ACADEMIC EMERGENCY MEDICINE • November 2000, Volume 7, Number 11
1263
TABLE 1. Questions Asked Regarding Each of the 11 Video Cases Presented
What is the quality of the image? 10 (1 is very poor; 10 best quality)
1
2
3
4
5
6 6 8 9
Is there a pericardial effusion present?
If an effusion is present, how big is it:
yes no
small (5 to 9 mm) moderate (10 to 15 mm) large (16 mm or more)
How sure are you?
Do you feel there is evidence of cardiac tamponade?
How sure are you about tamponade?
Would finding change your management?
1
2
3
4
5
6
7
8
9
10
(1 is not sure at all; 10 is has to go to the OR now)
yes no
1
2
3
4
5
6
7
8
9
10
(1 is not at all; 10 is very sure)
yes no
Would you like other views to feel more certain?
If you would like other views, which ones?
yes no
parasternal long parasternal short apical four chamber
presented with repeating digitized video clips of data were collected. All patient information was
real-time ultrasound examinations on trauma pa- entered into a Microsoft Excel 5.0 spreadsheet (Mi-
tients with penetrating wounds to the chest, back,
or upper abdomen. The video clips were specifically
selected because of the difficulty in interpretation
they presented to an experienced ultrasonographer
(MB)* during actual patient evaluation. Each par-
ticipant was provided with a total of 11 cases,
which were randomly ordered. The cases contained
normal examinations, examinations showing peri-
cardial effusions, and examinations demonstrating
epicardial fat pads. Two clips were of true pericar-
dial effusions, both caused by penetrating wounds
crosoft Corporation, Redmond, WA). Data were
pooled and then analyzed as independent varia-
bles. Data were analyzed using descriptive statis-
tics from a commercially available software pack-
age (Analyse-it 1.44, Analyse-it Inc., Leeds, Great
Britain). Accuracy, sensitivity, and specificity with
95% confidence intervals were calculated.
RESULTS
to the heart. Five were of patients with epicardial Five attending physicians and 17 residents partic-
fat pads and no pericardial fluid. Four cases were ipated in this study. All participants completed 11
of patients with no pericardial fluid or epicardial
fat pad.
video case presentations. Each video clip was ap-
proximately 20 seconds long and was repeated six
times in a continuous loop. Participants were ques-
tioned about presence and significance of any pe-
ricardial effusion. Tables 2 through 4 summarize
the results. The difficult video segments proved
challenging for all participants in discerning epi-
cardial fat pad from effusion. The overall accuracy
for discriminating an epicardial fat pad from a pe-
ricardial effusion was 30% (95% CI = 23% to 38%).
Normal examinations and those with pericardial
effusion were detected in 73% (95% CI = 64% to
82%) and 73% (95% CI = 58% to 84%) of cases,
respectively.
Sensitivity ranged from 63% to 93% for groups
depending on level of training and experience.
Specificity ranged from 31% to 61%. Both sensitiv-
ity and specificity tended to increase with increas-
ing level of training or experience. The level of con-
fidence also increased with experience and
training level. Attending physicians were the most
confident regardless of whether their answers were
correct or incorrect. Each group appeared to have
similar confidence levels for right and wrong an-
swers.
The study participants were given a clinical sce-
nario that applied to all 11 cases. The study sub-
jects were asked to imagine that they were in a
trauma bay with an unstable and critically ill
trauma patient who had multiple injuries, includ-
ing a penetrating chest, upper abdominal, or back
wound. The patient’s vitals were worsening and
the presiding trauma surgeon requested a rapid
FAST examination. The focus of the examination
is the subxiphoid view of the heart, which is part
of the standard trauma ultrasound examination.
All ultrasound examinations used in this study
had been previously corroborated with either chest
computed tomography or echocardiography pro-
vided by the cardiology service. Study participants
were asked to give their level of training as PGY1,
PGY2, PGY3, or attending. They were also asked
for the number of ultrasound examinations they
had performed and the number of hours of didactic
lectures and hands-on classes attended. Standard
questions (Table 1) were asked for each of the 11-
video cases. The participant’s level of confidence in
interpreting the exam was measured with a ten-
point Likert scale.
The difficulty of the video cases was illustrated
when the majority of participants requested addi-
tional views to help define the presence and extent
of the effusion. The most common additional view
requested was the parasternal long approach.
When the study participants thought that an ef-
fusion was present, all believed that their manage-
Data Analysis. Standardized data collection
sheets were given to participants. No identifying
*MB and MBP are RDMS (registered diagnostic medical so-
nographer)-certified.