I spoke at length today with the nice folks from Nuance Healthcare and John Vasicek, Senior Engineering Director, sold me on their new Speech SDK (software development kit). Called “Speech Anywhere,” the SDK is part of the new Nuance Healthcare Development Platform which will hit the street by the end of the month. I’ve been a long-time user of Dragon Naturally Speaking and I’m continuously amazed as to how much better their recognition is for medical / clinical speech vs. other tools. It’s still not quite as good a very experienced medical transcriptionists but of course the technology is much cheaper than the human counterparts.
The new cloud-based Speech Anywhere SDK easily enables speech on mobile phones, workstations, iPads / tablets, and web applications. As a developer that has to deliver solutions regularly, I think they have done their homework and are trying to automate as much of the integration as possible. What I like about the design is that it’s cloud-based; this means that the EHR or clinical app you’re building doesn’t need a separate local installation of speech recognition software: it’s all done in the cloud through a service call. I will be giving the software a try later this month but it sounds promising because there’s very little code that needs to be injected to enable speech recognition. It almost sounds too good to be true but of course a good API design can be like magic.
The tech seems nice and the business models seem flexible enough: there are two pricing models — the direct billing model and the reseller model. Direct is where the application developer simply enables the functionality in their software and then Nuance charges customers directly for the number of “minutes used” (like a cell phone provider). Each minute used for recognition is charged a fee (of course minutes can be bundled or charged at tiers where the more you use the less you pay). The reseller model allows the software developer to package the functionality into their app, charge their customers, and then pay Nuance on their own (instead of making the customers pay Nuance directly).
Both seems to have their merits but I like that they don’t seem greedy and are making it easy for developers to get started. There’s a free 90 day trial that will be coming out followed by fee-based support as well as enterprise support for the SDK.
So, is speech enabling your clinical app or EHR in a day really possible? If you take Nuance’s word for it — yes. And, the story seems believable so it’s worth a shot.