Cover Story
The role of voice-activated technology in today’s dental practice
GREG L. DREVENSTEDT, Ph.D.,
JENNIFER C. McDONALD and
LINDA W. DREVENSTEDT, M.S.
 |
ABSTRACT
|
|---|
Background. Voice-activated modules are available for most leading clinical software applications. This article briefly charts the evolution of this technology, describes how it is used in dental practices and discusses possible future developments.
Conclusions. Improvements in speech recognition and microphone technology have helped voice-activated charting and clinical note dictation become more accurate, faster and easier to carry out than previously possible. As the technology continues to evolve and is applied more broadly in clinical settings, voice activation is likely to become a mainstream part of dental computer technology.
Practice Implications. Voice activation is a user-friendly and cost-effective way to increase clinical productivity and maintain an innovative, up-to-date practice image.
Key Words: Speech recognition software; computer peripherals; patient records, computerized
In television’s long-running series "Star Trek: The Next Generation," set far in the future, Captain Picard spoke commands for his spaceship to go to warp speed or to request information about an alien life-form, and the Starship Enterprise’s computer responded dutifully. Although much of the technology in "Star Trek" existed solely in the realm of science fiction, the voice-activated technology it presented exists in the real world—and in the dental industry.
This article briefly charts the evolution of voice-activated computer technology, describes how this technology currently is used in dental practices and discusses possible future developments.
Voice activation is a user-friendly and cost-effective way to increase clinical productivity.
|
THE EVOLUTION OF VOICE RECOGNITION SYSTEMS
|
|---|
The earliest computer speech recognition systems were hardware-based. Although these systems provided a promising start, labor intensive "training" of the systems and frustratingly low levels of accuracy hindered their widespread use. Before these speech recognition systems could be used, the system had to be trained painstakingly to the unique characteristics and vocabulary of each user. Such training usually took several hours and required users to recite long lists of terms. Furthermore, because older systems rarely were networked, users had to train the speech recognition systems on each computer they used. Hardware-based systems also had difficulty adapting to temporary changes in a user’s voice—due to nasal congestion, for example—which limited accuracy and sometimes required retraining. Surgical masks, fatigue and stress further affected the users’ voices, confusing the system and reducing accuracy.
Most of today’s speech recognition systems are software-based solutions that are more adaptive than their hardware-based forebears, leading to reductions in training time and increases in accuracy. Speech recognition engine (SRE) software relies on complex mathematical algorithms to assign to each sound an identifier, which enables the software to distinguish voice sounds from background noise. New SREs perform "context" evaluations during the communication process much as humans do. Context evaluations follow a series of steps, according to Kaprielian1: "(1) listen for speech, (2) identify the phonemes [basic units of sound used to distinguish different words], (3) match the phonemes to the words, and (4) make an educated guess as to the context."
The adaptive nature of software-based speech recognition allows most systems to achieve moderate levels of accuracy without prior training of the system. For example, the customer support centers of some large companies now use voice-activated systems that listen and respond to spoken words or phrases rather than the number tones produced by pressing a phone’s keypad. Dental software companies that offer voice-activated clinical modules recommend that users undergo a brief training session to ensure high levels of accuracy. These training sessions take only 10 to 20 minutes and usually involve the more enjoyable experience of reading excerpts from novels rather than reciting vocabulary terms. Once trained, speech recognition systems continue to "learn" by updating the user profile, thus improving accuracy. Furthermore, because user profiles now can be stored on a network, voice-activated systems on all connected computers can be used after a single training session, and multiple user profiles can be stored on a single network.
According to Pam Hemmen, director of the Patterson Technology Center, Patterson Dental Supply (St. Paul, Minn.) (manufacturers of Eagle-Soft), "When voice activation started, accuracy was about 50 to 60 percent, but now it’s up to 85 to 90 percent. That makes a huge difference in effectiveness and efficiency, which are the main reasons why voice systems are used" (P. Hemmen, oral communication, October 2003).
Although nearly 100 percent accuracy can be achieved with low background noise and clearly spoken commands, manufacturers and users of clinical software voice modules say that accuracy of about 90 percent is typical. Innovation in several areas has boosted the accuracy of speech recognition systems. First, improvements in filtration have made it possible to separate different components of sound signals, enabling the isolation of speech and the cancellation of echo. In addition, advancements in speech detection algorithms make it easier for SREs to isolate vowels and identify speech sounds. Finally, many high-quality microphones now have noise-reduction capabilities that transmit a clearer signal to the speech-recognition system.2
|
SPEECH RECOGNITION IN DENTISTRY
|
|---|
Voice-activated systems used in clinical settings are of two types.
- – The predominant type is a command-driven voice system used to record patient data for treatment plans, existing conditions and clinical charts. This type of system allows hands-free navigation of clinical software during patient examinations and serves as a separate input device, whereby a microphone headset replaces a keyboard or mouse.
- – The second, less common type is a dictation system that translates clinical notes directly to text and attaches them to the appropriate patient record.
Voice-activated systems typically are sold as add-on modules to existing clinical software applications. As shown in the table
, seven of the eight leading dental software applications currently offer voice-activated modules for clinical charting and two offer modules for clinical note dictation.3 All of the voice-recognition systems offered by dental software manufacturers rely on third-party SREs, typically either Microsoft Speech Technologies (Microsoft, Redmond, Wash.) or Dragon NaturallySpeaking (ScanSoft, Boston). Because manufacturers of clinical software develop the interface between their application and the SRE rather than the SRE itself, the performance of voice modules in different clinical software systems using the same SRE generally is the same.
|
USING A VOICE-ACTIVATED SYSTEM
|
|---|
The great strides made in speech recognition technology have compressed the learning curve for using voice-activated charting and dictation systems. Following several guidelines can greatly improve accuracy and ease of use.
- – Experience helps. It is useful to have experience and familiarity with the charting system used by a particular clinical software application to avoid having to learn simultaneously how to use the chart and the voice module. Becoming comfortable with a voice-activated system mostly involves learning the system’s commands and the appropriate sequence in which the commands should be spoken. Users also are assisted by on-screen menus that list the commands.
- – Train the system under normal working conditions. Since the system must learn to distinguish speech from noise and adapt to the unique characteristics of a user’s voice, the speech-recognition system should be trained under conditions of normal use, such as when the user is wearing a surgical mask.
- – Boost accuracy with continued system training. Dr. Steven Rinaldi, who uses Eagle-Soft’s voice module in his Andover, Mass., practice, recommends that users of voice-recognition systems spend time training the system over and above what the manufacturer recommends, which should further boost accuracy. "I didn’t want to use the system live and have it not work efficiently. Spending extra training time up front saved me time and aggravation later" (S. Rinaldi, D.M.D., oral communication, April 2004).
- – Clarity counts. The system’s accuracy also depends on clarity, pace and volume of speech. According to Patrick McGibbon of Dentrix Dental Systems (American Fork, Utah), "Any drawback of this technology can be overcome with one simple recommendation: speak clearly and deliberately" (P. McGibbon, oral communication, October 2003). Although regional accents or other variations in speech patterns can be challenging for voice systems, those variations usually can be accommodated through proper training of the system and speaking clearly at a consistent pace and volume.
- – The microphone makes a difference. Invest in a high-quality microphone that reduces extraneous noise to help ensure that the system receives a clear sound signal.
Voice-activated systems pose virtually no threat to maintaining compliance with the privacy regulations stipulated in the Health Insurance Portability and Accountability Act.
Pros and cons.
Clinical software manufacturers and voice module users claim that voice-activated charting and dictation save time and boost productivity. Clinical chart data or notes are entered directly into the computer, thereby saving the clinician the extra step of entering handwritten notes or transcribing a tape at a later time. A reasonable concern, however, is accuracy. Rather than asking for clarification for commands that are not understood, voice systems available today either will guess or will do nothing. However, mistakes can be caught quickly and corrected through the use of a feedback option that repeats the commands into an earpiece after the clinician says them. If the feedback becomes distracting, the volume can be turned down; in some modules, it can be turned off. To fix mistakes, the user simply tells the system to back up, clear the information and input new data. Users also can glance at the computer screen while working or review the chart at a later time to make sure the data are correct.
Voice-activated systems pose virtually no threat to maintaining compliance with the privacy regulations stipulated in the Health Insurance Portability and Accountability Act. Patient records are accessed by navigating through a menu, not by speaking patient names. Clinical information communicated into the system becomes merely "a series of numbers which is meaningless because it can’t be associated with a patient" (L. Greenspan, director of dental development, PracticeWorks, Eastman Kodak, Rochester, N.Y., oral communication, October 2003). Furthermore, since patient data entered by voice activation are stored on the practice’s computer system in the same way as is all other information, there is no need for a specialized backup and recovery system.
|
ADVANTAGES OF USING VOICE ACTIVATION
|
|---|
There are various reasons why a practice would adopt a voice-activated system, the most important of which is to increase productivity. By taking advantage of hands-free navigation of clinical software, dentists and hygienists can examine patients more efficiently and with fewer distractions. Because clinical data are recorded directly into the computer, the voice-activated system eliminates the redundancy involved in taking down handwritten notes and then transferring the information to a computer, as well as the associated potential for data entry error. Since a charting assistant no longer is needed, that staff member can be assigned elsewhere and practice resources can be used more efficiently. Such gains in productivity and efficiency also are cost-effective. Although prices and features vary among dental software systems, voice-activated modules can be purchased for about $1,000.
In addition to potential productivity gains, other reasons for using a voice system include improving patient retention and practice image. By reducing examination time, dentists and hygienists can spend more time talking with patients and discussing their treatment needs. According to Dr. Chris Hammond, who uses Dentrix’s voice-activated charting and dictation in his Provo, Utah, practice, "Technology is often considered impersonal, but in a dental office it can be the opposite. Using a voice system, hygienists can conduct thorough perio exams in just a few minutes, allowing them to spend more time visiting with patients, establishing trust and building relationships" (C. Hammond, D.D.S., oral communication, October 2003).
Using state-of-the-art technology such as speech recognition also can promote an image of the practice as innovative and up-to-date. Dr. Steven Rinaldi uses EagleSoft’s voice recognition system because, he says, "it provides the ‘wow’ experience that other technology such as digital radiography and intraoral cameras also provides. My ultimate goal is to get more patients, and 98 percent of new patients will come from word of mouth. If I can create that ‘wow’ experience and voice recognition contributes to that, then it is very valuable" (S. Rinaldi, D.M.D., oral communication, April 2004).
|
THE FUTURE OF VOICE-ACTIVATED TECHNOLOGY
|
|---|
Although voice-activated systems for clinical software have been available for several years and now are more accurate, faster and easier to use than they were before, they still are not a mainstream part of the technology used in most dental practices today. According to the 2001 Technology Census in Dental Products Report, 94 percent of practices are computerized, yet only 29 percent of practices use computers in the clinical setting.4 Of the practices using computers in the treatment rooms, even fewer are using the voice technology that is available. As speech recognition becomes more mainstream—as have Internet access, e-mail and instant messaging—interest in voice automation in dentistry will increase and drive innovation.
Continued improvements in speech recognition and microphone technology will further increase accuracy and reduce the time required to train voice-activated systems. Voice activation soon will expand beyond charting and taking clinical notes to include image acquisition, intraoral photography and other clinical applications. Furthermore, systems soon will become wireless, making them even easier to use.
Although these particular innovations are on the horizon, others see the future of voice more broadly. Dr. Arlen Lackey, who has used Dentrix’s voice-activated charting in his Pacific Grove, Calif., practice since 1992, asks, "Why are we still using a keyboard or a mouse? Why isn’t everything voice[-driven]? Voice technology is the future of computer technology—it needs to leapfrog the logjam of yesterday’s way of entering information into a computer" (A. Lackey, D.D.S., oral communication, October 2003).
|
CONCLUSION
|
|---|
The dental practice of the future may resemble the science fiction realm of Captain Picard, with an entire dental treatment room controlled by voice commands. There certainly will be some technological similarities. Dentists, hygienists and assistants will instruct the computer system to adjust lights, change chair positions (or possibly turn on heat or vibration), take and store digital radiographs, request a patient education module, retrieve before-and-after digital photographs, play specific music in the room, and check out and schedule the patient—all hands-free and wireless.
|
FOOTNOTES
|
|---|
Dr. Drevenstedt is co-founder and partner of Valmont Research, 5721 Magazine St., #122, New Orleans, La. 70115, e-mail "greg{at}valmontresearch.com". Address reprint requests to Dr. Drevenstedt.
Ms. McDonald is president of McDonald Harbor Consulting, Suwanee, Ga.
Ms. Drevenstedt is president of Drevenstedt Consulting, St. Augustine, Fla.
The authors gratefully acknowledge the time and contribution of several people who agreed to be interviewed for the research presented in this article: Chris Hammond, D.D.S.; Arlen D. Lackey, D.D.S.; Steven J. Rinaldi, D.M.D.; Pam Hemmen and Carrie Koenig of Patterson Dental Supply; Patrick McGibbon of Dentrix Dental Systems; and Larry Greenspan, Steve Mallot and Bonnie Pugh of PracticeWorks. They also acknowledge the editorial assistance of Adrienne Lamb and Theodore Mathews.
|
REFERENCES
|
|---|
- Kaprielian S. Speech recognition in practice. Microsoft Developer network webcast. Available at: "www.microsoft.com/usa/webcasts/ondemand/2366.asp". Accessed Sept. 24, 2003.
- Terry M. How input devices enhance speech. Speech Technol Magazine Nov/Dec 2002:22–5.
- Office management software—user survey 2003. CRA Newsletter 2003;27(2):1–3.
- Goff S, Kehoe B. 2001 technology census. Dent Products Rep 2001;35(6):18–25.
TOP
ABSTRACT
THE EVOLUTION OF VOICE...
