iAssist: A Software Framework for Intelligent
Patient Monitoring
Christopher Brouse∗, Guy Dumont∗, Ping Yang∗ Joanne Lim∗∗, J. Mark Ansermino∗∗
Abstract—
A software framework (iAssist) has been developed for
intelligent patient monitoring, and forms the foundation of
a clinical monitoring expert system. The framework is ex-
tensible, flexible, and interoperable. It supports plugins to
perform data acquisition, signal processing, graphical dis-
play, data storage, and output to external devices. iAssist
currently incorporates two plugins to detect change point
events in physiological trends. In 38 surgical cases, iAssist
detected 868 events, of which clinicians rated more than
50% as clinically significant and less than 7% as artifacts.
Clinicians found iAssist intuitive and easy to use.
Index Terms—Clinical Monitoring, Patient Safety, Soft-
ware Framework, Expert System
I. Introduction
The unintelligent alarm systems currently employed in
operating rooms are a significant hindrance to informa-
tion transfer. Abnormal clinical conditions are signaled
by primitive auditory or visual alarms, automatically trig-
gered when a single parameter fluctuates beyond preset
upper or lower thresholds. The thresholds assume that
each parameter (e.g. heart rate, blood pressure, etc.) will
remain within the same range for each patient over time.
Natural physiological rhythms cause these parameters to
fluctuate, however, and the acceptable range varies for each
individual. False alarms are therefore inevitable, especially
since the alarm is based on a single parameter. Known clin-
ical interventions and artifacts from electrocautery noise
or movement exacerbate the situation; more than 90% of
alarms can be dismissed as insignificant, 1/3 of which are
triggered by artifacts [1]. This high false positive rate ren-
ders alarms ineffective as they are frequently ignored [2].
The clinician is unassisted in continuously observing the
parameters on the monitors, and in determining when a
clinically significant event has occurred. Unlike the alarm
systems, the clinician is able to understand and integrate
the underlying human physiology and its rhythms. Yet,
maintaining attention to detect subtle changes in param-
eters across multiple monitors is very difficult—especially
during very long surgical procedures. To further compli-
cate matters, the clinician is required to manage many
considerations simultaneously, including the patient, mon-
itors, fluid and drug delivery, and teaching students. A
balance must be reached between each responsibility, and
these demands can exceed the information processing ca-
pacity of even highly trained, focused clinicians [3]. Indeed,
attentional overload has been identified as a fundamental
cause of human error in clinical monitoring [4], often lead-
ing to poor clinical outcomes [5], [6].
The University of British Columbia, Vancouver BC, CANADA
Department of Electrical and Computer Engineering∗
Department of Anesthesiology, Pharmacology & Therapeutics∗∗
We aim to provide assistance to the clinician, with the
long-term goal of developing a clinical monitoring expert
system that can be adopted into standard clinical practice.
We believe we can improve patient outcomes by employing
a multi-stage approach to diagnosing patient conditions:
1. Noise and artifact detection and rejection. En-
sure that high quality data are used for processing,
and advise when only poor quality data are available.
2. Change point detection in physiological
trends. Detect clinically significant changes in pa-
tient state, which are often indicative of adverse con-
ditions.
3. Decision support system. Perform rule-based rea-
soning on patient data and change point detection re-
sults. Will incorporate an extensive clinical knowledge
base, acquired from domain experts.
The system will receive raw waveforms and trends from
standard physiological monitors for processing. Each stage
of the system will extract key information from the data for
use in the next stage, until the system reaches a diagnosis.
Diagnoses will be presented to the user as unified alarms,
aspiring to both high sensitivity and specificity.
Thus far, we have created algorithms to serve as part of
the first and second stages of the system. We have also cre-
ated a software framework (iAssist) based on the Java 5.0
platform, to incorporate these algorithms. As iAssist does
not yet employ rule-based reasoning, it cannot be called an
expert system. Instead, iAssist serves as an intelligent pa-
tient monitor, for testing and validation of our algorithms.
We aim to construct a solid foundation of event detection
algorithms, before incorporating high-level decision sup-
port.
In this paper, we discuss iAssist’s design relevance to our
current and future research. We also present the results of
a real-time clinical study of iAssist.
II. Design
A. Requirements
In order for iAssist to serve as an intelligent patient mon-
itor, it must satisfy a diverse set of requirements. We sep-
arate these into five functional areas:
• Input: acquire patient data as input from any number
of monitoring devices.
• Signal Processing: process data using any number
of artifact detection/rejection and change point detec-
tion algorithms.
• Data Storage: store patient data and signal process-
ing results for subsequent analysis.
• GUI: display patient data and signal processing re-
sults on-screen.
Proceedings of the 29th Annual International
Conference of the IEEE EMBS
Cité Internationale, Lyon, France
August 23-26, 2007.
FrP2D3.13
1-4244-0788-5/07/$20.00 ©2007 IEEE 3790