Medical Informatics
Oregon Health Sciences University
Medical informatics: better technology for better health care
Reproduced with permission from NeXT
Computer, Inc.
A Reference
Guide to NeXT in Higher Education, Fall 1992
ยช
1992 NeXT Computer, Inc
The 11 faculty members in Biomedical Information
Communication Center (BICC) at Oregon Health Sciences
University use their cluster of 45 NeXT machines to conduct
research in medical informatics, a discipline focusing on
the use of information technology in the health-care
industry.
Researchers use NeXTSTEP developer tools, including the
Application Kit, Interface Builder, and Objective C,
according to Kent Spackman, BICC's associate director for
research and development. "NeXT has confirmed many times
over that it is the fastest development and prototype
platform around," he says. "X-Windows and the Macintosh are
slow to program and more difficult to operate. I've been
using NeXT machines since 1988 and was convinced of its
value early on."
BICC has several NeXT-based projects under way. The
following are a sampling:
End-user computing. Led by Spackman,
researchers are developing physician-station software to
enable rapid, high-quality health-care decisions. The
software will integrate medical information found in
various hospital and clinical databases. Health-care
professionals will not only be able to access the database
but also to network with colleagues locally and statewide.
The NeXT-based workstation will also evaluate laboratory
data and monitor the appropriateness of requests for
laboratory tests and blood products.
"The workstation will help doctors overcome the problem of
information overload in making decisions," says Spackman.
"It's a way for doctors to gather information in a form
that's very easy to assimilate. Previously, doctors had to
consult with other doctors, talk with the nurse and track
down a number of other personnel to gather information,
which caused a large amount of administrative overhead. All
the information available to make a decision was either on
paper or on numerous computer systems. The workstation will
provide summary information from a variety of sources,
which saves time and improves decision making. In the past,
treatment decisions were based on a single point in time
because doctors did not have time to trace the case
history.
"The workstation is to the doctor what Lotus Improv is to
the businessperson," he adds. "It's a presentation
mechanism and a way to do what-if analyses on medical
data."
Image Analysis. Today, most medical
decisions depend on visual images. Through their image
analysis research, BICC faculty hope to develop
methodologies for computer-assisted quantitation of this
visual data, which will improve physicians' ability to
detect subtle changes in patients' conditions.
Assistant Professor of Dermatology Douglas Perednia is
investigating uses of digital images in dermatology. With a
National Cancer Institute grant, Perednia is working on
various image analysis techniques to improve melanoma
screening. Using NeXTdimension, he analyzes and tracks
images of patients' skin over time. With this method,
Perednia expects small changes-that would not have been
detected using traditional methods-will be caught so the
melanoma can be removed early.
"With NeXTdimension, we have state-of-the-art, high-end
color capabilities," says Spackman. "Plus, the large,
high-resolution screen lets us not only view various parts
of our research, but a variety of menus and tools as well."
Information retrieval. Assistant Professor
of Medicine William Hersh is using the NeXT platform to
develop methods for indexing and retrieving information
from biomedical literature so answers to clinical and
research questions can be more easily and completely
acquired. According to Spackman, this work builds on the
National Library of Medicine's metathesaurus for
translations among different medical terminologies and
coding schemes.
"We're using NeXT as a generic UNIX box for this
application," says Spackman. "We hope to create a
rapid-search indexing option to obtain medical information
that's available over the Internet. The application will be
similar to Digital Librarian."
Outcomes research. Outcomes research assesses which
techniques do or do not work in health care. It provides
comparative evaluations of different treatment methods for
a given condition. For example; Is a person with a
particular ailment best treated as an in-patient or an
out-patient? Is it recommended that a patient in a certain
condition be transported from a rural location to a
hospital or not?
To help doctors make the best decisions when faced with
these dilemmas, BICC researchers have devised a series of
neural networks that are trained to use data collected from
the real world to predict the outcome of medical
procedures. These predictive models are then
cross-validated on BICC's NeXT network using Zilla, a
computationally-intensive application.
"With its massive virtual memory and efficient
communication facilities, running Zilla on a network of
NeXT machines is like having the capabilities of a
supercomputer," says Spackman. "Zilla is a very powerful
computational tool."
For more information, please contact:
Kent Spackman, M.D., Ph.D.
Associate Director for Research & Development
Biomedical Information Communication Center
Oregon Health Sciences University
3181 S.W. Sam Jackson Park Road
Portland, OR 97201-3098
(503)494-4502
spackman@ohsu.edu