Voice Recognition

Its Upside and Its Downside

Reliable Clinical Documentation Needs Clinical Narrative

Voice recognition technology has made great strides over the years.  We all have come to rely on it with our own mobile devices.  This is no less true in the medical field where physicians are being coerced into incorporating it in their clinical documentation of patient encounters.  If you and/or your practice find yourself facing this choice, first consider what many physicians have come to learn as to the upside and downside of this growing technology. 

UPSIDE:  The voice recognition system has the ability to learn your particular pattern of speech.  You will have a “written” draft instantaneously available to you on the screen, and you will find the draft will actually contain 100% spelling accuracy. 

DOWNSIDE:  Any tendency on your part to use less than perfect diction will greatly impair the ability for accurate voice recognition.  The presence of any background noise could also present a challenge.  Understand as well VR may generate up to 20% of incorrectly recognized words and has particular problems with deciphering homonyms.  In addition, as your system transforms the spoken word into a formal record, it cannot correct grammar.  So while an instantaneous draft is produced while speaking, time and effort will be required to edit the draft into an acceptable record of fact.

Voice recognition - an amazing technology, but with limitations.
Transcription - a time-proven method providing quick and reliably accurate results.  
David Pearson in an article for CMIO quotes Dr. Elizabeth Todd as reporting: "Yesterday, someone sent me a letter about an amputee patient he sent to a podiatrist. He got a report back on both the patient’s feet. This patient only has one foot."  He quotes another physician's concern:  "I have a sinking heart every time a specialist gets on the EMR because I realize I won’t have any new information about how to manage my patient; it will just be cut-and-paste."* Aside from the lack of any new information, this type of input into the system has the potential for propagating incorrect patient information many times over.

In addition, there is an issue of correct yet incomplete information through the use of such a documentation style.  The entire clinical narrative can be lost.  A case in point, a retired physician himself, Charles Battig states:  "My answers to my own physician's questions were inputted to the laptop via a stylus touchpad.  No typing skills necessary for the keyboard-challenged.  One can only hope that the pre-programmed list of choices accurately reflected my potential responses.  Chest discomfort?  Would that be on exertion?  How much exertion?  Was that discomfort actually ribcage discomfort, as in a pesky costochondral rib joint, and not heart-related?"^

In addition, Dr. Todd in her interview with Mr. Pearson also observed that as many as 30 to 50 percent of patients discuss with their primary care physicians symptoms that are directly related to the stress in their lives or worsened by such stress.  This is where the clinical narrative fills in what EMR documentation options leave off.
While the EMR is a permanent fixture in the foreseeable future, many physicians readily see the need to be able to incorporate the entire account of the medical encounter.  The value of dictation and transcription is again being proven as the overall most reliable clinical documentation method and now is being used in tandem with electronic medical records.

* EMRs are intrusive, encourage false patient info CMIO-June 25, 2012
ObamaCare and Laptop Medicine American Thinker-September 24, 2012