A VP of IT at Catholic Healthcare West explains the impact that clinical genomics could have on electronic records.
Healthcare Technology: You are a leader in the clinical genomics
space. Can you tell us a little bit about your work with genomics?
Scott Whyte: A few years ago, I became interested in the topic of genomics
and its ultimate impact on clinical information systems (CIS). I had heard one
of the leading researchers with the National Institute of Health, Jeffrey Trent,
comment that one of the biggest challenges was the exchange of genomic information.
Independent researchers would work on sequence and profiling information but
it was difficult to share that information electronically with others without
human intervention to reconfigure the information before it got loaded into
another database. He said that, to a certain degree, this slowed down research
collaboration and progress.
Although Im not a researcher, I predicted we would have a more significant
problem when it ultimately came to patient care. We are now moving from bench
research to translational research, where researchers are looking at specific
drug therapies or diagnostic techniques; some of which are in place and being
used by doctors and hospitals.
Today the volume of these therapies is relatively low, but as the volume builds,
we need to be able to exchange genomic information from one machine to another
without having a human intervene. I have a background on solving these types
of problems, so I volunteered with Health Level 7 (HL7), which is an international
standards organization. They had not yet begun work in genomics, so I helped
form the team, and Im currently a co-chair of the Special Interest Group on
HCT: Can you explain the connection between EMRs and clinical genomics?
SW: Today most of the genomic-related information is not stored in an
electronic medical record (EMR) at all, and if it is stored, its stored in
a text-based note. There arent discrete data fields that can be sorted and
queried or reported upon.
So if you think about a routine lab test, a blood chemistry, for example, those
values come back from a lab instrument into a laboratory information system
and then through an interface over to the main EMR. When the results get to
the EMR, theyre stored in very discrete values there are numeric measurements
for the glucose, calcium, electrolytes, etc. Each of those is stored as a value
in a little bucket within the EMR. As a result, you can graph that information
and trend it, and you can sort across patients and do all kinds of work it information.
Quantitative information is much more valuable to the clinician. Since the genomic
information now currently comes in as text, you cant do any searching, graphing
or comparison between patients or for a given patient over time.
Theres very little use of genomic information within the EMR today. There
are very few exceptions where more leading-edge institutions are building discrete
attributes in which to store genomic information. The Mayo Clinic is one, for
instance. The Marshfield Clinic in Wisconsin is also doing this. So there are
some institutions that are starting to integrate genomic information into the
HCT: What are some of the security and privacy concerns.? What kind
of barriers are we encountering?
SW: Well, the security and privacy concerns are essentially the same
areas of concern for all other healthcare information. Federal regulations,
such as HIPAA, are mandating privacy and security standards and compliance with
those standards. However, the genomic area is of particular concern because
of the predictive nature of the information. You can predict that theres a
higher likelihood for a particular disease to occur prior to it occurring. And
a patient may have the concern that if an insurer knew future likelihood of
disease based upon genomic information that the insurer might make his/her policy
either more expensive or effectively make them uninsurable. You can start to
extrapolate other implications there as well; for example, relating to employability.
The areas of concern are the same for all health information. The degree of
the concern, and sensitivity related to privacy is just much higher for genomic
I personally think that we will overcome the security and privacy issues.We
have exposures today that are still of great concern that we are addressing
with technology, laws and prosecution. I also think theres a certain scientific
humility as we learn more about the genomics.We learn about the value of genomics
and also learn more about the limits of the meaning of the genomic information.
For instance, certain gene markers tell us disease likelihood, but there absolutely
are behavioral and environmental factors that affect the disease. So while you
may have a particular single nucleotide polymorphisms (SNPs), it is not a foregone
conclusion that youre going to get breast cancer or some other disease.
The value of genomics in medicine will outweigh privacy concerns. Today breast
cancer and the BRCA1 and BRCA2 genes are talked about the most. In addition,
theres work with other cancers, diabetes, heart disease and obesity; there
are many other research and development efforts related to predisposition. But
much of the work goes beyond predispositions to evaluating which particular
therapy is effective on the person. Theres significant work being done on pharmacogenomics,
which doesnt predict your ability to acquire the disease but assists in the
development and predicts the effectiveness of a particular drug.
HCT: Whats on the horizon for EMRs and clinical genomics in the
next five to seven years?
SW: I would say within a year or two we will see a number of pilots
of genomic information in EMRs. Theres will still be an early adopter phase,maybe
in three to five years. Ultimately I do see that the information will be more
routinely stored in the EMR. And theres a question about what that means. Initially
only very small fragments of information about a persons genome will be stored
only the fragments that result from testing for specific conditions. Those
test results, identifying the presence of a particular SNP or mutation for example,
will be placed in the EMR. Your whole genome sequence will not be in the EMR
at least not initially. I think thatll be a long way off. As the volume of
information increases, security and storage issues become much more significant.
We still wrestle with the issue that there is not a common vocabulary in the
clinical setting and I think thats really the biggest barrier.We need a common
vocabulary to describe the whole topic of clinical genomics and all the various
pieces and parts. There arent universally accepted terms for genomic components
and how they interact with each other from a care delivery perspective. Until
that common vocabulary is better defined, its going to be harder for widespread
adoption to occur. So its really more of a clinical issue or medical terminology
issue than it is a computer issue. Right now the advances are so rapid, there
are billions of dollars being invested worldwide by government entities and
by pharmaceutical companies. The knowledge is growing at an incredible pace.
The method to categorize and agree on the meaning of those findings is not keeping
pace with the progress. There is no shortage of challenges and opportunities
to including genomic information in the EMR.