Title 22”x3.5” - University of Virginia

Download Report

Transcript Title 22”x3.5” - University of Virginia

Developing an Electric Flowsheet for Pediatric Intensive Care Units
David Bauer and Stephanie Guerlain*
University of Virginia Department of Systems and Information Engineering
School of Engineering and Applied Science, Charlottesville, VA 22904-4747
*Co-Director, Medical Informatics Training Program [email protected]
Flowsheet Analysis
Introduction
•
By studying PICU flowsheets in the context of their actual use, we have
defined requirements based on how clinicians track patient data.
In the pediatric intensive care unit (PICU), large amounts of data are
generated by distributed sources. Paper flowsheets bundle data in a central
location, but may not be the optimal design for electronic systems.

•
•
Electronic medical records have the potential to extend health care providers’
abilities to view, integrate and share information. However, the available
electronic products have largely been designed to replicate paper-based
systems without taking advantage of what we know about human perception
and decision making and the “user centered design process.”
The prototype electronic flowsheet surpasses the capabilities of available
systems. The proposed design enables better tracking of patient state by
providing and showing relationships among variables. Novel graphical
elements have been developed for visualizing data trends, acid-base balance,
fluid balance, and the Glasgow Coma Scale.



Objective
The purpose of this study was to apply user-centered design principles to
create an electronic flowsheet for pediatric intensive care units that will
address the shortcomings of available paper and electronic systems.
Paper flowsheets are designed for
ease of data entry, grouping data by
source.
Physician daily notes are grouped by
system, integrating lab values,
medications, vital signs, etc.
Strengths
 Bundles and summarizes key
information
 Both structured and flexible
 Portable (locally)
Weaknesses
 Static
 Spatial constraints
 Determining trends and ranges
requires many comparisons
A sample page from a flowsheet containing arterial
blood gas and other laboratory results.
Design Concepts
Zones for common
diagnoses
Other input
Drips
Mouse-over data
points for time and
values.
Urine
Other output
Data point shade one
indicator of time
Cumulative
Slider for finding
cumulative totals of
any time range
Axes show data range
A graphic for hourly and cumulative fluid balance.
Click-and-drag target
value. Data points
resize based on
distance from target
Adjusted (purple) and
normal O2-Hemoglobin
dissociation curves.
Reference
ranges appear
on mouse over
Grey bar (normal range) is
a 24 hour timeline. Lighter
grey shows 48 hours.
Acid-base map and O2-Hemoglobin Curve
Mouth = verbal response
Eyes = eye response
Pupils = pupil dilation
Eyebrows = pain
Ears = motor response
Nose = turn/position
Black lines is the patient’s
range over a given time period,
and the dot is the currently
selected value. The text is also
the patient’ range.
Most recent value, or value for
time period on mouse over
Critical values appear in red.
Faces represent GCS and other nursing assessments
Trend and range graphics for lab data and vital signs.
Proposed Design
Design Goals
• Consolidate as much information on a single screen as possible
• Use graphics that provide an overview with details-on-demand
• Place related data elements on the same time scale
• Show the relationships among data
Future Work
• Continue to add functionality for filtering and customizing views
• Create patient cases for a formative evaluation with attending physicians
and for an experiment with residents
• Develop training scenarios for residents, particularly focused on acid-base,
electrolyte and fluid balance
In addtion to the design concepts shown above, the interface contains a list of most recent values
of significant variables, a timeline for events and data collection, and problem, event, and
medication lists.
Acknowledgements
This work was supported by the National Science Foundation under Grant No. 0092985 and by the National Library of Medicine Medical Informatics Training Grant