UK reporting radiographers' perceptions of AI in radiographic image interpretation - Current perspectives and future...

Area of Science:

• Medical Imaging Informatics
• Artificial Intelligence in Healthcare
• Radiography

Background:

• Radiographer reporting is established in the UK, facing challenges from radiographer and radiologist shortages.
• Artificial intelligence (AI) offers potential to reduce unreported image backlogs.
• Understanding end-user perceptions of AI is crucial for ethical integration and trust.

Purpose of the Study:

• To investigate UK reporting radiographers' perceptions of AI in image interpretation.
• To understand potential future interactions between radiographers and AI systems.
• To identify features necessary for building appropriate trust in AI for reporting.

Main Methods:

• A Qualtrics survey was designed and piloted by UK AI expert radiographers.
• The survey targeted reporting radiographers, with this study analyzing the third part of the survey.
• 86 responses were collected and analyzed.

Main Results:

• A majority of respondents (62%) felt confident in AI's decision-making process.
• Less than a third felt confident communicating AI decisions to stakeholders.
• AI affirmation improved confidence (57%), while disagreement prompted second opinions (70%); moderate trust exists, improvable with performance data and visual explanations.

Conclusions:

• AI is expected to significantly impact future reporting radiographer decision-making.
• Confidence in AI decision-making contrasts with lower confidence in explaining these decisions.
• Explainable AI solutions are key to improving trust levels in AI for image interpretation.