Radiological reporting that combine continuous speech recognition with error correction by transcriptionists.

Area of Science:

• Radiology
• Medical Informatics
• Speech Recognition Technology

Background:

• Radiological reporting is crucial for patient diagnosis and treatment.
• Traditional dictation and transcription methods can be time-consuming and prone to errors.
• Advancements in speech recognition technology offer potential improvements in reporting efficiency.

Purpose of the Study:

• To evaluate the effectiveness of combining continuous speech recognition (CSR) with manual transcriptionist error correction for radiological reporting.
• To compare the accuracy and time efficiency of this combined method against conventional transcription.
• To assess the impact of transcriptionist experience on the performance of the combined method.

Main Methods:

• Four transcriptionists (varying experience levels) created radiological reports from 100 dictation files using two methods: conventional transcription and CSR with manual error correction.
• Accuracy and report creation time were measured for both methods.
• Inter-transcriptionist variability in accuracy and time was analyzed.

Main Results:

• No significant difference in accuracy was found between conventional transcription and the CSR-combined method.
• While experienced transcriptionists were faster with conventional methods, error correction speed with the CSR-combined method showed no significant difference based on experience.
• The CSR system used did not require voice-specific training.

Conclusions:

• The integration of CSR with manual error correction by transcriptionists provides a practical and accurate solution for radiological reporting.
• This combined approach streamlines the reporting process without compromising accuracy, benefiting both novice and experienced transcriptionists.