Female radiology residents tend to commit transcription errors at a significantly higher rate compared with male residents when using commercial voice recognition applications, according to a recent study.
Female radiology residents tend to commit transcription errors at a significantly higher rate compared with male residents when using commercial voice recognition applications, according to a recent study.
Error rates among male residents were found to range from 2.5% to 13.9% on 50 dictated reports, while female error rates on the same number of dictated reports ranged from 1.5% to 20.6%. Interestingly, both the lowest error rate (1.5%) and the highest error rate (20.5) on a single report were by female residents.
Results of the study were presented at the May meeting of the American Roentgen Ray Society.
"The discrepancy may have a significant negative impact on reporting accuracy and productivity for female radiologists," said lead author Dr. Syed Ali, a University of Maryland radiology resident.
The discovery was made after the university's radiology department made the transition recently to speech recognition from digital dictation. The move prompted department officials to investigate how to increase accuracy rates and what factors adversely affected speech recognition.
During the investigation, the department noticed some of the residents had higher accuracy rates than others, which raised questions about the relative importance of various U.S. or foreign accents on speech recognition accuracy. This subsequently prompted the researchers to also wonder whether error rates were gender related.
"We are not aware of any previous reports in the literature pertaining to this topic," Ali said.
The immediate impact of the study for radiologists is an increased level of awareness that women may need to spend more time training on the system than their male counterparts and may have to work somewhat harder to make the system successful. This could include more time training the system to recognize particular words or phrases or using macros or actually altering their dictation style to increase recognition rates, he said.
"As an increasing number of nonradiologist physicians use speech recognition as part of the medical record, we anticipate a similar bias toward higher recognition rates for males," Ali said.
Any efforts to improve recognition rates will have a positive impact on physicians and patients by reducing error rates and improving productivity, he said.
It may also be possible to "preprocess" speech for women to increase accuracy using software that can alter the pitch and volume of voice files, Ali said.
MRI-Based AI Radiomics Model Offers 'Robust' Prediction of Perineural Invasion in Prostate Cancer
July 26th 2024A model that combines MRI-based deep learning radiomics and clinical factors demonstrated an 84.8 percent ROC AUC and a 92.6 percent precision-recall AUC for predicting perineural invasion in prostate cancer cases.
Breast MRI Study Examines Common Factors with False Negatives and False Positives
July 24th 2024The absence of ipsilateral breast hypervascularity is three times more likely to be associated with false-negative findings on breast MRI and non-mass enhancement lesions have a 4.5-fold likelihood of being linked to false-positive results, according to new research.
Can Polyenergetic Reconstruction Help Resolve Streak Artifacts in Photon Counting CT?
July 22nd 2024New research looking at photon-counting computed tomography (PCCT) demonstrated significantly reduced variation and tracheal air density attenuation with polyenergetic reconstruction in contrast to monoenergetic reconstruction on chest CT.