Zusammenfassung
Mit der voranschreitenden technischen Entwicklung werden Spracherkennungssysteme (SES) — gerade vor dem Hintergrund der aktuell unabweisbaren Kostenreduktion bei gleichbleibender Qualität in der Patientenversorgung — eine zunehmend attraktive Alternative zur traditionellen Befunderstellung.
Die 2 Hauptkomponenten eines SES sind das akustische und das Sprachmodell. Merkmale kontinuierlicher SES mit Realtimeerkennung umfassen vorformulierbare Befund(blöck)e, Standardbefundvorlagen und Sprachkommandos (Navigation im Text, Steuerung von SES und RIS). Sinnvoll für eine optimale Nutzung des SES-Potenzials ist die Integration von SES, RIS und PACS. Wichtige Leistungsparameter eines SES sind Befundverfügbarkeit und Zeiteffizienz des Befundungsprozesses (Erkennungsrate, Editier- und Korrekturaufwand, Wortschatzpflege) für den Radiologen.
In der Praxis wird die Erkennungsrate über die Fehlerrate (Einheit „Wort“) abgeschätzt. Fehlerraten liegen zwischen 4 und 28%. Etwa 20% davon sind Wortschatzfehler, die u. U. zu einer falschen Befundinterpretation führen können. Sie unterstreichen die Notwendigkeit einer sorgfältigen Textkorrektur und Wortschatzpflege.
Die Einführung eines SES erbringt eine drastische Verbesserung der Befundverfügbarkeit. Dagegen nimmt der individuelle ärztliche Zeitbedarf bei digitaler Befunderstellung um ca. 20–25% (Projektionsradiographie, CR) bzw. ca. 30% (CT, MRT) zu. Die Entlastung des Schreibbüros (Hintergrunddiktat) hängt von dessen Qualifikation ab. Das Onlinediktat führt zu einer Umverteilung von Arbeitsschritten vom Schreibbüro auf den Befunder.
Abstract
With ongoing technical refinements speech recognition systems (SRS) are becoming an increasingly attractive alternative to traditional methods of preparing and transcribing medical reports.
The two main components of any SRS are the acoustic model and the language model. Features of modern SRS with continuous speech recognition are macros with individually definable texts and report templates as well as the option to navigate in a text or to control SRS or RIS functions by speech recognition. The best benefit from SRS can be obtained if it is integrated into a RIS/RIS-PACS installation. Report availability and time efficiency of the reporting process (related to recognition rate, time expenditure for editing and correcting a report) are the principal determinants of the clinical performance of any SRS.
For practical purposes the recognition rate is estimated by the error rate (unit “word”). Error rates range from 4 to 28%. Roughly 20% of them are errors in the vocabulary which may result in clinically relevant misinterpretation. It is thus mandatory to thoroughly correct any transcribed text as well as to continuously train and adapt the SRS vocabulary.
The implementation of SRS dramatically improves report availability. This is most pronounced for CT and CR. However, the individual time expenditure for (SRS-based) reporting increased by 20–25% (CR) and according to literature data there is an increase by 30% for CT and MRI. The extent to which the transcription staff profits from SRS depends largely on its qualification. Online dictation implies a workload shift from the transcription staff to the reporting radiologist.
Literatur
Antiles S, Couris J, Schweitzer A et al. (2000) Project planning, training, measurement and sustainment: the successful implementation of voice recognition. Radiol Manage 22:18–31; quiz 32–16
Arndt H, Petersein J, Stockheim D et al. (1999) [The use of automated speech recognition in diagnostic radiology]. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 171:400–404
Callaway EC, Sweet CF, Siegel E et al. (2002) Speech recognition interface to a hospital information system using a self-designed visual basic program: initial experience. J Digit Imaging 15:43–53
Devine EG, Gaehde SA, Curtis AC (2000) Comparative evaluation of three continuous speech recognition software packages in the generation of medical reports. J Am Med Inform Assoc 7:462–468
Ernst R, Carpenter W, Torres W et al. (2001) Combining speech recognition software with digital imaging and communications in medicine (DICOM) workstation software on a Microsoft Windows platform. J Digit Imaging 14:182–183
Gale B, Safriel Y, Lukban A et al. (2001) Radiology report production times: voice recognition vs. transcription. Radiol Manage 23:18–22
Hayt DB, Alexander S (2001) The pros and cons of implementing PACS and speech recognition systems. J Digit Imaging 14:149–157
Herman SJ (1995) Accuracy of a voice-to-text personal dictation system in the generation of radiology reports. AJR Am J Roentgenol 165:177–180
Holman BL, Aliabadi P, Silverman SG et al. (1994) Medical impact of unedited preliminary radiology reports. Radiology 191:519–521
Hundt W, Stark O, Scharnberg B et al. (1999) Speech processing in radiology. Eur Radiol 9:1451–1456
Kanal KM, Hangiandreou NJ, Sykes AM et al. (2001) Initial evaluation of a continuous speech recognition program for radiology. J Digit Imaging 14:30–37
Lai J, Vergo J (1997) Med speak: report creation with continuous speech recognition. American Association for Computing Machinery Special Interest Group — Computers 97 conference proceedings, March 1997, pp 431–438
Langer SG (2002) Impact of tightly coupled PACS/speech recognition on report turnaround time in the radiology department. J Digit Imaging 15 [suppl 1]:234–236
Langer S (2002) Radiology speech recognition: workflow, integration, and productivity issues. Curr Probl Diagn Radiol 31:95–104
Leeming BW, Porter D, Jackson JD et al. (1981) Computerized radiologic reporting with voice data-entry. Radiology 138:585–588
Makhoul J, Schwartz R (1995) State of the art in continuous speech recognition. Proc Natl Acad Sci U S A 92:9956–9963
Mehta A (2002) Voice recognition. In: Dreyer K, Mehta A, Thrall J (eds) PACS: a guide to the digital revolution. Springer, Berlin Heidelberg New York, pp 281–302
Mehta A (2003) The Internet for radiology practice. Springer, Berlin Heidelberg New York
Mehta A, McLoud TC (2003) Voice recognition. J Thorac Imaging 18:178–182
Ramaswamy MR, Chaljub G, Esch O et al. (2000) Continuous speech recognition in MR imaging reporting: advantages, disadvantages, and impact. AJR Am J Roentgenol 174:617–622
Reed RA (1992) Voice recognition for the radiology market. Top Health Rec Manage 12:58–63
Robbins AH, Horowitz DM, Srinivasan MK et al. (1987) Speech-controlled generation of radiology reports. Radiology 164:569–573
Robbins AH, Vincent ME, Shaffer K et al. (1988) Radiology reports: assessment of a 5000-word speech recognizer. Radiology 167:853–855
Rosenthal DI, Chew FS, Dupuy DE et al. (1998) Computer-based speech recognition as a replacement for medical transcription. AJR Am J Roentgenol 170:23–25
Seltzer SE, Kelly P, Adams DF et al. (1997) Expediting the turnaround of radiology reports in a teaching hospital setting. AJR Am J Roentgenol 168:889–893
Sferrella SM (2003) Success with voice recognition. Radiol Manage 25:42–49
Vorbeck F, Ba-Ssalamah A, Kettenbach J et al. (2000) Report generation using digital speech recognition in radiology. Eur Radiol 10:1976–1982
Zafar A, Overhage JM, McDonald CJ (1999) Continuous speech recognition for clinicians. J Am Med Inform Assoc 6:195–204
Zafar A, Mamlin B, Perkins S et al. (2004) A simple error classification system for understanding sources of error in automatic speech recognition and human transcription. Int J Med Inform 73:719–730
Interessenkonflikt:
Keine Angaben
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Glaser, C., Trumm, C., Nissen-Meyer, S. et al. Spracherkennung: Auswirkung auf Workflow und Befundverfügbarkeit. Radiologe 45, 735–742 (2005). https://doi.org/10.1007/s00117-005-1253-7
Issue Date:
DOI: https://doi.org/10.1007/s00117-005-1253-7