Biomedical Engineering Reference
In-Depth Information
purposes. When available, diagnosis codes entered
for patients who present themselves to acute care
clinics or emergency departments and who are
subsequently admitted to a hospital can be used to
track the types of presenting symptoms or illnesses
that are most likely to result in admissions. Other
information collected at the time of patient admis-
sion, such as text fields containing admitting diag-
nosis, current procedure terminology (CPT) codes,
and even nursing admission orders could be used
to categorize patients syndromically.
A second approach to the syndromic coding
and evaluation of all admitting diagnosis codes is
the tracking of admissions by different wards in
the hospital (medical versus surgical), severity of
illnesses (e.g., admission to the intensive care unit),
age of patients, or even time of day of admissions,
as all of these factors may provide an indication
that something unusual is occurring in the health
of the community. Even if diagnosis codes are
not available upon admission, discharge diagnoses
can be useful in the development of syndromic
systems, although timeliness of event detection
would be affected. Evaluation of whether any
discharge diagnoses of acute infectious diseases of
interest correspond to initial chief complaints may
help define which presenting complaints are most
important to monitor.
some. Typically, either the ED staff is required
to fax chief complaint logs to public health enti-
ties, or public health staff is physically detailed
to hospitals to enter data during the surveillance
period. When the data collection consists of faxing
ED logs, substantial work has to be performed by
health department staff upon receipt of the data:
Staff is required to review the logs and manu-
ally code each patient into the selected syndrome
groups and enter the information into computer-
ized databases. Following this data preparation,
staff must analyze the data, follow up on important
findings, and report the results of their work. This
process is difficult to sustain over long periods of
time due to the large burden imposed on already
resource-limited health departments—particularly
given the increased workload brought about by the
threat of bioterrorism [15].
However, advances in informatics can expedite
the entire process with new capabilities to capture
chief complaint data directly from hospital data
sources, transmit the information electronically,
and automate the coding process. Such automation
removes the burden from emergency department
staff and dramatically reduces the burden on health
department staff. Indeed, a number of jurisdictions
have systems in operation with completely auto-
mated data transmission, coding and preliminary
analyses, and thereby allowing health department
staff to focus their efforts on reviewing important
findings [3]. The development of chief-complaint
based syndromic surveillance has been developed
for use in the public health sector. However these
systems also provide a valuable resource for hospi-
tals in the absence of any such system by local
health authorities. Ideally, however, syndromic
surveillance systems can, and should, be developed
as a joint endeavor between hospital and public
health authorities for coordinated use by both.
Chief-complaint data often provide more timely
information than diagnosis-coded information
because of the significant delay in many hospi-
tals in classifying complaints into diagnosis coding
schemes. Chief complaints, however, are available
as free or fixed text at the time of patient regis-
tration. The free text data can be transmitted to
a surveillance system that will parse it into the
3.2 Hospital Emergency Department
Data
Emergency Department (ED) chief complaint data
represent a second major source of hospital infor-
mation that may be useful to hospital institutions
and public health officials for routine surveil-
lance. Such efforts began as “drop-in” surveil-
lance systems for focused periods of time such
as following major disasters (natural and other-
wise) and during high profile sporting and political
events [13]. This type of surveillance assumed a
high level of visibility and scrutiny immediately
following the attacks on New York City's World
Trade Center in 2001 [14].
For most of these drop-in surveillance systems,
collection and processing of data by hospital staff
and public health officials is enormously burden-
