Healthcare News & Insights

Speech recognition software leads to major errors, says new report

Healthcare technology has many benefits for patients and doctors, but if it’s not used properly, it can create risks as well. A new study warns providers of how one common technology can lead to critical errors.

In medical reports, especially ones as critical as breast MRI reports, mistakes are unacceptable. However, recent studies found that automatic speech recognition (ASR) software often contains major errors. In fact, there were significantly more errors (23%) when the radiologist used ASR than when the radiologist used conventional dictation transcription (only 4% of these had errors).

ASR software automatically transcribes reports dictated by a radiologist. Traditionally, conventional dictation transcription is when the radiologist dictates the report to a team, who will then transcribe and review the report.

All of the errors counted were ones that could be classified as “major” because they affected patient management or provided incorrect measurements. In breast MRI reports, the major error rate was even higher than the average (35% for ASR and 7% for conventional).

This doesn’t mean ASR software should be ignored altogether. Experts say with proper dictation and navigation of the software, many errors can be avoided. Doctors and other staff members must be trained to check, double check, and triple check reports for errors; careful editing is a must.

Also, many problems can be avoided by choosing the correct vendor — most importantly, one that meshes well with your workflow.

Subscribe Today

Get the latest and greatest healthcare news and insights delivered to your inbox.