Summer is finally here, even in Boston. Lilac and honeysuckle scent the air and the warm sun lingers until almost 9 pm. It’s time for flip flops, t-shirts and s’mores. There’s nothing better than a summer evening with friends around a campfire, roasting marshmallows until golden brown, dropping them onto a chocolate bar, and pressing them into gooey graham-cracker sandwiches. If you are a patient roaster and really get the marshmallows warm through and through, they melt the chocolate just enough to create the supreme s’more. The result makes even a non-Girl Scout like me love a campfire.
However, this blog post isn’t about indulgent campfire desserts; it’s about the FHIR (or, really the fire, as it’s pronounced) that makes it all possible. At the National Forum on Data and Analytics last month Charles Jaffe, the CEO of HL7, talked about the new HL7 FHIR product as a modern web technology, an “implementation delight,” that makes FHIR to data sharing what a real campfire is to a marshmallow—an opportunity to turn something cold and hard to swallow into something toasty and delicious.
HL7
Fast Healthcare Interoperability Resources or FHIR is the brainchild of the HL7 organization. HL7 is a non-profit that has been an international Health IT leader in the development of data standards including HL7 v2, HL7 v3, RIM, and CDA. Data standards create the opportunity for data exchange, integration, sharing, and retrieval. The ability to share and exchange data will be essential to achieving Health IT-driven improvements in healthcare delivery, workflow improvements, and knowledge transfer. Standards for health information exchange (HIE) will transform the multitude of EHR data silos into a functional collaborative network.
Data interoperability made easy
According to Dr. Jaffe, two-and-a-half years ago HL7 took a fresh look at Health IT standards and interoperability and asked, “What would you do if you could do HIT over?” The consensus response was that data exchange, or interoperability, should be easier, faster, and less costly which led to the development of FHIR. HL7 provides the FHIR specifications for free and assembling it as a framework for data management is purported to be easy and fast. Today the HL7 website describes FHIR as a “draft standard for trial use” and it appears that HL7 is still closely monitoring FHIR functionality and performance.
In very basic terms, since that is all I can manage for sure, FHIR collects healthcare data including clinical data into “Resources” which become exchangeable data packets. Each Resource has a standard structure or “pattern of elements” and specific metadata that describes its structure and/or content. One aspect of FHIR that enhances the usability is that it is “human readable,” meaning that the narrative summary of the data packet is in plain English not programming language.
The individual Resources may have enough meaningful data to be useful, but the plan is for the FHIR to bring together standard base “Resource sets” that can be exchanged among healthcare entities. For example a hospital may determine that a core information set should be sent to the appropriate recipient when a patient is discharged from the hospital. The set may contain data from various hospital information systems including devices, labs, and health records.
HIE in the real world
During the national analytics conference Dr. Jaffe mentioned 2 point-of-care initiatives underway with FHIR. One involves using FHIR to parse vast amounts of genomic data and present only pertinent selective data at the point of care. The other focuses on helping over 3,000 public health organizations transition from faxing documents to the CDC to having real-time data capture and exchange.
During the Q&A after Dr. Jaffe’s session, “Connected Health: Linking Records Across Health Systems,” the director of the Kansas State Health Information Exchange spoke from the audience to request that proprietary systems solve HIE problems, since the average $17,000 cost per interface is a barrier to data sharing. Dr. Jaffe replied that in the future there will be no interfaces; instead, distributed computing will provide access to data when and where it is needed. That will definitely be a warm and toasty marshmallow, melted-chocolate, graham-cracker-sandwich, campfire day.
Leave a Reply