Abstract
Chris Anderson, Editor in Chief
Scarcely a day passes without a new announcement of some health system or diagnostics company launching a precision medicine initiative. Unfortunately, precision medicine isn’t measured by what you could do in theory or hope to do, but by what can actually be delivered by a doctor to a patient.
Perhaps the biggest limiter of the spread of precision medicine sits right in every doctor’s office, or on their tablets—the electronic health record (EHR). On the face of it, marrying a patient’s genomic data with the phenotypic, medication, imaging, and even unstructured data housed in the EHR is an elegant solution. Right now, it is anything but that.
Granted, there are a handful of providers delivering a brand of precision medicine to their patients. Many have tackled the problem themselves—Penn Medicine’s internally built pathway for incorporating genomic data in the EHR comes to mind—but these efforts have been relatively few, and significantly expensive. When it comes to delivering precision medicine to the masses, the EHR is the logical conduit, but there is still a mountain of work to be done.
Luckily, significant players have taken notice. Witness the pilot project launched recently by the Office of the National Coordinator for Health Information Technology (ONC), called Sync for Genes (see story page 40). Quite simply, Sync for Genes will be looking for ways to leverage a data-sharing standard called FHIR (Fast Healthcare Interoperability Resources). It’s a standard that has received broad praise from a notoriously fractious lot—healthcare data interoperability tech wonks.
Better yet, the five pilots announced under Sync for Genes are backed by some healthcare and genomic high-flyers including Illumina, Foundation Medicine, and Intermountain Healthcare. The hope is these leaders in the field can find ways to crack two of the biggest problems standing in the way to incorporating genetic data into the EHR: the lack of a standard format for the capture of genetic data; and the sheer size of genomic data files.
In the end, the success of implementing FHIR (pronounced “fire”) for this task may come down to a simple change in method. FHIR is a “pull” technology, meaning it asks for only the data it needs, only when it needs it, and pulls it in from wherever the data are stored. Older technologies were “push”, meaning the holder of the data pushed the entire dataset to the user. Suffice to say, doctors are neither inclined, nor trained, to distill an entire 30-page genomic data report.
Instead, using FHIR and an application-based approach, queries for much smaller portions of the data can be pulled back to the doctor’s EHR, analyzed, then presented in a short digestable format to support the doctor’s care plan. It’s a small change in thinking that could yield big results.