Level of practice for ECG interpretation skills should be ‘expert’

See also pp. 143–150 In this edition of EMA Hoyle, Walker and Thomson study ECG interpretation among a large cohort of Victorian ACEM trainees. The main question ‘does ECG interpretation accuracy improve with advancing years of emergency training?’ is supplemented by further questions that surround the exposure to, type of, and satisfaction with ECG diagnostic training.1 It is comforting that ECG interpretation skills do seem to improve with training as shown by comparing ‘senior trainees’ (n = 48) in their last 2 years of advanced training with ‘other trainees’ (n = 74). It should be noted that the ‘other trainees’ comprise of mainly provisional and basic trainees; 53 of 74 (71.6%). These trainees are clearly junior in terms of ACEM training but also might represent a diverse group. Doctors are eligible for provisional training in their third postgraduate year but some doctors might have decided on, or swapped to, Emergency Medicine (EM) at a later stage. Some provisional trainees might have clearly decided to train in EM whereas some have enrolled with ACEM considering their options. Many are preparing for or have just completed their primary examination and consequently might not have ECG interpretation skills as a priority. So in many ways it is not surprising that ‘senior trainees’ did better in this study but interestingly the findings were independent of postgraduate years of medical practice. Not so comforting is that the overall accuracy for ‘senior trainees’ was 67.5%. In particular, life-threatening arrhythmias such as ventricular tachycardia (VT) and ventricular fibrillation (VF) were 43.8% and 70.8%, respectively. This is despite the new fellowship exam (FE) curriculum under section 4.1 k (Medicine; Cardiovascular: Disturbances of cardiac rhythm) designating the level of practice (LP) required as ‘expert’ for almost all of the many items linked to ECG interpretation skills. Expert for clinical topics and their learning objectives have been defined as: ‘able to demonstrate the knowledge and skills required by the objective at a level necessary to provide all aspects of acute care without the need for consultation’.2 The reasons for this designated LP should be obvious to every trainee and emergency physician (EP). Some confounders to the study results might apply. The ECG not being put into a clinical context is one. The pros and cons of this have been well addressed in the article. With respect to the VF ECG the reader might wonder whether the additional presence of an AMI could confuse trainees as to what was ‘the main diagnosis’, the question that was asked in regard to each ECG. The authors indicate that any mention of VF was taken as correct and that the most common incorrect answers were VT and torsades, not AMI. Another confounder might have been that a rhythm strip, which is a standard feature in the clinical setting, is not shown. So should we be surprised that many of our trainees find ECG interpretation difficult? I would argue probably not for a number of reasons. First, the literature supports that junior doctors require improved ECG interpretation skills. Hoyle quotes a number of articles that illustrate diagnostic variance in residents and non-cardiologist physicians and in particular comment on problems distinguishing wide complex tachycardias.3–8 Various studies have addressed the difficulties with interpretation of ST elevation (STE) syndromes. The combined percentage accuracy for pericarditis of 48% in this study is low but not completely dissimilar to other studies. Brady et al. investigated diagnostic accuracy by American EP in the setting of chest pain and 11 STE syndromes. Forty-two per cent (193) of these EP were said to be in their second or third postgraduate year, therefore actually trainees; and 58% (265) were attending. The following ECG were interpreted correctly 100% of occasions: AMI with typical STE; right bundle branch block (RBBB); RBBB with AMI and ventricular paced rhythm. The following ECG were misdiagnosed in the percentage of cases allocated with the second percentage in parenthesis indicating the degree of inappropriate thrombolysis (TL) for non-AMI syndromes: left ventricular aneurysm (72%; TL 28%), AMI with atypical STE (63%), benign early repolarization (53%; TL 23%), pericarditis (37%; TL 21%), left ventricular hypertrophy (13%), left bundle branch block with AMI (11%) and left bundle branch block without AMI (11%; TL 5%). The article does not address the trainee to attending differences.9 Another interesting American study compared ECG interpretation and management skills in 31 internal medicine (IM) residents with 31 EM residents. Programme residents were equally likely to misinterpret left ventricular hypertrophy (23% IM vs 16% EM) and benign early repolarization (48% IM vs 52% EM). The diagnosis and management of pulseless VT, unstable supraventricular tachycardia, complete heart block and posterior AMI were also examined. They concluded that the majority of both residents diagnosed and managed complete heart block and pulseless VT well but EM residents performed better with unstable supraventricular tachycardia and posterior AMI.10 Second, how do our senior trainees perform in examinations with ECG interpretation? I went back and had a look at the last 10 FE, January 2002–February 2006, to specifically research this question.11 The visual aid questions (VAQ) provide a particularly good section for this purpose. Eight questions are examined in 60 min with an additional 10 min of reading time. Questions can be categorized as being related to an ECG, radiology, clinical image or investigational visual aid. Over the 10 years only one question did not fit any of these categories. ECG questions just about always have a first question of describing and interpreting the ECG and are followed by some type of assessment or management question. The overall pass rate for VAQ excepting those that involved ECG was 78.2% (3245 passes of 4146 questions attempted by 634 candidate sittings; 65 individual questions). Fifteen individual ECG questions, as outlined in Table 1, were examined with a significantly different lower pass rate of 71.7% (Fishers exact test P < 0.0001; 95% confidence interval for the difference in proportion 3.5–9.8% calculated by Wilson’s method; 664 passes of 926 questions attempted by 634 candidate sittings). The statistics quoted are somewhat crude as the methods assume independence, do not account for the same individuals answering questions in both groups and cannot reflect individual-specific candidates performing less well with ECG questions. Candidates who have sat on more than one occasion contribute to these statistics also. Other category pass rates include radiology 78.9%, clinical images 76.6% and investigations 70.2%. ECG and investigation interpretation seem to be performed less well despite potential problems with the analysis. A total of 262 ECG VAQ questions were failed. Examiner comments about reasons for failure in VAQ exam reports include the following: ‘The examiners felt that this question was not well done by candidates overall, especially as this was regarded as a straight forward ECG’; ‘Failing answers misinterpreted the ECG, including being unable to estimate the rate’; ‘Failures were due to basic errors of description’; ‘Candidates who failed did not appreciate the significance of the ECG changes’; ‘some candidates were not able to appreciate the extent of the infarct’; ‘failures occurred due to … the unstructured approach to the ECG with VT’; ‘candidates failed to pick up pacemaker spikes and/or pacemaker dysfunction’; and ‘most common reason for failure was inability to diagnose sinus bradycardia – candidates diagnosed as CHB or 2:1 HB’. These comments are clearly not applicable to the majority but certainly significant numbers of trainees sitting our FE in the past years have displayed insufficient knowledge and skills in this area. Table 2 summarizes the Structured Clinical Examination (SCE) section. This examination involves six 10 min stations that proceed in an evolving scenario cross-table oral format and might involve clinical props or skills testing. Eleven individual SCE involved some type of ECG interpretation with an overall pass rate of 83.1% (466 passes of 561 questions attempted by 494 candidate sittings). This is compared with 49 SCE not involving ECG with an overall pass rate of 87% (Fishers exact test P = 0.017; 95% confidence interval for the difference in proportion 0.7–7.5% by Wilson’s method; 2090 passes of 2403 questions attempted by 494 candidate sittings). Higher pass rates for the SCE orals reflect the fact that those candidates who have not passed at least two or three written examinations are not invited to the orals. It could also be argued that passing an ECG SCE, as compared with an ECG VAQ, is more complex and less reliant on just ECG description. This is particularly the case when 50% of an ECG VAQ mark might depend on the ‘description’ and interpretation component alone. Finally, my own personal anecdotal evidence backs up the concept that many trainees find ECG interpretation difficult. This includes my experience of teaching in a large regional training programme, running ECG workshops at the New Zealand annual pre-FE course, and acquiring ECG interpretation skills over many years via clinical encounters and a special interest. I would imagine many of my colleagues have had similar experiences. It is concerning that ‘most trainees reported being dissatisfied with their ECG training’. Our trainees certainly require specific teaching and experiential learning in this area. Adult learning theory applies and there should be a partnership between fellows and trainees. Fellows, especially Directors of EM Training (DEMT), in accredited departments have an obligation to teach. At the same time trainees have an obligation to seek and facilitate teaching and experiences that will acquire the LP required. The Hoyle article suggests a number of viable options for improving trainee education and departmental teaching.1 Trainees, DEMT and departments are certainly encouraged to collect a variety of ECG that could be used for teaching and FE preparation. FE preparation ECG and questions clearly need to be pitched at consultant level. Identification of the types of syndromes and ECG that could apply to this should be given in-depth consideration. So what should ACEM do? A specific ECG textbook has been recently introduced into the recommended textbook list over and above the general comprehensive EM texts.12 The FE curriculum clearly identifies the expected learning objectives and LP. Should ACEM have an ECG course or online module or even an earlier specific examination? Perhaps, but logistics and the subsequent begging questions ‘Well then do we have to have a module for every aspect of EM?’ and ‘Why is ECG interpretation any more important than trauma or respiratory EM or any other aspect of EM?’, will need to be addressed. Certainly an Australasian ECG resource or text would be useful. In the meantime our exit FE, and hopefully in-training assessments supported by DEMT courses, will reflect the competence required and the LP of ‘expert’. Apparently assessment drives learning! I am currently the Censor-in-Chief of ACEM and an ACEM examiner. Information about the FE from FE reports has been done so with the consent of the FE committee chair. Opinions expressed in this article are my own and not those necessarily embraced by the ACEM, Board of Censors, Court of Examiners or FE committee.