For those of you who aren’t data geeks, feel free to browse over to buzzfeed and figure out which Ghostbusters character you are.
For those of you still with me, I’m sure you share my passion for data standards. I’ve been in the trenches for years dealing with data quality issues and data that looks similar but is semantically different. As mentioned in my previous post on interoperability, we still have a long way to go in the area of health information technology (HIT). In that post, I mentioned that ONC (and its parent organization, HHS) has recently stated more publicly efforts for driving interoperability. There are some general principles, but a March 2014 GAO report was critical of the work over the last year: “…While HHS officials intend these principles to lead to future actions that have the potential to address key challenges, the HHS strategy does not specify any such actions, how any actions should be prioritized, what milestones the actions need to achieve, or when milestones need to be accomplished…” – Ouch!
A huge barrier to interoperability and, by extension, health information exchange (HIE) is data standards. That is, if something is syntactically equivalent but semantically different, the receiving entity will interpret the data incorrectly…and as everyone can guess, getting health data wrong can be very bad for you and I. I believe there are two problems to this: 1. it’s both new and difficult enough that few want to dive into it; and 2. there’s no money in it.
Let’s first look at the new and difficult part of this: I truly feel for everyone in healthcare IT over the last few years. Not only has everyone had to go through a lot of system implementations, there’s meaningful use attestation to do, and a conversion to ICD-10…and then the ICD-10 implementation date was delayed. Regardless of whether anyone thinks that’s a good idea or not, the fact is that IT professionals (you, that is) will still have to deal with ICD-10 implementation or testing or regression testing (at the least) for a long time to come. Given all the initiatives that HIT professionals have to deal with, it’s not surprising that data standards haven’t risen to the top. In most of the country, HIE is a luxury and the norm. As healthcare organizations are forced (through meaningful use) to engage in HIE, they are finding that data standards aren’t as well defined as they need to be.
While there are standards (e.g., LOINC, CPT, SNOMED, NDC, HL7), they aren’t cohesive enough to be a “plug-and-play” type of activity. That is, one might think that by representing a lab test and result with LOINC in an HL7 message, any system in the world would be able to interpret it correctly. Unfortunately, the standards are somewhat loose and were not develop in coordination with one another. Therefore it takes time and effort (aka: money) to make systems interoperable even with existing standards. At the present time, it’s painful enough to put the frustration front and center in providers’ minds. In that same March 2014 GAO report, this was called out: “…many providers told us that the cost of developing, implementing, and maintaining interfaces with others to exchange health information is a significant barrier….”
This then leads to the second point that there’s no financial incentive to work on data standards and HIE. In an August 2013 strategy report, HHS does generally speak to pushing this through a variety of ways; but currently there’s only “the stick” aspect of meaningful use if an organization doesn’t get on-board with health information exchange. The EHR incentives will not be enough for interface or other related HIE expenses. The EHR incentive – as it was intended – will enable providers to recoup some of the cost for implementing the EHR, not other costs. Additionally, providers interviewed in the GAO report indicated that many times the benefits do not yet outweigh the costs. While there has been an incentive program for EHR adoption, there is no incentive program for creating interfaces or doing significant health information exchange.
There is a desire for real HIE, though. Consider the EHR certification: first was the requirement to support Direct. However, that is much the same as secure e-mail and doesn’t add much value in the whole health information exchange scheme of things. In fact, the American Hospital Association politely reinforced this idea with the GAO: “…The AHA noted that it is appropriate to require certified EHRs to support both the Direct Protocol and other standards that would allow for more robust exchange so providers can choose which standard is appropriate for their needs when exchanging health information….” and then “…several providers and stakeholders commented that the Direct Protocol allows for limited exchange, such as exchanging a secure email message, rather than enabling certain other functionalities, such as the ability to query another EHR system…”. Fortunately, the 2014 EHR certification (summary here) does include more than just Direct and calls out more standards like LOINC for test results.
Now, this post isn’t just identifying problems. Let’s explore an idea for a solution. What I propose is a phased approach to ramp up organizations and providers that don’t have ready access to established HIE products like we have in Indiana. That is, this assumes a lower HIE maturity level of the community.
First, organizations and providers need to start to work with CCD’s either for transition of care or more general medical history purposes. CCD’s are being used for this purpose and can add much value quickly. Through use of a certified EHR or using the open-source Connect software, these can be implemented fairly readily and provides a much needed, standard vehicle for patient history documents. By starting with this, it will require both the sending and receiving parties to start to focus on what data is being sent and how. In my experience, this level of review/involvement by the parties that benefit is needed as there isn’t always a deep understanding of the data being produced by EHR and other systems. In just one example from my experience, a provider was sending lab result date that was prior to the order date. It was due to a slight disconnect in systems and wasn’t immediately caught until a more detailed review was done. For any significant progress to be made, many organizations will need to “roll up their sleeves” and become more familiar with the specific data their systems are generating and work through the gaps. I also propose that this be supported by a payment model or other incentive from CMS to drive this. With so many other initiatives going on, only compelling (aka: money) reasons will drive new work…or at least, make it move at a good pace.
Once many organizations are starting to produce and share CCD’s, then real-time, on-going interfaces should be tackled. Instead of sending a summary for an individual patient, it is much more efficient and effective to have a steady stream of data available to any and all providers treating a patient (in a secure manner, of course). So this will involve the ramp up of third party Health Information Exchanges (the noun) like Indiana Health Information Exchange and Thrive HDS to enable real-time health information exchange (the verb). As noted above, providers are pointing out that interfaces are expensive to build and maintain. To drive this part, CMS should also provide some incentives. Currently, CMS has “advertised” to states that the Federal financial participation (FFP) or Federal medical assistance percentage (FMAP) is 90/10 for HIE efforts. That is, CMS will pay 90% of the cost – not a bad deal at all! Having the previous experience of working through data issues through the use of CCD’s, the funding will have more bang for the buck…and be less expensive overall…AND will have “raised all boats” in data quality as each organization has had to work on their own data. The willingness and benefits of CMS funding can already be seen in this recent example from Philadelphia.
If, at the same time, ONC can work on improving standards overall through their standards and interoperability committee, we’ll actually see real health information exchange take root. Otherwise, increased health information exchange outside of an integrated system will be painfully slow and very spotty due to the fact that will be secondary to other initiatives.