Accuracy of detecting changes in auditory heart rate
in a simulated operating room environment*
E. Chou,1 J. Lim,2 R. Brant,3 S. Ford4 and J. M. Ansermino5
1 Medical Student, Faculty of Medicine, The University of British Columbia, Vancouver, Canada
2 Research Manager, Department of Pediatric Anesthesia, British Columbia Children’s Hospital, Vancouver, Canada
3 Professor, Department of Statistics, The University of British Columbia, Vancouver, Canada
4 Specialist Registrar in Anaesthesia, Department of Anaesthesia, University Hospital of Wales, Heath Park, Cardiff,
UK
5 Assistant Professor, Department of Pediatric Anesthesia, British Columbia Children’s Hospital, Vancouver, Canada
Summary
The threshold for the identification of changes in heart rate and the accuracy in estimating heart
rate were compared between 20 anaesthetists and 20 non-anaesthetists in a simulated operating
theatre, both with and without distraction tasks. Typical operating theatre distractions were sim-
ulated by requiring anaesthetists and non-anaesthetists to perform secondary tasks. There were no
differences found between the groups in identification of heart rate changes. The distraction tasks
reduced performance in both groups (to a greater extent in the anaesthetists group). A change of
> 10 beats per minute was required for 80% of the changes to be detected. An upward heart
rate change was more easily detected than a reduction. Anaesthetists were found to be marginally
better at estimating the heart rate change from an auditory tone alone. However, the study
did not confirm that anaesthetists have a superior ability to detect changes in heart rate than
non-anaesthetists.
.......................................................................................................
Correspondence to: Dr J. M. Ansermino
E-mail: anserminos@yahoo.ca
*Presented at the Society for Technology in Anesthesia Annual
Meeting, San Diego, CA, USA; January 16–19 2008.
Accepted: 2 May 2008
The anaesthetist working in the operating theatre envi-
ronment must gather information from multiple sources
to ensure the safety of the patient. These sources include
interaction with coworkers, direct observation of the
patient, and observation of the monitor display showing
the measured physiological variables. In current moni-
toring systems, information is predominantly communi-
cated visually using the clinical monitor display. The
amount of data displayed by these monitors often exceeds
the maximum human monitoring capacity [1, 2]. The
sonification (transfer of information by the use of sound)
of heart rate (by a change in the interval between tones)
and oxygen saturation (by a change in pitch) is used to
augment information transfer to anaesthetists. This multi-
modal information transfer assists clinical performance
during periods of high cognitive workload [2].
The transfer of precise values of oxygen saturation
using sonification has been shown to be poor, and of
greater concern is that a 8% change in oxygen saturation
was required before 95% of anaesthetists noticed a change
[3]. In an abstract [4], the threshold for detecting a change
in heart rate sonification has been described as 8 beats per
minute (bpm), a rate at which < 50% of anaesthestists
could correctly identify changes. The addition of a ‘click’
sound at predetermined time intervals did not improve
the accuracy of heart rate estimation. This study used a
quiet environment, unlike the everyday theatre environ-
ment, and did not consider the effects of concomitant
tasks, which are common in the normal clinical setting.
In this study, we investigated the anaesthetist’s ability
to detect changes in heart rate and their accuracy in
estimating heart rate. We hypothesised that anaesthe-
tists could detect a smaller increment in heart rate
change than non-anaesthetists; that a distraction would
reduce the accuracy of heart rate change detection;
and that anaesthetists would be more accurate than
Anaesthesia, 2008, 63, pages 1181–1186 doi:10.1111/j.1365-2044.2008.05629.x
.....................................................................................................................................................................................................................
 2008 The Authors
Journal compilation  2008 The Association of Anaesthetists of Great Britain and Ireland 1181

non-anaesthetists in estimating the heart rate from the
heart rate tone.
Methods
Following approval from the Children’s & Women’s
Hospital Research Review Committee and the Univer-
sity of British Columbia Clinical Research Ethics Board,
a group of trained anaesthetists (experts) and a group
of subjects with no anaesthesia training or previous
experience with using continuous physiological monitors
(non-experts) were compared. Written informed con-
sent was obtained from all subjects. Information on
gender, age range (years) (20–29, 30–39, 40–49, 50–65,
> 65), duration of sleep in the past 24 h, musical
training, and number of caffeinated drinks consumed in
the past 12 h was recorded for each subject. The expert
group indicated the number of years of anaesthesia
training and experience for each individual anaesthetist.
All subjects were screened for hearing deficits of 25 dB
or more in the speech range of 1000, 2000, and
4000 Hz, using the MAICO Audiometer MA39 (Maico
Diagnostics, Eden Prairie, MN, USA). This is in line
with the American Speech Language and Screening
Association guidelines for audiological screening [5].
Subjects were asked to verbally identify the sound
coming into either the left or right ear.
Environment
Subjects were seated in a quiet room with a 15.4’’ laptop
computer (Acer Aspire 5602WLMi; Acer America Cor-
poration, Missisauga, ON, Canada) displaying a simulated
monitor similar to the A ⁄ 3 Anaesthesia Display (Datex-
Ohmeda, Madison, WI, USA) without the heart rate
value displayed. The simulated monitor was created using
the FLASH (Adobe Systems Inc., San Jose, CA, USA)
programming environment. The monitor displayed
changing physiological variables previously recorded from
a real clinical case. A prerecorded auditory track of
operating theatre noise was played to simulate the theatre
environment. The noise soundtrack was set at 60 dB in
keeping with previously reported noise levels [6] con-
firmed by our own observations.
Training
All subjects completed a 5-min training phase with use of
the keypad (Kensington Computer Products Group,
Redwood Shores, CA, USA) to input responses for
Phases I and II. Each key was labeled with degree of
change in heart rate of 3, 5, 7, 10, 15, and 20 bpm. The
non-expert subjects were instructed in recording physi-
ological values on a standard anaesthetic chart. In train-
ing all subjects observed a 2-min simulation, with the
numeric heart rate value displayed on the monitor. This
simulation demonstrated six heart rate changes, occurring
at a fixed time interval of 10 s, to give an introduction to
the heart rate changes during the study. In Phase III
training, subjects observed a 2.5-min simulation, starting
at 60 bpm and increasing 10 bpm every 15 s to 160 bpm.
The keypad was not used, and the heart rate numeric
value was displayed on the monitor.
Task
The study consisted of three sequential phases. The
numeric heart rate value was not displayed during the test
phases.
Phase I: detection of heart rate increment without distraction
All subjects recorded the variables heart rate, blood
pressure, oxygen saturation, end tidal CO2, and anaes-
thetic agent on an anaesthetic chart once every 3 min for
the 15-min duration of Phase I. An auditory tone
representing 100% oxygen saturation and the heart rate
was played throughout the 15-min study period at a level
of 75 dB, which is consistent with outputs from several
commercially available heart rate tone generators [7].
Changes in heart rate were taken from data recorded
from real cases to reproduce an authentic rate of change.
The auditory heart rate was reproduced as the identical
clinical case for each subject. Increases and decreases in
heart rate occurred at 3, 5, 7, 10, 15, and 20 bpm
increments, superimposed on a heart rate of 100 bpm,
introduced in a random order. Each of these heart rate
changes occurred four times, for a total of 24 rate changes.
These changes were introduced at random time intervals
of 15, 20, 25, 30, 35, 40, 45, and 50 s; each interval
occurred three times for a total of 24 changes in time
intervals. The random order and timing of the changes
were necessary to prevent subjects from beat counting.
The subjects were asked to listen for changes in the
audible heart rate tone, and to input the results using a
numeric keypad with six labeled keys to denote the
detected change in heart rate (e.g., a 5 bpm increase
change detected, press ‘+’ and ‘5’ then press ‘Enter’ on
the keypad).
Phase II: detection of heart rate increment with distraction
In Phase II, subjects repeated the experiment as described
in Phase I whilst simultaneously performing a secondary
task. The secondary task used to create manual dexterity
and visuo-spatial demands was to complete two advanced
puzzles (IQubes by Toylogic, Hong Kong) which
required subjects to return uneven shapes back into a
cube within the specified time limit of 15 min. Two
different puzzles were provided for each subject to
complete.
E. Chou et al. Æ Detecting changes in auditory heart rate Anaesthesia, 2008, 63, pages 1181–1186.....................................................................................................................................................................................................................
 2008 The Authors
1182 Journal compilation  2008 The Association of Anaesthetists of Great Britain and Ireland