Another step forward for data liberation; Check out the SMART C-CDA Scorecard!
My job at HealthVault is to be a pragmatist. If somebody has information they want to get into HealthVault, my #1 job is to get it there --- no matter what format it’s in. That means we accept hand-entered data, scans, uploaded files, CDs and DVDS, Blue Button files, structured CCR and CCD documents, device formats, you name it.
But while just “taking” the data is a good first step --- the real magic only starts to happen when that data comes in as some machine-readable format, with codes that allow HealthVault applications to actually interpret the information and do great things with it.
Take for example a medication list. Having a list of medication names is great --- in an emergency or when reviewing with my doctor, just being able to SEE the list puts me light-years ahead. But what if I want to use an application that tells me if there are dangerous drug-to-drug interactions in my regimen? In order to do that … the medications have to be “coded” (assigned a number vs. just a name, like an ISBN number on a book) so that the application can look them up in a database of interactions.
And boy does the medical industry stink at reliable coding. We do a TON of work to try to clean up incoming data as best we can, but it’s a hard job. Then again, maybe I shouldn’t complain because a huge part of the value that HealthVault provides to applications is that we try to make it simpler to deal with the variations in data quality they otherwise have to face on their own. Job security for me! :)
The Meaningful Use program is doing a great job at pushing the industry towards more codes and standardization of data … which is really exciting. The “CCDs” we see coming into HealthVault today are WAY better than they’ve ever been. But that’s not saying much, and there’s still a ton of work to do.
Which is why I’m excited about the new SMART C-CDA Scorecard tool that Josh Mandel and his team (yes, these guys) have created. It’s a simple web-based form that does a bunch of semantic checks on these documents. For example, not only checking if codes exist in a document, but making sure they are good. “Good” means a lot of things --- the codes actually match the text in the documents, they aren’t out of date, they make sense for the section of the document they’re in, and so on.
This is a great step forward --- if all the EMRs out there would simply use this tool before checking in their Meaningful Use Stage 2 updates, it would make for a HUGE advancement in data interoperability. And better interoperability, very specifically, means safer families --- in HealthVault and beyond.
Kudos to the SMART team for the great work!