUS perceptions and adoption.
• Significant enthusiasm exists alongside critical barriers in training, validation, and support.
• Addressing these challenges is vital for responsible AI implementation in POCUS.