|
ATMs for healthcare
DCLnews talks
to IT expert Landen Bain, who is using XML and other standards to digitize clinic
notes and clinical research documents.
Landen Bain is a
prime mover in the development of IT infrastructure in healthcare. A former
chief information officer at Duke University Health System, North Carolina,
and at Ohio State University Hospitals in Columbus, Bain has a big vision to
transform the delivery of healthcare. He wants to digitize the country's health
records and make them easily accessible to doctors and medics across the nation.
That way, if you
are taken ill in Michigan, but live in LA, paramedics will be able to get your
personal health details up on screen in a flash and give you the appropriate
treatment faster and more safely.
This, however, is
not as straightforward as you might think - even in these days of Internet banking
and other information exchange systems. Security and privacy issues add a great
deal of complexity. In the past, the big problem was the lack of information
availability, along with the difficulty in deciphering the multiple data formats
being used. If medical information becomes more easily transportable and standardized
it becomes critical to set up suitable protection of private medical data.
To find out more,
DCLnews caught up with Bain during a hiatus in his busy schedule and asked him
about the main issues involved in implementing health e-records.
Landen
Bain: People are always saying, "Look
at how automated banking is, what with ATMs and internet banking - why can't
you do the same with healthcare?" And it's true, you can access your records
from a hamlet in the Himalayas, if that's where you happen to be. But for a
doctor to access your medical records from another state in America you need
phone calls, faxes, transcriptions, and often a good deal more. So it's understandable
that people want to know why healthcare isn't automated like banking.
|
About Landen Bain |
|
Landen Bain is currently involved in two standards bodies in healthcare - HL7 (Health Level Seven) and CDISC (Clinical Data Interchange Consortium). Both are non-profit. HL7 focuses on its Clinical Document Architecture (a standard for clinical vocabularies) and on developing data standards for clinical trials.
While CDISC is committed to the development of industry standards to support the electronic acquisition, exchange, submission and archiving of clinical trials data for medical and bio-pharmaceutical product development.
Bain is exploring the common area between HL7 and CDISC.
"It's surprisingly unexplored," he says. "The pharmaceutical companies and the clinical trials they sponsor are much more disconnected from patient care than they should be. I aim to change that."
|
All I can say is,
it certainly could be if all you had to deal with had two digits to the right
of the decimal place and was a number. But that's not the case. Medicine is
language-based; and the sharing of data in medicine is not the exception, it's
the rule. In banking information is only occasionally transferred from bank
to bank by users. In healthcare it's a constant occurrence. Added to that, medical
and healthcare information covers a myriad of different types of information,
in many different formats, across a wide geography, and spanning many years
worth of information. People who say healthcare is behind in its technology
are only looking at the surface.
DCLnews:
What are you currently working on to
implement health e-records?
Landen
Bain: Our project is called Single
Source. What we mean by that is a single data capture that creates a single
source document. This in turn populates the electronic health record on the
patient care side, and also populates the clinical trial manager system, which
constitutes the other aspect of our work.
DCLnews:
So you're focusing on two avenues
- patient care and clinical research?
Landen
Bain: That's right. We're trying
to come up with a way to gather data simultaneously, which meets the needs of
these two realms.
DCLnews:
Is XML involved?
Landen
Bain: XML is certainly part of it.
There are two standards and they're both XML-based. The key one is CDA, or Clinical
Document Architecture. This is under the HL7 family of standards. We believe
it can be the schema for the single source document. This in turn can produce
XML that complies with the CDISC (Clinical Data Interchange Consortium) standard,
which is called ODM or Operational Data Model.
|
Why XML?
Medical records
are more than just numbers (the medium banking deals with), which is why XML
will prove a critical element in tying healthcare "e-records" together.
XML provides a standard architecture that allows you to include text, numbers,
metadata, and explanatory materials - as well as X-Ray images, electrocardiograms,
voice recordings, and other forms of information not yet invented. Since other
standards' bodies are dealing with XML, the healthcare community, including
HL-7 and CDISC, can use it without having to reinvent the wheel
Data Conversion
Laboratory has been involved with the healthcare application of XML. One project
concerns Structured Product Labeling (SPL), which is being developed under the
HL-7 and CDISC umbrellas. This is an XML standard to standardize pharmaceutical
data, making it easier for pharmaceutical companies to keep information up to
date and to provide access to those who need it. This is part of an FDA initiative
to make such information more easily available (see FDA draft guidance at: http://www.fda.gov/OHRMS/DOCKETS/98fr/04-2536.htm).
|
DCLnews: What
are you doing in practice to make this a reality?
Landen Bain: We're
about to conduct a single clinical trial at Duke University in North Carolina,
and we'll use that as a proof of concept to show the feasibility of jointly
capturing data for clinical trials and health records. Two types of document
are being captured: Case Report Forms (CRF) for clinical trials, and Clinic
Notes for health records. Case Report Forms are highly structured; every data
element is defined precisely. Whereas Clinic Notes are very loose in structure;
they might include off-the-cuff observations or recommendations from doctors,
for example. You have to be able to accommodate both.
DCLnews: When
will this be rolled out?
Landen Bain: We
have the technical architecture in place. We have the tools selected. We've
begun the process of designing the forms capture for both the Case Report Form
and the Clinic Note. So we expect to implement the proof of concept within two
or three months.
DCLnews: Where
do you see it going after that?
Landen Bain: Following
the proof of concept at Duke, we will look at testing the system in a number
of physicians' offices. This could be put into action in a year.
DCLnews: How
do you see all this benefiting the world?
Landen Bain:
The
main benefit will be it will help pharmaceutical companies get new drugs tested
and to market sooner. Currently there are a lot of hold ups. Not in the bench
research, but in the clinical trials. Getting enough patients enrolled to complete
these trials is what's holding things up. Our system will help get data through
faster and break the bottleneck. It will also save money - especially when you
consider the drug company cliché that for every day a trial is delayed
it costs a million dollars.
DCLnews editorial
|