Content

THE NEW ZEALAND
MEDICAL JOURNAL
Journal of the New Zealand Medical Association
NZMJ 3 April 2009, Vol 122 No 1292; ISSN 1175 8716 Page 9
URL: http://www.nzma.org.nz/journal/122-1292/3536/ ©NZMA
Are our medical graduates in New Zealand safe and
accurate in ECG interpretation?
Nigel A Lever, Peter D Larsen, Mathew Dawes, Annie Wong, Scott A Harding
Abstract
Aim We aimed to assess the skills of final year medical students and resident medical
officers in recognising and interpreting important common or life-threatening
abnormalities in the electrocardiogram (ECG).
Methods 102 participants at two study sites (52 of whom were final year medical
students) attempted to determine the heart rate and rhythm and identify and interpret
any abnormalities present in 15 ECGs in a 30-minute time period.
Results Accurate determination of heart rate was poor, ranging from 0% to 89%
correct across the 15 ECGs. Normal sinus rhythm in 8 ECGs was identified 81% to
95% of the time, and ventricular tachycardia was identified by 98% of participants.
Atrial fibrillation (55%), second degree heart block (19%) and ventricular pacing
(9%) were not well identified. Four ECGs showed acute ischaemic ST segment
changes, and these were correctly identified in 87% to 93% of cases, although
interpretation of these abnormalities was less accurate. Long QT interval (7%) and
pre-excitation (WPW pattern, 11%) were not well recognised. Nearly half of the
participants rated their ability to interpret ECGs as less than satisfactory while just
over half rated the ECG teaching they had received as less than satisfactory.
Conclusions Overall study participants did not achieve what we would consider an
adequate standard in recognising and interpreting important common or life-
threatening abnormalities in the ECG. To address this we need to define minimum
standards in ECG interpretation, to improve our teaching to meet these standards, and
to assess our graduates against these.
A century after its introduction into clinical practice, the electrocardiogram (ECG)
remains one of the most commonly used tests for the assessment of cardiac disease.
ECG interpretation is therefore a core skill and medical practitioners are expected to
be proficient at reading ECGs. This is particularly important when dealing with
acutely ill patients where failure to recognise an abnormality may have life-
threatening consequences.1,2
There is increasing concern that ECG teaching and the
ability of medical professionals to interpret ECGs are declining.3
Teaching ECG interpretation is complex, and the best methods for ECG teaching to
recognise and interpret abnormalities remains unclear.4–6
The use of pattern
recognition versus understanding the principles of vector analysis has occupied
commentary, as has the role of computer-aided learning. Whilst there are guidelines
on the competency requirements for practitioners who are in training programmes for
various disciplines (cardiology, emergency medicine7
), guidelines are lacking for
teaching and assessment of medical graduates.

As part of our role in providing undergraduate ECG teaching, we undertook this study
to assess the skills of final year medical students and house officers in recognising and
interpreting important common or life-threatening abnormalities. We were also
interested in how this group rated their own ability to interpret ECGs and the teaching
they had received.
Methods
Final year medical students from University of Auckland and University of Otago, Wellington as well
as first and second year resident medical officers (RMOs) from Auckland and Wellington Hospitals
were invited to participate in the study between March and June 2007 during their regular teaching
programmes. The study was approved by the local institutional ethics committees. Respondents
consented to participate in an anonymised fashion.
Participants were given 15 ECGs (Table 1; Appendix 1—view ECGs at www.nzma.org.nz/journal/122-
1292/3536/appendix.pdf) to interpret over a 30-minute time period. For each ECG they were asked to
determine the rate and rhythm, and to identify any abnormalities present. In the presence of
abnormalities, they were asked to interpret what the abnormality meant. Participants were instructed to
be as specific as possible when describing the rhythm and abnormalities. Following the 30-minute time
period they were then asked to complete a short survey providing demographic details, and answering
questions related to their ECG teaching.
The ECGs used were selected from a pool of departmental teaching ECGs and were considered
representative of important ECG abnormalities. The selection was reviewed with general physicians,
experienced medical educators and cardiologists. The group defined a marking schedule that was
applied to each participant’s answers. For heart rate, a range of ±5 beats per minute for regular rates
below 100, ±10 beats per minute for regular rates between 100 and 150 beats per minute, and ±15 beats
per minute for heart rates greater than 150 beats per minute was allowed. For irregular heart rates the
above ranges were applied to the fastest and slowest heart rates on the ECG rhythm strip.
When assessing the interpretation of rhythm and abnormalities present, only complete answers were
considered correct (for example, atrial fibrillation was considered correct where as irregularly irregular
rhythm was not). Responses for each participant were assessed against the marking schedule by two
physicians.
Statistical analysis—For each participant we calculated the percentage of correct answers and
examined the relationship between this score and self-reported level of confidence using Pearson’s
correlation. We also compared the scores of the final year medical students with those of the house
surgeons, and the scores between the two study sites (Wellington and Auckland) using Mann Whitney
U Test. A p value <0.05 was considered statistically significant. Statistical analysis was performed
using SPSS v11.0 software (SPSS, Chicago, IL).
Results
All 102 subjects approached participated in the study. Of those surveyed, 49 were
male, 52 were final year medical students, and 50 were house officers. The results of
the interpretation of the study ECGs are presented in Table 1. All subjects completed
the first 8 ECGs, but only 73% were able to complete all 15 ECGs in the 30 minutes
allocated.
The ability to determine heart rate was highly variable, with responses ranging from
0% to 89% correct. Where heart rate was inaccurately determined, the tendency was
for it to be underestimated by 10–20 beats per minute. For example, in ECG 4, we
accepted answers between 85 and 95 beats per minute as correct, and the median
answer given was 75 beats per minute; and for ECG 8, we accepted answers in the
range 60–70 beats per minute, and the median answer given 50 beats per minute.

Heart rhythm accuracy ranged from 9% for ventricular pacing to 98% for ventricular
tachycardia. The eight ECGs showing sinus rhythm were identified with accuracy
ranging from 81% to 95%. Atrial fibrillation (55%) and atrio-ventricular conduction
disease (second degree heart block type 1, 19%) were poorly recognised.
Table 1. Description of each ECG, percentage of responders answering each
question, and the percentage of those correctly determining heart rate, rhythm,
abnormality, and interpreting abnormalities other than rhythm correctly if
present
ECG ECG Description Responders
(%)
Rate
(%)
Rhythm
(%)
Abnormality
(%)
Accurate
Interpretation (%)
1 SR, CHB with acute inferior
STEMI
100 85 72 87 54
2 SR with trifascicular block 100 48 90 22 3
3 Ventricular tachycardia 100 89 98 – –
4 SR with LBBB 100 0 87 44 44
5 Atrial flutter 100 79 59 – –
6 SR with acute antero-lateral
STEMI
100 10 93 93 46
7 Ventricular pacing 100 79 9 – –
8 SR with pre-excitation (Wolff-
Parkinson-White pattern)
100 3 95 11 11
9 SR with acute inferior STEMI 92 42 90 93 61
10 SR, 2nd
degree type 1AV block 90 73 19 – –
11 Supraventricular tachycardia 86 5 79 – –
12 SR with Long QT interval and
repolarisation abnormalities
80 11 89 7 7
13 Sinus Rhythm, normal ECG 77 18 81 – –
14 Atrial fibrillation with old
anterior Q wave MI
78 60 55 61 61
15 SR with acute ischaemic
antero-lateral ST depression
73 35 86 92 66
CHB = complete heart block, LBBB = left bundle branch block, MI = myocardial infarction,
SR = sinus rhythm, STEMI = ST elevation myocardial infarction.
There were 4 ECGs showing acute ischaemic ST segment changes (3 with ST
elevation and 1 with ST depression). These changes were identified correctly by
between 87% and 93% of those answering. The interpretation of the ST segment
changes was considerably less accurate, with correct answers in between 46% and
66% of cases. Conduction defects were poorly identified, although a more lenient
marking schedule that accepted any identification of a conduction abnormality would
have seen accuracy increase to approximately 80%.
Prolongation of the QT interval and pre-excitation (WPW pattern) were signs that
were not well detected despite the fact that the examples selected had marked
abnormalities. Also of note is that 34% of those who interpreted the normal ECG
(number 13) reported that an abnormality was present.
The majority (76%) of participants had reported that they had recorded less than 5
ECGs in the previous year. More than half (54%) had examined fewer than 50 ECGs

with 24% reporting that they had interpreted fewer than 10 ECGs in the previous year.
When asked how many ECGs they had discussed with a registrar or consultant, 87%
reported that they had discussed fewer than 50 with 38% reporting that they had
discussed fewer than 10 ECGs in the previous year.
Participant’s ratings of confidence in their ability to interpret ECGs, and of the quality
of their teaching are given in Figure 1. Participants reported that they had received a
mean of 6.3 hours of formal teaching on ECGs during their medical school training.
Figure 1. Participants rating of their own abilities to interpret ECGs (A) and the
quality of the teaching they had received (B)
A. Rating of own abilities in ECG interpretation
0
5
10
15
20
25
30
35
40
45
50
poor less than
satisfactory
satisfactory good excellent
Percentage
B. Rating of quality of ECG teaching
0
5
10
15
20
25
30
35
40
45
50
poor less than
satisfactory
satisfactory good excellent
Percentage

There was a positive correlation between levels of confidence and overall mark
(r=0.242, p=0.03). Final year medical students (median score 52%, range 20–65%)
scored significantly higher than house officers (40%, 18–61%, p=0.001). There was
no difference in score between the two centres in the study (Wellington median score
50%, range 18-65%, Auckland median score 48%, range 21–61%, p=0.67).
Discussion
The major finding of our study is that there are significant deficiencies in the ability
of final year medical students and resident medical officer to interpret common and
important ECG changes. However, the majority of participants were able to recognise
the most important rhythm disturbance, ventricular tachycardia, accurately as well as
identify ischaemic ST segment changes. Furthermore, nearly half of the respondents
rated their ability to interpret ECGs as less than satisfactory while just over half rated
the ECG teaching they had received as less than satisfactory.
Currently there are no clearly defined standards in ECG interpretation for medical
graduates for us to compare the results of this study against. A New York study of
residents in emergency medicine and internal medicine reported 93% accuracy in
identifying VT8
while an Australian study of emergency medicine trainees reported
40% accuracy in identifying VT.9
In this context, the 98% accuracy seen in this New
Zealand study is very good.
With respect to identifying signs of myocardial infarction, previous studies have
reported levels of accuracies of around 90% in registrars8,9
and a retrospective audit of
1684 patients with acute MI in 5 emergency departments in the United States reported
an overall accuracy of 88%.1
The level of accuracy for our medical graduates in
recognising ST segment changes (87–93%) is probably reasonable in comparison.
Clearly, failure to recognise ST segment changes could potentially lead to
inappropriate patient management. Similarly, there are concerns that simple
recognition of ST segment changes without any understanding of what these changes
mean may also lead to inappropriate management decisions being made.10
Given this,
we would have hoped to see considerably better performance in the interpretation of
ST segment changes in this study.
The recognition of pre-excitation and long QT syndrome we observed was poor, but
in Hoyle et al’s study of emergency trainees9
these were identified in only 24% and
43% of cases respectively so it is perhaps not surprising that more junior doctors and
medical students struggle with these ECGs. Again, failure to recognise these
abnormalities has the potential to result in inappropriate patient management.
It is difficult to account for the systematic inaccuracy in calculation of heart rate
observed in the present study. Perhaps heart rate is regarded as clinically unimportant,
and rounding down is viewed by the study participants as a reasonable approximation.
While it is difficult to argue that errors of the magnitude we observed in the current
study are likely to lead to errors in patient management it is none the less
disappointing that a fairly basic skill was not performed to a higher standard.

The time pressure and the lack of a clinical scenario may have detrimentally
influenced the overall level of accuracy. There is good evidence that accuracy is
improved when clinical vignettes are provided in addition to the ECG7
—although
conversely there is evidence that where ECG changes occur in the context of atypical
symptoms, inaccuracy increases.1
Previous studies do not appear to have limited time in any way.6,8,9
We chose to do so
on the basis that 2 minutes per ECG was considered a reasonable time period given
that in clinical practice there are time constraints.
Our medical graduates expressed a lack of confidence in their abilities in ECG
interpretation, and this was related to their accuracy levels. Previous studies of New
Zealand graduates in other areas have identified knowledge gaps and low levels of
confidence similar to those we report.11,12
Perhaps of slightly greater concern to those of us involved in the provision of ECG
teaching is that this was rated as less than satisfactory or poor by more than half the
study participants. It is possible that we have tended to assume that students interpret
and discuss a considerable volume of ECGs in the course of various clinical rotations.
However, our results suggest than more than half the study participants looked at less
than one ECG per week in a clinical context.
The fact that medical students performed better than the residents again suggests that
skills in ECG interpretation gained in medical school are not been built upon by
normal clinical duties or post graduate teaching. The teaching implications are that
more has to be done to assist the students to consolidate the teaching that they get, and
to provide additional teaching opportunities.
While some aspects of performance in the current study were encouraging, overall our
final year medical students and house officers do not achieve what we would consider
an adequate standard in recognising and interpreting important common or life-
threatening abnormalities in the ECG. Consistent with this, the majority of
participants rated both their own ability and the teaching they had received as less
than satisfactory.
To address this problem we need to define minimum standards in ECG interpretation,
improve our teaching to meet these standards, and assess our graduates against them.
Competing interests: None known.
Author information: Nigel A Lever, Senior Lecturer, Department of Medicine,
University of Auckland and Green Lane Cardiovascular Service, Auckland City
Hospital, Auckland; Peter D Larsen, Senior Lecturer, Department of Surgery and
Anaesthesia, University of Otago, Wellington; Mathew Dawes, Senior Lecturer,
Department of Medicine, University of Auckland, Auckland; Annie Wong, House
Surgeon, Wellington Hospital, Wellington; Scott A Harding, Cardiologist, Wellington
Hospital, Wellington
Correspondence: N A Lever, Green Lane Cardiovascular Service, Auckland City
Hospital, Private Bag 92024, Auckland 1030, New Zealand. Fax: +64 (0)9 3078941;
email: nlever@adhb.govt.nz

References:
1. Masoudi FA, Magid DJ, Vinson DR, et al. Implications of the failure to identify high-risk
electrocardiogram findings for the quality of care of patients with acute myocardial infarction:
results of the Emergency Department Quality in Myocardial Infarction (EDQMI) study.
Circulation. 2006;114:1565–71.
2. Todd KH, Hoffman JR, Morgan MT. Effect of cardiologist ECG review on emergency
department practice. Annals of emergency medicine. 1996;27:16–21.
3. Hurst JW. The interpretation of electrocardiograms: pretense or a well-developed skill?
Cardiology Clinics. 2006;24:305-7, vii.
4. Childers R. Teaching electrocardiogram interpretation. Journal of Electrocardiology.
2006;39:426–9.
5. Hurst JW. Thoughts about the abnormalities in the electrocardiogram of patients with acute
myocardial infarction with emphasis on a more accurate method of interpreting ST-segment
displacement: part I. Clinical Cardiology. 2007;30:381–90.
6. Pines JM, Perina DG, Brady WJ. Electrocardiogram interpretation training and competency
assessment in emergency medicine residency programs. Acad Emerg Med. 2004;11:982–4.
7. Salerno SM, Alguire PC, Waxman HS. Competency in interpretation of 12-lead
electrocardiograms: a summary and appraisal of published evidence. Annals of Internal
Medicine. 2003;138:751–60.
8. Berger JS, Eisen L, Nozad V, et al. Competency in electrocardiogram interpretation among
internal medicine and emergency medicine residents. The American Journal of Medicine.
2005;118:873–80.
9. Hoyle RJ, Walker KJ, Thomson G, Bailey M. Accuracy of electrocardiogram interpretation
improves with emergency medicine training. Emerg Med Australas. 2007;19:143–50.
10. Hurst JW. To simply identify "up or down" displacement of the ST segment in the
electrocardiogram can lead to serious errors. Journal of Electrocardiology. 2008 Sep-
Oct;41(5):436–7. Epub 2008 Mar 19.
11. Price CS, Bell SF, Janes SE, Ardagh M. Cardio-pulmonary resuscitation training, knowledge
and attitudes of newly-qualified doctors in New Zealand in 2003. Resuscitation. 2006;68:295–
9.
12. Subramaniam R, Hall T, Chou T, Sheehan D. Radiology knowledge in new medical graduates
in New Zealand. N Z Med J. 2005; 118(1224). http://www.nzmj.com/journal/118-1224/1699

Content

Recomendados

Recomendados

Mais conteúdo relacionado

Mais procurados

Mais procurados (20)

Destaque

Destaque (11)

Semelhante a Content

Semelhante a Content (20)

Mais de Yabniel Lit Jingga

Mais de Yabniel Lit Jingga (20)

Content