viewpoint article in the Journal of the American Medical Association (JAMA). This approach, the authors argue, could enable doctors to benefit from the latest genomic advances without waiting for EHR vendors to catch up with this scientific revolution. But eventually, they add, EHRs will have to change.
In "Crossing the Omic Chasm: A Time for Omic Ancillary Systems," the three authors -- Justin Starren, MD, PhD; Marc S. Williams, MD; and Erwin P. Bottinger, MD -- cite a recent Institute of Medicine report noting that storage of genomic, epigenomic, proteomic and metabolomic data (the "omics") is not feasible in the current generation of EHRs. EHRs are designed to display only clinically relevant information, they point out. That's why radiological images are typically stored in picture archiving and communication systems (PACS), and only the radiology report is sent to the EHR.
While a medical image typically represents nearly 300 times as much data as the report based on it, whole-genome sequencing of a particular individual requires 5 to 10 GB of storage, 50 times more than an image. So an ancillary system is required to store that amount of data, the article says.
In addition, the authors observe, EHRs do not include the analytics needed to interpret genetic variations in light of the latest scientific research. And the genetic data will have to be stored for a long time, they point out, to reanalyze and reinterpret the genomic results in the context of evolving knowledge.
[ Genetics-based CDS will be routine five years from now, experts say. Learn more about Health IT In 2018: Crystal Ball Predictions. ]
The authors predict that large organizations will likely operate their own omics ancillary systems. Small practices will probably use reference labs, which will add omics ancillary services to their current service lines.
The article offers three ways that decision support based on genomic analysis could inform medical decision making:
1. The results of the analysis could be converted to a text report that would go to the clinician.
2. "Computable observations" could be created and stored within the EHR, where the observations could be used to trigger conventional CDS rules.
3. An external CDS system could be queried by the EHR user in the clinical workflow.
In an interview with InformationWeek Healthcare, Justin Starren, chief of the division of health and biomedical informatics, department of preventive medicine, at Northwestern University Feinberg School of Medicine in Chicago, said that Northwestern is using the second option. To explain how this works, he cited a genomic-based clinical decision support tool that has been piloted at Vanderbilt University. In this project, Vanderbilt University researchers use genetic markers to predict how patients will react to the drug clopidogrel, which is used to prevent blood clotting after a stent has been inserted. If the patient's genomic data indicates that they don't have the ability to metabolize clopidogrel quickly into the active compound needed for clotting, it's recommended that they receive a different medication with a similar effect after a stenting operation.
At Northwestern, when genomic data comes in from the lab, the data is processed into the attributes that the researchers want to store, whether or not they know what those mean. Then they parse the data further to decide which of these attributes is clinically relevant. If they notice that a patient has a mutation that makes him or her a low metabolizer of clopidogrel, for example, they can send that "computed observation" to the EHR, where it can be stored in the system's observation table. Then, if clopidogrel is prescribed to a particular patient with the mutation, the EHR's CDS engine can send an alert to the doctor.
The Electronic Medical Records and Genomics (eMERGE) network is doing further research in this area with funding from the National Human Genomic Research Institute (NHGRI). The first phase of eMERGE's research, Starren noted, showed that clinical and genomic data could be combined to do scientific work. In the second phase, the consortium is continuing in the same direction, but progressing to whole-genome sequencing and the implementation of genomic CDS in the EHRs of the target sites, he said.
Participants in phase one of eMERGE included Northwestern University, Mayo Clinic, Vanderbilt University, the Marshfield Clinic and the University of Washington's Group Health Cooperative. In phase two, the consortium added Mount Sinai Medical Center, Children's Hospital of Philadelphia, Boston Children's Hospital, Cincinnati Children's Hospital Medical Center and Geisinger Health System.
Regulatory requirements dominate, our research shows. The challenge is to innovate with technology, not just dot the i's and cross the t's. Also in the new, all-digital The Right Health IT Priorities? issue of InformationWeek Healthcare: Real change takes much more than technology. (Free registration required.)