EMR and Vista speech

I'm going to look into this. The native Vista Speech appears to work very well. If it had a way to easily import a medical vocabulary, it would reach new limits.

We have eClinicalworks, and I have been using Vista speech recognition for a few months.  Works reasonably well, definitely faster than typing for the most part and definitely cheaper than dictating.  I usually dictate the HPI and the treatment plan, type most of the other stuff, put in a physical exam template (I am a subspecialist and tend to use the same general exam frequently).  I am pleased with the setup, training wasn't onerous.  It still occasionally drives me crazy, of course, but it's not bad.

Issues:  You need to be in a quiet space for it to work.  Ambient noise is a significant issue.
There are a few words that I use frequently that it refuses to learn, so I've learned to just type those (posterior = post your ear).  It locks up once in a while, but usually doesn't cause a loss of data.

Overall, a good experience. 

wes

Wes7373 wrote: We have eClinicalworks, and I have been using Vista speech recognition for a few months.  Works reasonably well, definitely faster than typing for the most part and definitely cheaper than dictating.  I usually dictate the HPI and the treatment plan, type most of the other stuff, put in a physical exam template (I am a subspecialist and tend to use the same general exam frequently).  I am pleased with the setup, training wasn't onerous.  It still occasionally drives me crazy, of course, but it's not bad.

mmarkoe wrote: cmwilkerson wrote: I have used the WSRToolkit to import medical charts and exams in MS Doc formats using, "Add From File." . I also took a medical vocabulary list of thousands of words, converted it to a doc file and imported it, using the toolkits, "Add From File." The Add From File description (right on the page) specifically says it is not meant for adding lists of words. Its purpose is to recognize a specific style of speaking based on the context of how you use words. It will update your speech model to reflect your style and usage and not add specific words.   I also tried to add a list of words, left formatted one word per line, to the Add to Dictionary section of the toolkit. This was not workable for me,  as I pasted hundreds of words, and the program wanted me to deal with and verify each entry, one at a time. I could not get Windows Speech Recognition to recognize, "a bad" as in a bad microphone and sound card. It constantly came up with, "a bed." Just adding it to the dictionary one time along with its pronunciation solved the problem.   The description for the Add to Dictionary feature describes its purpose for improving the recognition of difficult to understand words. It does this by not only adding the word (it helps if you use the word within a short phrase) but also record a pronunciation associated with it.   You can add hundreds of words, but as you found out, this will work best if you add a pronunciation for each one.  Most words, even medical words are already in the speech dictionary. Usually, correcting a word one or two times in context will mean it will be recognized from then on.   I've done the above, at home, on a 2.8 Imac with 2 gigs of ram,  dual booting to Vista. I'm using the iMacs built in mic but will try a better mic.  (I have a 3.2 quad 4 at work with 4 gigs of ram that is better suited for this but I have more time at home to experiment). There is certainly nothing shabby about your system specifications. However, and accurate/noise canceling microphone will mean that the software will understand your words and commands with much higher frequency than a built-in microphone or commodity microphone purchased in a retail store.

I have discovered that my EMR at work, (www.writepad.com) does not like Vista Voice in my charting section. Vista Voice will pop up with numbered choices ONLY in my EMR? DNS Medical Pro 9.5 works fine in this same application, and Vista Voice works in other applications. Not sure what that is all about but presently this prohibits my using Vista Voice in my EMR, which is my main goal.[/quote]

cmwilkerson wrote: I'm sitting here with DNS 9.5 enabled and dictating supplemental text into my EMR chart. When I turn off DNS and switch to Vista Voice, dictating the exact same words (or any words) results in the Alternates Panel showing up and giving me a choice of numbered answers. This does not occur in Notepad or Word, with Vista VR. Both of these applications work just fine and the dicataion below is perfect. I'm at a loss. I assume this Alternates Panel ONLY appears when recognition is failing?   Here is the exact text I am dictating: "The patient presents today with a follow up examination, with respect to a diagnosis of Cervical radiculitis." The Alternates Panel pops up even with dicating only, "The patient presents."

I may have an idea about bypassing the Alternates Panel while you're dictating into your EMR chart.  The concept is for a speech macro to determine when the EMR chart application is active then do a little fancy footwork (via <sendKeys>) to bypass the Alternates Panel.

First we'll need to determine what is the exact name of your EMR chart application.  If you're not sure, with it loaded, go to Windows Task Manager (CTRL-SHIFT-ESC).  On the 'Processes' tab you should find the exact .EXE name of your app.  Let me know this and I'll craft a simple macro for you to try.

Brad

cmwilkerson wrote: Hi BRad,   I'm open for any ideas.   My EMR executable is writepad.exe.

Well darn.  After having played around with what I thought would work, it's turning out to be quite a kludgy workaround (You have to preface every group of words with the phrase "EMR dictate".    Guess it's time for Plan B.  But anyway, here's the macro... 

You'll first need to make sure you have WSRMacros loaded and running (see other post) then copy/paste this macro into notepad then save it in your Speech Macros directory.

<speechMacros>
<command>
<appIsInForeground processName="writepad.exe"/>
<listenFor>E M R dictate [...]</listenFor>
<sendKeys>{[...]}{250 WAIT}{u+0020}</sendKeys>
</command>
</speechMacros>

Brad

cmwilkerson wrote: Notepad? Don't you mean the Macro Editor of the Toolkit, which autosaves in appropriate directory, and proper extension .wsrmac), which in my case appears to be C:\Users\Chris\Documents\MacroBuilder

cmwilkerson wrote: Brad,   How do I know my macros are working?   Verbalizing, "EMR dictate," does nothing but open the Alternatives panel.   As you probably remember, there is a question whether my macros are working at all given the error in opening the WSRTookit and my installed path of C:\Users\Chris\Documents\MacroBuilder

At this point I don't believe macros are working for you... and it sounds like WSRMacros hasn't actually installed.  Here are some things to check:

Is there a WSRMacros.exe file on your HD?  Assuming you have 32-bit Vista (not 64-bit) it should be in this folder: "C:\Program Files\WSRMacros".

If you do have that folder and .exe file, does it put the WSRMacros icon in the system tray when you run it?  The WSRMacros icon looks like this:






Edited by BradT - 10/Oct/2008 at 3:44pm

cmwilkerson wrote: Brad,   I found the problem and it is with me. I assumed, erroneously that when I started your WSRToolkit that the macros would be loaded. Not the case. So now I start the Windows macros program separately from the run command.   Having said that, I am still at a loss for how to start your EMR dictate macro which I made in the Toolkit??

The WSRToolkit will automatically run WSRMacros.exe if it's in the C:\Program Files\WSRMacros directory.  I'm still not convinced that WSRMacros is running on your PC.  There are two folders involved in a successful install of WSRMacros.  The executable file will reside in C:\Programs Files\WSRMacros and all macros must reside in C:\Users\Chris\Documents\Speech Macros.  Without the Speech Macros directory, WSRMacros won't function properly. 

When it runs, WSRToolkit looks for  this Speech Macros directory.  If it doesn't exist, WSRToolkit creates a "safe" directory C:\Users\Chris\Documents\MacroBuilder to hold macros created before WSRMacros is installed.  Once WSRMacros is installed, you have to manually copy the macros in the MacroBuilder folder  to the Speech Macros directory. 

The mystery is, where is the Speech Macros directory on your PC? 

If you're certain WSRMacros is running, create the folder C:\Users\Documents\Speech Macros and copy the EMR Dictate macro to that folder.  Then try opening your EMR app (writepad) and say "E M R dictate this is a test".

Brad

cmwilkerson wrote: Hi BRad,   I'm open for any ideas.   My EMR executable is writepad.exe.

Another thing to try:

How about having notepad open along side of writepad, dictate into notepad.  We can craft a simple macro that will allow you to say "copy to EMR" (or whatever you'd like to say) and it will automatically copy everything in notepad, make writepad the active application and paste it at the current cursor location. 

Brad

cmwilkerson wrote: Let's just make a macro that works.

Aw, come on, where's your sense of adventure! 

cmwilkerson wrote: Now I am getting another error when I open the windows macros: "no signature was present in the subject."   This error appears to be related to the macro you made for me as when I remove the macro, no error. When I place it back, the error occurs.

Now we're getting somewhere!  While experimenting (and tweaking) your macros, it's recommended to set macro security to 'Low'.  Right-click the WSRMacros icon in the system tray and choose 'Security -> Set Security Level -> Low'.  That will get rid of the error messages for now.  Once things settle down a bit, you can always go back and resign your macros if worried about malicious programs.

Brad