BACKGROUND: Barriers to incorporation of simulation in cardiothoracic surgery training include lack of standardized, validated objective assessment tools. Our aim was to measure interrater reliability and internal consistency reliability of a coronary anastomosis assessment tool created by the Joint Council on Thoracic Surgery Education. METHODS: Ten attending surgeons from different cardiothoracic residency programs evaluated nine video recordings of 5 individuals (1 medical student, 1 resident, 1 fellow, 2 attendings) performing coronary anastomoses on two simulation models, including synthetic graft task station (low fidelity) and porcine explant (high fidelity), as well as in the operative setting. All raters, blinded to operator identity, scored 13 assessment items on a 1 to 5 (low to high) scale. Each performance also received an overall pass/fail determination. Interrater reliability and internal consistency were assessed as intraclass correlation coefficients and Cronbach's α, respectively. RESULTS: Both interrater reliability and internal consistency were high for all three models (intraclass correlation coefficients = 0.98, 0.99, and 0.94, and Cronbach's α = 0.99, 0.98, and 0.97 for low fidelity, high fidelity, and operative setting, respectively). Interrater reliability for overall pass/fail determination using κ were 0.54, 0.86, 0.15 for low fidelity, high fidelity, and operative setting, respectively. CONCLUSIONS: Even without instruction on the assessment tool, experienced surgeons achieved high interrater reliability. Future resident training and evaluation may benefit from utilization of this tool for formative feedback in the simulated and operative environments. However, summative assessment in the operative setting will require further standardization and anchoring.
BACKGROUND: Barriers to incorporation of simulation in cardiothoracic surgery training include lack of standardized, validated objective assessment tools. Our aim was to measure interrater reliability and internal consistency reliability of a coronary anastomosis assessment tool created by the Joint Council on Thoracic Surgery Education. METHODS: Ten attending surgeons from different cardiothoracic residency programs evaluated nine video recordings of 5 individuals (1 medical student, 1 resident, 1 fellow, 2 attendings) performing coronary anastomoses on two simulation models, including synthetic graft task station (low fidelity) and porcine explant (high fidelity), as well as in the operative setting. All raters, blinded to operator identity, scored 13 assessment items on a 1 to 5 (low to high) scale. Each performance also received an overall pass/fail determination. Interrater reliability and internal consistency were assessed as intraclass correlation coefficients and Cronbach's α, respectively. RESULTS: Both interrater reliability and internal consistency were high for all three models (intraclass correlation coefficients = 0.98, 0.99, and 0.94, and Cronbach's α = 0.99, 0.98, and 0.97 for low fidelity, high fidelity, and operative setting, respectively). Interrater reliability for overall pass/fail determination using κ were 0.54, 0.86, 0.15 for low fidelity, high fidelity, and operative setting, respectively. CONCLUSIONS: Even without instruction on the assessment tool, experienced surgeons achieved high interrater reliability. Future resident training and evaluation may benefit from utilization of this tool for formative feedback in the simulated and operative environments. However, summative assessment in the operative setting will require further standardization and anchoring.