https://youtu.be/ccg0hX-FKqM INTRODUCTION The use of electrocardiograms (ECGs) in preparticipation evaluations (PPE) for athletes remains a topic of debate in the United States. Over the years, athlete-specific ECG criteria have evolved, leading to greater specificity and reduced false positive rates. Recent comparisons have been made between local physicians and those from specialized centers in their consistency of ECG interpretation using the 2017 International Criteria. However, there has been limited research on whether novice ECG interpreters, such as undergraduate students, can be trained to accurately interpret athletes’ ECGs using the 2017 International Criteria and further, be able to identify the abnormal condition seen on the ECG. This study aims to evaluate the overall and individual inter-rater reliability of novice ECG interpreters and exercise physiologist when compared to a cardiologist. METHODS Three novice ECG interpreters (undergraduate exercise science students) received training in interpreting athlete ECGs using the 2017 International Criteria over the course of one semester under the guidance of an expert reader. During an annual high school sports screening event, 1,350 ECGs were collected and assigned unique identification numbers. After the screening event, three novice interpreters (students), a cardiologist, and an Exercise Physiologist (exercise science professor) independently classified the ECGs as either "normal" or "abnormal" and identified the findings on abnormal ECGs based on the International Criteria. All participants were blinded to cardiologist classifications made during the screening event. Information on the athlete’s age, gender, race/ethnicity, and sport was included with the ECGs. Overall agreement between the cardiologist, exercise physiologist, and three students was assessed using a Fleiss’ kappa analysis and individual pair agreement was calculated using Crosstabs. RESULTS 1,350 athlete ECGs (males = 879; females = 471, age (mean + SD) 15.09 + 1.3y) were reviewed and 51 (3.8%) abnormal cases were identified and further diagnosed. The overall inter-rater agreement between a cardiologist, an exercise physiologist, and three student readers in classifying an ECG as normal or abnormal was rated as good (k = 0.711, p < .001). Individual agreement between which ECGs were classified as normal ranged from moderate, (k = .585) to very good (k = .845). Abnormal readings were further analyzed, and overall abnormal diagnosis agreement was rated as moderate (k = .432, p = 0.00) with individual agreement on abnormal diagnosis ranging from poor (k = .196) to very good (k = .851). CONCLUSION This study demonstrated that novice ECG readers could correctly classify ECGs based on the International Criteria as “normal and abnormal” but may need further training in accurately identifying ECG abnormalities in athletes.
Read full abstract