voice recognition in the electronic health record diane luedtke nursing informatics, nsg600ina...

19
Voice Recognition in the Electronic Health Record Diane Luedtke Nursing Informatics, NSG600INA November, 2010

Upload: elwin-chambers

Post on 23-Dec-2015

213 views

Category:

Documents


0 download

TRANSCRIPT

  • Slide 1
  • Slide 2
  • Voice Recognition in the Electronic Health Record Diane Luedtke Nursing Informatics, NSG600INA November, 2010
  • Slide 3
  • Speech Recognition Definition The process of converting an acoustic signal, captured by a microphone or a telephone to a set of words.
  • Slide 4
  • History 1952 - Recognition of single digits 1964 Device exhibited at NY Worlds Fair 1980s 1,000 to 20,000 word vocabularies Early 90s Accuracy 10% to 50% and discrete voice recognition 1997 Recognition of normal speech Early 2000s Accuracy 80%
  • Slide 5
  • Types of Speech Recognition Isolated - pause between words Continuous no pause between words Spontaneous extemporaneous most difficult to recognize
  • Slide 6
  • Properties Speaker enrollment Speaker independent Finite state network General language models Perplexity External parameters
  • Slide 7
  • Variables Phonemes Acoustic variables Within speaker variables Across speaker variables Zue, V., Cole, R., Ward, W. Speech recognition. Retrieved from http://cslu.cse.ogi.edu/HLTsurvey/ch1node4.html on 10/6/2010.http://cslu.cse.ogi.edu/HLTsurvey/ch1node4.html
  • Slide 8
  • http://www.google.com/imgres?imgurl=http://static.howstuffworks.com/gif/speech-recognition-process.gif
  • Slide 9
  • Speech Recognition in Health Care Earliest users radiologists Most successful early users radiologists, pathologists and emergency physician s Photo source:www.google.com/imgres?imgurl=http://www.rsna.org/Publications/RSNAnews/November -2010/images_speech_recognition_1.jpg
  • Slide 10
  • Other Healthcare Settings Primary care clinicians Psychiatrists IV nurses - AccuNurse http://www.google.com/imgres?imgurl=http://1stprovidersc hoice.com/images/Medical-Voice-Recognition-Software.jpg
  • Slide 11
  • Primary Care Trial at US Army Medical Command in 2009 10,000 copies of voice recognition software Installed 42 healthcare facilities Software tutorial and face-to-face training offered Champions trained Accuracy rated 90% by all participants Not used with patient in exam room, but used immediately after seeing patient by majority of users Hoyt, R., & Yoshihashi, A. (2010, Winter). Lessons learned from implementation of voice recognition for documentation in the military electronic health record system. Perspectives in Health Information Management, 7(Winter). Retrieved from http://www.ncbe.nlm.nih.gov/pmc/articles/PMC2805557/?tool=pubmed.
  • Slide 12
  • Primary Care Clinic from Wellspan Health implemented electronic health records with voice recognition included Voice recognition treated as component of EHR Used in exam room with patient Baker, R.H. (2010, May). Voice recognition assists clinicians. Health Management Technology. Retrieved from http://healthmgttech.com.
  • Slide 13
  • The VA Early trial in late 1990s Cost $2,000 per work station Compare 3 word recognition systems using 12 physicians Evaluation from scripted charting Error rate ranged from 6.6% to 14.6% Estimate current use by 7000 nurses and physicians Devine, E.G., Gaehde, S.A., & Curtis, A.C. (2000, Sept-Oct). Comparative evaluation of three continuous speech recognition software packages in the generation of medical reports. Journal of the American Medical Informatics Association, 7(5), 462-468.
  • Slide 14
  • Psychiatry Health record includes dense narrative In mandatory implementation, providers who do not type notes more inclined to accept voice recognition Providers would not dictate in front of patient Providers found no perceived benefit in speech recognition Half of the evaluators favored the use of speech recognition Derman, Y.D., Arenovich, T, Straus, J. (2010). Speech recognition software and electronic psychiatric progress notes: physicians ratings and preferences. BMC Medical Informatics and Decision Making, 10:44. Retrieved from http://www.biomedcentral.com/1472-6947/10/44.
  • Slide 15
  • IV Nurses Used at Butler Memorial Hospital, Butler, PA Pilot project with 3 IV nurses Lightweight headset and pocket sized wireless device Computer entry of IV needs sent to nurses headset On completion of patient care, nurse uses voice recognition system to record what was done in patients record Receive next order over headset for next patient McGee, Marianne Kolbasuk. (2009, September 17). Voice recognition tools make rounds at hospitals. InformationWeek Healthcare. Retrieved from http://www.informationweek.com/news/healthcare/EMR
  • Slide 16
  • Patient Interactive Voice Response System Automated telephone calls made to patients on day following surgery Patients respond to questions via speech Speech recognition software updates database based on to patients response If response indicates follow-up telephone call by nurse, nurses will be prompted to complete contact System reported to be 97% accurate Foster, AJ; LaBranche, R; McKim, R; Faught, JW; Feasby, TE; Janes-Kelley, S; Shojania, KG; van Walraven, C. (2008). Automated patient assessments after outpatient surgery using an interactive voice response system. The American Journal of Managed Care, 14(7), 429-36.
  • Slide 17
  • Benefits of Speech Recognition Reduction of transcription expense Improved patient care Reduction in time documenting care Increase per patient revenue Allows physician to dictate in their own words Does not add recurring labor costs
  • Slide 18
  • Barriers to Speech Recognition Capital cost of EHR with speech recognition Cost in time (users) Security or confidentiality issues Costs to maintain EHR Interference with doctor-patient relationship Difficulty with learning new technology Lack of tech support Lack of perceived benefit
  • Slide 19
  • Problems with Speech Recognition Accuracy rate approximating 90% requires editing Upgrade of processor speed and/or random access memory may be required Change in method of documenting encounter notes Not all users receiving appropriate training
  • Slide 20
  • Thank You!