Interoperability

“I meant close the fridge, not the bridge,” a prominent comedian said.

Interoperability standards picked as key to the greatest leap forward. Why? 

Recently, 240 participants in a KLAS interoperability study overwhelmingly picked standards as key to the advancement of interoperability. Could that be interpreted as a need to develop more standards before real progress will be made? Thankfully participants’ voices fine-tuned the clamor around standards, noting that unclear standards are the greatest hurdle to sharing. Virtually no respondents reported a lack of standards; they instead focused frustration on the lack of depth and utility within existing standards, as well as poor clarity regarding the application of current standards. EMR vendors corroborated the view. Providers and vendors confirm that HL7, LOINC, and RxNorm facilitate better sharing of the bread-and-butter data exchange most providers expect. Yet, there is room to greatly improve these existing standards (see Stan Huff’s insights below). 

So what is generally missing? Could proactive EMR vendors make dramatic progress if they worked together to leverage the state of existing standards? What would happen if competitive vendors quietly collaborated with each other to open up the flow? If MEDITECH connected to Cerner and Epic to athenahealth so that all of their respective clients could easily turn on sharing? How much would that expand interoperability? Would that collaboration frustrate HIE proponents or diminish the value of The Sequoia Project or CommonWell? Could it possibly mean that the best minds will knock down competitive walls and come up with amazing solutions? Does a focus on standards really mean vendor cooperating to fine-tune and use available standards to improve interoperability? Who will be first to set the path and light the fire?

Standards and getting to data liquidity, a corollary note from Stan Huff: 

Standards are still needed depending on what you mean by standards. I see the single biggest task in data exchange as terminology mapping, either mapping from my local codes to standard codes or, worse, point-to-point mapping of local codes to local codes. This problem is not solved by using LOINC or SNOMED. With LOINC and SNOMED, there are nearly an infinite number of ways that data can be sent. For easy and interoperable data exchange, we will need to say exactly which subset of LOINC codes and SNOMED codes we are using and define the "shape" of the data (degree of pre- or post-coordination). So we don't necessarily need more things like FHIR, HL7 V2, CDA, CCD, C-CDA, HL7 V3, NCPDP, X12, etc. In that sense, we don't need more standards. But we need explicit detail on how codes are used in the standards. I don't know whether that should be called an implementation guide, detailed clinical models (my favorite), configuration information, or something else, but whatever it is, it is needed in addition to the existing standards or we won't make much progress on faster development of data-sharing applications. Stan Huff’s two cents (Health Information Technology Standards Committee [Office of the National Coordinator for Health Information Technology], board member HL7, co-chair of the LOINC committee, and HSPC collaborator)