Comparative Evaluation of Three Continuous Speech Recognition Software Packages in the Generation of Medical Reports

Devine, Eric; Gaehde, Stephan A.; Curtis, Arthur C.

doi:10.1136/jamia.2000.0070462

Cited by 69 publications

(49 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Past user studies reported WERs of 6-11% [6], 7-15% [2], and 15% [4]. As these studies used older recognizers, we did our own testing using Dragon v9.…”

Section: System Speech Recognizermentioning

confidence: 99%

On the benefits of confidence visualization in speech recognition

Vertanen

Kristensson

2008

Proceedings of the SIGCHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

In a typical speech dictation interface, the recognizer's bestguess is displayed as normal, unannotated text. This ignores potentially useful information about the recognizer's confidence in its recognition hypothesis. Using a confidence measure (which itself may sometimes be inaccurate), we investigated providing visual feedback about low-confidence portions of the recognition using shaded, red underlining. An evaluation showed, compared to a baseline without underlining, underlining lowconfidence areas did not increase user's speed or accuracy in detecting errors. However, we found that when recognition errors were correctly underlined, they were discovered significantly more often than baseline. Conversely, when errors failed to be underlined, they were discovered less often. Our results indicate confidence visualization can be effective -but only if the confidence measure has high accuracy. Further, since our results show that users tend to trust confidence visualization, designers should be careful in its application if a high accuracy confidence measure is not available.

show abstract

“…Past user studies reported WERs of 6-11% [6], 7-15% [2], and 15% [4]. As these studies used older recognizers, we did our own testing using Dragon v9.…”

Section: System Speech Recognizermentioning

confidence: 99%

On the benefits of confidence visualization in speech recognition

Vertanen

Kristensson

2008

Proceedings of the SIGCHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

show abstract

“…Several previous studies have confirmed recognition rates for speech recognition software from 90% to 98%. 5,6,7,9 At WP, all dictated reports by all radiologists were included. At this site, only one tier of verification is required, that of the dictating radiologist.…”

Section: Datamentioning

confidence: 99%

Speech recognition interface to a hospital information system using a self-designed visual basic program: Initial experience

et al. 2002

View full text Add to dashboard Cite

Speech recognition (SR) in the radiology department setting is viewed as a method of decreasing overhead expenses by reducing or eliminating transcription services and improving care by reducing report turnaround times incurred by transcription backlogs. The purpose of this study was to show the ability to integrate off-the-shelf speech recognition software into a Hospital Information System in 3 types of military medical facilities using the Windows programming language Visual Basic 6.0 (Microsoft, Redmond, WA). Report turnaround times and costs were calculated for a medium-sized medical teaching facility, a medium-sized nonteaching facility, and a medical clinic. Results of speech recognition versus contract transcription services were assessed between July and December, 2000. In the teaching facility, 2,042 reports were dictated on 2 computers equipped with the speech recognition program, saving a total of US $3,319 in transcription costs. Turnaround times were calculated for 4 first-year radiology residents in 4 imaging categories. Despite requiring 2 separate electronic signatures, we achieved an average reduction in turnaround time from 15.7 hours to 4.7 hours. In the nonteaching facility, 26,600 reports were dictated with average turnaround time improving from 89 hours for transcription to 19 hours for speech recognition saving US $45,500 over the same 6 months. The medical clinic generated 5,109 reports for a cost savings of US $10,650. Total cost to implement this speech recognition was approximately US $3,000 per workstation, mostly for hardware. It is possible to design and implement an affordable speech recognition system without a large-scale expensive commercial solution.KEY WORDS: computer, speech recognition, picture archiving and communication systems, interface, composite health-care system THE EFFORT to improve patient care by .1 collapsing the diagnostic and therapeutic timeline has driven computer applications development in a variety of areas. Tremendous improvements in hardware and software over Journal of Digital Imaging, Vol 15, No 1 (March}, 2002: pp 43-53 the last decade have stimulated this progress. With picture archiving and communication system (PACS) technology, images are available immediately throughout the health care system, but there continues to be a lag in the transmission of the corresponding completed radiology reports. 1,2 The purpose of this report is to relate our experience with the development and integration of an off-the-shelf speech/voice recognition application into a hospital information system (HIS) using a graphical interface program developed by one of the authors.Speech recognition in the radiology department setting decreases overhead expenses by reducing or eliminating transcription services or as a means to improve patient care by reducing report turnaround times. 2 ,3 Significant problems can arise in facilities that attempt to integrate a speech recognition system into the HIS. 4 This can be difficult particularly in the setting of a training program...

show abstract

“…Many studies on speech recognition in the area of healthcare in other languages than Finnish have been presented. For example, Devine et al [2] compared the performance attributes of three commercial English-language speech recognition software with physicians, and found that utilizing such a technology in healthcare transcription showed feasibility already in 2000. Borowitz [1] compared human and computerized dictation transcription at a pediatric gastroenterology clinic in terms of dictation, editing and transcription completion times, e.g.…”

Section: Introductionmentioning

confidence: 98%

Mobile dictation for healthcare professionals

Keskinen

Melto

Hakulinen

et al. 2013

Proceedings of the 12th International Conference on Mobile and Ubiquitous Multimedia

View full text Add to dashboard Cite

We demonstrate a mobile dictation application utilizing automatic speech recognition for healthcare professionals. Development was done in close collaboration between human-technology interaction and nursing science researchers and professionals working in the area. Our work was motivated by the need for improvements in getting spoken patient information to the next treatment steps without additional steps. In addition, we wanted to enable truly mobile spoken information entry, i.e., dictation can take place on the spot. In order to study the applicability we conducted a small-scale Wizard-of-Oz evaluation in a real hospital environment with real nurses. Our main focus was to gather subjective expectations and experiences from the actual nurses themselves. The results show true potential for our mobile dictation application and its further development.

show abstract

Comparative Evaluation of Three Continuous Speech Recognition Software Packages in the Generation of Medical Reports

Cited by 69 publications

References 1 publication

On the benefits of confidence visualization in speech recognition

On the benefits of confidence visualization in speech recognition

Speech recognition interface to a hospital information system using a self-designed visual basic program: Initial experience

Mobile dictation for healthcare professionals

Contact Info

Product

Resources

About