(P068) COMMUNITY-TO-INSTITUTION ATHLETIC CARDIOVASCULAR SCREENING: VALIDATION OF AN ELECTROCARDIOGRAM WORKFLOW MODEL

Background: To prevent sudden cardiac death in young athletes, many sporting organizations and institutions require that athletes undergo pre-participation screening (PPS). The inclusion of electrocardiogram (ECG) testing in PPS remains controversial due to cost implications, resource requirements, and the potential for inaccurate interpretations. At most centres, ECGs are performed internally by health care professionals trained in the interpretation of athlete-specific ECG findings. Since 2017, Queen’s University has implemented a PPS workflow model which outsources ECG requisitions to the local care networks of athletes undergoing screening, thus reducing institutional costs and resource expenditure. The objective of this study was to evaluate the accuracy of ECG interpretation within this workflow model by comparing interpretations between community physicians (i.e. any physician outside of Queen’s University) and an institutional sports cardiologist.

METHODS AND RESULTS: This was a retrospective, single-centre observational study of all athletes undergoing cardiovascular PPS at Queen’s University between 2017-2021 (n=1,100). A total of 740 athlete ECGs met the inclusion criteria and were reinterpreted by a sports cardiologist using the International Criteria athletic ECG recommendation (gold-standard). Percent agreement and Cohen’s kappa statistic were used to compare community physician ECG interpretations with gold-standard interpretations. Among this sample of young athletes (mean age: 18.5 years, female: 39.6%), a self-reported history of syncope (12.7%), angina (1.5%), dyspnea (1.3%), and familial cardiac complications (5.2%) were relatively uncommon (Table). A total of 181 unique community physicians performed the initial ECG interpretations. The proportion of ECGs interpreted as abnormal/borderline by community physicians and an institutional sport cardiologist was 5.5% and 0.14%, respectively (p < 0.01). The most common ECG finding interpreted as abnormal/borderline by the community physicians but ultimately deemed as normal by the sport cardiologist was right axis deviation (n=8). There was a 94.6% agreement in ECG interpretation between the two groups, correlating to a slight agreement in interobserver reliability (k = 0.045, 95% CI: -0.040 to 0.131). All ECG interpretation disagreements were due to false-positive interpretations by community physicians.

Conclusion: Community physicians interpret athletic ECGs with a high rate of agreement to a gold-standard interpreter. However, false-positive interpretations are more prevalent among community physicians, which may prompt unnecessary follow-up testing. These findings suggest that a PPS workflow model that outsources ECG requisitions to the community may be a reliable approach to PPS, all while reducing institutional costs and resource requirements associated with screening.