ABOUT SPEECH RECOGNITION SOLUTIONS
What does Speech Recognition Solutions provide that I can’t do on my own?
Although some practices, particularly those with an experienced “champion” in the use of speech recognition software, have been able to take the plunge without assistance, this is often fairly riskly and is associated with a high risk of failure. Why? Speech Recognition Solutions offers the following:
- Assessment of your practice logistics, workflow and current EMR
- Assessment of your current hardware and its suitability for use with Dragon; if necessary we will make recommendation for hardware procurement
- Assistance in the installation of the software
- Assistance in configuring your software
- Initial user training (typically 2 hours per user – done either in person or during a personalized web based training session)
- Provision of a starter command/macro set to assist in provider efficiencyAfter sales support
What does Speech Recognition Solutions offer over another vendors?
We consider our primary advantage the high level of medical involvement of everyone involved in our company. Speech Recognition Solutions is physician owned and managed. Our training division is headed by a woman with years of nursing experience. In addition to the medical orientation of our company, we are a family owned and managed business. We take great pride in providing honest advice, fair prices, and excellent support. Our goal is to be rewarded for providing value to our customers. Absent the providing value, we are simply selling software and this is not our goal.
What support will Speech Recognition Solutions provide after the sale?
In general we will contract to assist in setting up the software for users, provide assistance in constructing an initial command set for use in your EMR (No Integration) , and provide initial (in depth) training for your users. We are happy to field questions after this time. For more extensive trouble-shooting, additional user training, and other requests, we will offer our technical and training services on a for-pay hourly basis.
GENERAL QUESTIONS ABOUT INSTALLATION AND USE OF DRAGON
What does speech recognition software do?
In the simplest terms, speech recognition software allows your voice to substitutes for keyboard entry on a personal computer. In so doing, it allows for entry of text in a word processing document or text box within an Electronic Medical Record (EMR) / HIS / HER /PACS and navigation on your computer simply by speaking into a microphone. In addition, speech recognition software has functionality which allows multi-step processes on the personal computer to be initiated with a single spoken command.
What are the hardware requirements?
In general, most recent vintage Windows-based computers support the use of speech recognition software. Specific requirements appear in the table below. Although a minimum of 3 GB of RAM (random access memory) is required, 4 GB or more is generally preferred.
Will speech recognition software work on a Macintosh computer?
If you are determined to run Dragon on a Mac, it is possible to do so in 2 ways if you have one of the new Intel Macs.
- Use boot camp to set up a hard drive partition on which you can install the Windows operating system (XP or Vista) and install Dragon on this partition. Dragon will work fine. The only problem with this arrangement is that you will be running the Windows operating system and not have access to your apple software and functionality
- You can use either Parallels or Fusion to run a Windows partition on your Mac “virtually”. This allows to you launch the Windows OS from within the Mac OS. The only issue with this arrangement is that there are occasionally some issues porting the sound from the Apple hardware to the Windows OS.
Will a faster computer provide better accuracy?
No. A faster computer will lead to your words showing up faster and tends to make the process a bit quicker and more enjoyable, but processor speed does not directly influence accuracy.
How much training is necessary to get the software working at a high level of accuracy?
Once the software is installed (a 20 minute process), you must create a new “user profile” by teaching dragon how you talk. You are required to read a 5-6 minute training text. At this point Dragon Medical will be 95% accurate with good enunciation. Higher levels of accuracy (99-100%) can be achieved as you use Dragon Medical and correct the occasional error that occurs.
What does the user need to do in order to achieve high levels of accuracy with Dragon Medical speech recognition software?
Many factors influence accuracy, including the quality of the computer sound card, the microphone quality, and the presence (or absence of contaminating external noise). By far the most important determinants of accuracy are dictation style and careful enunciation.
- Dictation Style: The type of dictation that works best is smoothly flowing words grouped in either phrases or complete sentences. It is helpful, therefore, to think about what you plan to say before dictating with speech recognition software.
- Enunciation: Careful enunciation is extremely important and requires that you speak a bit more distinctly than you might in casual conversation. Slurring of words and mumbling type of speaking will tend to reduce your accuracy. It is not necessary, however, to speak loudly or harshly. Copying the type of enunciation used by newscasters will generally lead to the best results.
How many software licenses will be needed in my office if there will be more than one user?
The Dragon End User Agreement requires that each user have a separate license.
What are Dragon commands? How will this help my efficiency?
The ability to create commands or macros is a functionality that can greatly extend the productivity of Dragon users that use repetitive language or perform repetitive actions on the PC in the course of a usual day. Here are some examples:
- Text commands: These allow you to insert a specified phrase, sentence, paragraph, image or other data into a text document by speaking a brief pre-specified word or phrase. As an example, if you customarily document the counseling discussion before performing a common procedure in the same way each time, you can make a command out of the procedure documentation and insert it any time you desire (with any patient) by simply speaking the command name. These commands take literally seconds to create and forever available for you to use. Such commands can be easily shared from one user to another.
- Advanced commands: these require a bit of programming knowledge and allow you to program Dragon to perform more complex and sequential actions upon a voice command.
What is the importance of vocabularies in using Dragon?
Vocabularies are extremely important to obtaining high levels of accuracy in two respects. First, it is critical that you choose the right baseline vocabulary (one appropriate to your medical specialty) when setting up your first user profile. This will serve as the base vocabulary from which Dragon will select words as it hears you talk. Second, it is extremely helpful to add personalized words to your vocabulary. Personalized words might include proper nouns that are probably not in the Dragon vocabulary such as the names of local hospitals, names of referring physicians (or your partners) and specialized words pertaining to your office or hospital. New words can be added to your user profile one word at a time using the “Words” menu, or can also be added as long lists of words in a .txt document. Like commands, specialized word can be shared from user to user and added to a user’s profile in a single step.
USING DRAGON WITH AN ELECTRONIC MEDICAL RECORD /HOSPITAL INFORMATION SYSTEMS
Will Dragon work with all electronic health records?
Dragon will work in most common word processing documents, including WordPad, Microsoft Word, WordPerfect, etc. Each EHR uses a specific embedded word processing program for text entry and Dragon tends to work well with each of these. There may, however, be some variation in the functionality of Dragon from one EMR/HIS/PACS to another. The makers of Dragon are constantly working to assure the integration of their product, particularly the medical version of Dragon, with the extensive selection of EMRs/HIS/PACS being utilized today. A current list of EMRs that are specifically certified for use with Dragon is shown in an attached document.
Is it necessary to use the Medical version of Dragon with an electronic health record?
In general, yes. For a number of reasons, Nuance, the maker of NaturallySpeaking, has arranged that non-medical versions become essentially inoperable when working in most common EMRs /HIS/PACS.
What special functionalities does Dragon medical the users?
The medical version of Dragon includes a variety of unique features that make it highly supportive of physician workflow.
- Dedicated medical vocabularies: Dragon 11 Medical includes nearly 80 specialized vocabularies. These are listed in the appendix.
- Medical formatting options: Examples include the ability to automatically expand medical abbreviations (CHF > congestive heart failure; b.i.d. > twice daily), automatic formatting of certain key terms (no code > NO CODE)
- HIPPA compliance
- Inclusion of “Dragon Templates” which allow quick navigation through a note in progress
Does Dragon include specific commands for use within my EMR?
Most of the native commands built into Dragon will work in any word processing environment, including the text entry windows within an EMR. Additional functionality which is fairly universal, such as the ability to click on a button, close a window, or open and select something from a navigation menu, will also work within an EMR. There may be special functions which you might want to accomplish within an EMR and which are not built into Dragon. These are generally easy to create and can be done so by Speech Recognition Solutions or a speech recognition champion within your practice.
How long will it take to recoup the costs of implementing speech recognition software in my office?
Obviously this can be highly variable and depends on whether your providers are currently using paid transcription or direct data entry. Likewise, if your users are currently hand-typing, the ROI will depend upon how you value the time, productivity and reduced billing from inefficient documentation. The following table shows ROI calculations for a pilot study of speech recognition software at Dartmouth-HitchcockMedicalCenter and shows the time to break even with speech recognition software compared with varying levels of annual paid transcription.
| Annual Transcription Cost per steno
|| Year One Break Even Point for Speech Recognition
| INR 150000
|| 7 Months after that saving per months Rs.12500/