Last month Academic Emergency Medicine published the article “Can emergency physician gestalt ‘rule in’ or ‘rule out’ acute coronary syndrome: validation in a multi-center prospective diagnostic cohort study”. The following is a critique by Dr. deSouza and Dr. Sinert:
In this large multi-center, prospective diagnostic accuracy study, the investigators conclude that gestalt (including ECG interpretation) is not sufficient to exclude acute MI (AMI) in patients who present to ED with suspected ACS. However, we feel the need to call attention to some issues in the study methods that may limit the strength of its findings.
First, Oliver et al (1) did not uniformly blind clinicians to the initial troponin level reasoning that this method may be a more pragmatic representation of gestalt in practice. However, an elevated troponin level was included in the diagnostic criteria for AMI, so physician knowledge of the troponin level will significantly bias gestalt (incorporation bias). If data were available, it would be useful to calculate the accuracy of gestalt from assessments during which clinicians were appropriately blinded to the initial troponin result. The accuracy of gestalt with an ECG only (without troponin result) would be informative considering the non-specificity of troponin assays and the downstream consequences of positive results in patients in whom no acute coronary occlusion is ultimately identified.
Second, Oliver et al (1) did not consider (or adjust for) physician experience in their analysis, yet sufficient clinical experience is necessary to develop gestalt cognition (2). This may have had numerous consequences. The authors appropriately state that “our findings do not apply to patients who did not pass the clinicians’ pre-test probability threshold for warranting investigation for ACS”. However, the experience of the clinician may impact his or her pre-test probability threshold. This threshold may be based upon confidently identifying an alternative diagnosis, and less experienced clinicians may be less equipped to do so. For example, in a similar study that did stratify the relationship of gestalt to AMI prevalence by clinician experience (and also blinded clinicians to initial troponin level), Body et al (3) reported that only less experienced (< 7 years) physicians initiated testing despite recording their gestalt as “definitely not ACS”. This relatively lower threshold for testing would result in the inclusion of a proportion of patients with a lower overall AMI prevalence and thus limit clinical applicability.
In Oliver et al (1) it is possible that a number of less experienced clinicians may have similarly overestimated pre-test probability for AMI, consequently inflating the Likert scale assessments. On the other end of the spectrum, Body et al (3) reported that experienced physician gestalt (gestalt of physicians with at least seven years of postgraduate experience) may be relatively more specific for the diagnosis of AMI, although notably, their data in each “years of experience” stratum was limited. Therefore, the participation of less experienced clinicians in both trials (1,3) may have resulted in a lower AUROC and diminished the reported accuracy of gestalt to predict acute AMI.
In conclusion, the available current evidence still does not have an answer to the question: Can a clinician with sufficient clinical experience use gestalt and an ECG alone to accurately exclude AMI in a patient with suspected ACS?
1. Oliver G, Reynard C, Morris N, Body R. Can emergency physician gestalt “rule in” or “rule out” acute coronary syndrome: validation in a multi-center prospective diagnostic cohort study. Acad Emerg Med 2019.
2. Cervellin G, Borghi L, Lippi G. Do clinicians decide relying primarily on Bayesians principles or on Gestalt perception? Some pearls and pitfalls of Gestalt perception in medicine. Intern Emerg Med 2014;9:513-9.
3. Body R, Cook G, Burrows G, Carley S, Lewis PS. Can emergency physicians ‘rule in’ and ‘rule out’ acute myocardial infarction with clinical judgement? Emerg Med J 2014;31:872-6.
Charles Murchison
Latest posts by Charles Murchison (see all)
- Benign Early Repolarization vs. Anterior STEMI - February 20, 2020
- Can experienced clinician gestalt + ECG accurately exclude acute MI? - August 20, 2019
- Does Observation for ACS Make Sense? Part 3: Risk Stratifying for Adverse Cardiac Events - June 20, 2019
0 Comments