Benchmarking a deep learning model against healthcare practitioners for hip fracture detection in the emergency department

Key Points

• Evaluating the performance of a deep learning model for hip fracture detection compared to radiographers' assessments.
• Benchmarking deep learning model against radiographers' diagnoses
• Assessing reading time for fracture detection
• Evaluating the integration of multimodal clinical data
• Deep learning model did not match radiographers' performance
• Short reading time may help in reducing diagnostic delays
• Further development of the model is required with higher-resolution images