I was at an Emergency Medicine Congress earlier this week, attending an all-day workshop on healthcare IT. After a morning spent discussing the advancement of the Electronic Health Record, the afternoon was dedicated to speech recognition technology and the various projects taking place across the province. One doctor was explaining how he was using off-the-shelf Dragon as an enterprise solution, much to the peril of his own productivity, let alone the protection of patient data. Another physician was presenting a facility-wide project involving SpeechMagic being integrated into a dictation/workflow system. By listening to these guys, I got the feeling that Dragon appears to be a great solution for individual doctors (i.e.: small physician offices) willing to spend the time “training” the software, while SpeechMagic clearly sets apart as a professional solution more equipped to fit the needs of large healthcare facilities. Here is why:
On a licence-to-licence basis, Dragon comes out cheaper than SpeechMagic. However, when looking at concurrent pricing, the gap shrinks. Concurrent pricing takes into account the fact that all physicians don’t use the system at the same time, a flexibility reflected in Philips’ pricing model, who offers 2:1 and 3:1 SpeechMagic licences (1 license for 2 or 3 physicians that don’t operate on the same shift). Dragon doesn’t seem to offer the concurrent pricing option to date.
Shared vocabulary / corrections
The question of shared vocabulary was raised at one point; another area where Dragon and SpeechMagic differentiate. I understood that you could share vocabularies and corrections by downloading a special file from the Dragon website, but it wasn’t very clear, while on the SpeechMagic side, corrections from one author are automatically shared with other authors, which significantly accelerates the system’s learning curve. Then again you couldn’t expect much less of a network solution, could you?
All in all, what blew me away is the amount of training required by Dragon. One of the physicians was suggesting other Dragon users to follow the initial dictation exercises offered with the software in order to reach acceptable accuracy: 50 dictations for the physician to do prior to being fully operational! Once again, I doubt that even 5% of the physician population would be willing to put in so much time and efforts, and who can blame them? Another physician in the audience using SpeechMagic explained that his team was operational right away, without having to “train” the software to their individual voice characteristics. So I find funny to see how the whole “initial training myth” is still very much alive – Dragon being the sole reason behind it – when professional technology has eradicated it long ago.
As important as the speech recognition software are the file management and workflow aspects, which are simply non-existent with Dragon. File transfer, electronic signature, distribution: you figure it out. That’s where healthcare facilities need to be careful upon investing: yes speech recognition can be valuable to their staff, as long as it comes along with other tools to ensure that technology adapts the way they work, and not the other way round:
- Workflow management automation: ability to activate speech recognition for certain times/users/work types/departments only, leaving all options open. For instance, a facility can decide that short reports can be reviewed by Authors in front-end mode, while more complex and detailed work can be routed to transcription for correction as a standard or on the fly.
- Electronic signature: ability to assign a legal value to all reports completed by a physician.
- Distribution: ability to fully automate the distribution of final reports as per rules set by the hospital (fax distribution to referring physician, e-mail distribution to nursing staff, etc.)
A standalone speech recognition software cannot deliver all this. And that’s not its “job”. That’s why companies like Philips do not sell SpeechMagic directly, but only through a number of dictation and workflow systems vendors that have worked with the medical field for years and that understand its complexity and unique requirements.
File storage and confidentiality
More distressing is the file storage issue, which directly falls under the HIPAA (US) /PIPEDA (Canada) regulations, both very strict on the subject. One doctor using Dragon explained how all of his completed reports were stored on his laptop. Then what happens if the laptop is stolen? Unless the hospital’s provides dedicated server space, those files are at risk, but then again, it’s up to the hospital to enforce a central storage policy. But is it really their role? I believe it is the technology provider’s to provide a clear and reliable answer to the confidentiality problem. The way a system like Crescendo-SpeechMagic addresses the issue is as follows: files are always kept on a central server at all times, leaving zero footprint of confidential information on users’ PCs. When a user needs to edit or listen to a dictation, voice files are streamed – as opposed to copied – to users’ desktops. If a PC is stolen, it’s a hardware loss, period.
Potatoe-Patatoe. Both products offer a similar choice of medical dictionaries, covering general medicine and a number of specialties, in a wide range of languages. The only question mark I have is regarding the French Canadian language: I know SpeechMagic offers a specific model for French Canadian, while it is not clear to me whether Dragon simply sells its French-from-France model in Quebec.
Well, I think I have covered the main points, which to me, only reflect the development history of both products. SpeechMagic was designed as a network solution while Dragon was originally made for individual users in the consumer world. That’s why a licence-to-licence price doesn’t make any common sense in my opinion, nor adding a “Pro” sticker on a product make it acceptable for professional use. I am even tempted to conclude this SpeechMagic-Dragon comparison thread by saying both products shouldn’t…be compared in the first place (duuhh!)