Study population and data collection
We enrolled all patients at a large academic echocardiography laboratory referred for TEE between the dates of 10/16/2017 and 11/22/2017 in which one of the three following TEE protocols was chosen by the physician/operator: aortic stenosis (AS), atrial fibrillation (AF), or mitral regurgitation (MR). The protocol was chosen by the primary operator (physician) based on the clinical indication for the TEE. While all patients did not have an AS, AF, or MR indication for the exam, the primary operator chose the TEE protocol that would best answer the clinical indication for the exam. All TEE exams were performed on the same echocardiogram (Siemens ACUSON SC2000, Mountain View, CA) using customized procedure-specific study protocol (Siemens eSie ScanTM workflow protocol software, Mountain View, CA). TEE procedural data, including exam duration, and sonographer/machine keystrokes were prospectively collected at the time of the TEE exam. Patient demographics and baseline data were obtained on subsequent retrospective chart review.
TEE image acquisition was performed for each TEE exam/patient using both a manual and an automated protocol image acquisition by the same sonographer and operator (physician). The physician operator was responsible for TEE probe insertion and manipulation (moving the TEE probe proximally and distally within the esophagus; ante-flexion, retro-flexion, sideways movement of the TEE probe; clockwise or counterclockwise rotation of the TEE probe). The sonographer was responsible for the operation of the echocardiogram cart/machine, including the use of all adjustments on the machine console and image acquisition. A third researcher was responsible for videotaping of the echocardiogram machine counsel during each exam for subsequent analysis of sonographer keystrokes and procedure duration.
Performing each protocol (manual and automated) for each patient allowed for control of patient-specific factors that may influence exam duration or complexity. The order for which type of protocol was performed in each patient (manual first or automated first) was alternated between patient exams ensuring that the same number of patients had the manual protocol performed first and the automated protocol performed first. The same sonographer was used for every acquisition with each operator. Operators were categorized as a novice (<3 years of TEE experience) or expert (>3 years TEE experience). At no point did any TEE exam need to be stopped or truncated due to patient or clinical factors. If additional, more complex, and non-standard views/images outside of the pre-specified protocol views were required, these views were obtained after the protocol was complete and were not included in analysis. No adverse events occurred during any of the TEE exams performed.
Statistical power analysis was performed to determine the sample size; 20 patients per cohort were required to detect a difference of\(2\pm 3\) minutes in mean exam durations between the two cohorts with 95% confidence interval and 80% power.