| Home | Free Registration | Resources | RSS |
|
|
| The Health Record Review by Patty Enrado |
Health IT vendor survey captures physician concerns
Posted on Thu, Dec 24, 2009 - 04:45 amNuance, provider of clinical documentation and communication products to healthcare organizations, released a survey it conducted with nearly 1,000 physicians. It is somewhat self-serving - more on that later - but it does highlight issues to which EHR vendors need to pay attention.
Nuance is advocating for EHRs that aggregate structured and narrative data. That stance benefits the company because their communications product would produce narrative data. If the survey sponsor had been a third-party entity, the results would have had more credibility. I'm not saying that the results shouldn't be taken seriously. I actually believe that the results are valid. It's just too bad the entity conducting the survey wasn't objective.
What's worse, the survey showed respondents two different physician notes on the same patient. One was with a template-driven EHR and the other using Nuance's own speech-recognition product. Ninety-seven percent of the physicians surveyed chose the document produced with the speech-recognition product as the note they would "consider more valuable in treating this patient." The respondents then checked off other clinical communication objectives met by the speech recognition-generated choice. It might have been better had the survey used another vendor's product or at least not identified the speech-recognition tool. The idea, after all, is to prove that speech recognition is to difference.
Try to look past that part of the survey because Nuance's assertion that template-driven EHRs don't capture a physician's thoughts at the point of care is something the industry needs to address. While some EHRs allow additional documentation to be captured, admittedly the process adds to the physician's workload and disrupts their workflow. EHRs that can seamlessly capture additional clinical narrative documentation through an integrated speech-recognition tool would be an ideal solution. It's a good idea and maybe many EHR vendors are already integrating the capability into their product. If not, maybe they should to capture a larger piece of the market.
The Nuance senior vice president was quoted as saying that the Department of Health and Human Services should require that electronic patient notes not be limited to templated text and structured data elements alone. Making it a requirement by HHS might be a bit much to ask for. What makes sense is working with the EHR certification bodies. If the survey results are on to something, and I think they are, the market will drive EHR vendors to incorporate multiple types of clinical documentation.
- Recent blog entries
- Login or register to post comments






Comments
As a EM physician having used Nuance's DNS with an EMR (which I believe is the "straw man" used in that survey) I believe that with respect to providing a quality narrative, the templated EMR HPI is inferior to the spoken narrative. Overall, Voice Recognition software (VRS) is an improvement, but despite industry claims is not nearly as as accurate as is claimed. All narratives produced by VRS, at least within the noisy Emergency Department, still require careful and time-consuming editing. Otherwise the VRS Narrative can prove inferior to template version, which, while stilted, provides at least some basic elements, while the VRS version can be structurally sound nonsense. Please note however that templated EMRs can't possibly anticipapte the idiosyncratic nature of the Medical Decision Making (MDM) narrative, except for the most minor of cases, and so here there must be some means of capturing the spoken narrative (or type -- yikes)
The ultimate EMR would follow a facilitated dictation model: utilize templated format for the PE, ROS and lab/diagnostics, while capturing a voice file for narrative elements: HPI & MDM. This voice file would be retained within the EMR, available via mouse-click for immediate review, until transcribed either entirely by a human or by voice-recognition software WITH subsequent verification of errors by a transcriptionist.
Is any vendor out there listening?
You've said it all! Great point! I must add that I really don't see VRS as a "mainstream" product in the near future, once "understanding the words" is just a very small part of the "extracting the full meaning" process. Human language is complex and unstructured, and a lot of research and development is still necessary. In the future, VRS might play a main role, of course, but a specialised review by a knowledgeable person will remain a must for some years. And, by the way, vendors are listening. But don't forget they are vendors, Ok?
regarding: "I really don't see VRS as a "mainstream" product in the near future, once "understanding the words" is just a very small part of the "extracting the full meaning" process...."
I don't follow.
And regarding my plea to a vendor, I'm just looking for one to hire me as a consultant, so I don't feel obligated to work so frequently with my present EMR.
I just meant that spoken language is always richer and more complex than the written version of it, even with the words being the same: there are tones, loud and soft voices, face expressions, body language, gestures and a myriad of details a piece of software would have a very hard time to capture in full. That's why I believe VRS systems still have a long road to go by. BTW, I wish you success in your intended consultant career and a good 2010.
Same to you. Happy 2010.