Determining the learning curve of transcutaneous laryngeal ultrasound in vocal cord assessment by CUSUM analysis of eight surgical residents: when to abandon laryngoscopy

Abstract

Transcutaneous laryngeal ultrasonography (TLUSG) is a promising alternative to laryngoscopy in vocal cords (VCs) assessment which might be challenging in the beginning. However, it remains unclear when an assessor can provide proficient TLUSG enough to abandon direct laryngoscopy. Eight surgical residents (SRs) without prior USG experience were recruited to determine the learning curve. After a standardized training program, SRs would perform 80 consecutive peri-operative VCs assessment using TLUSG. Performances of SRs were quantitatively evaluated by a composite performance score (lower score representing better performance) which comprised total examination time (in seconds), VCs visualization, and assessment accuracy. Cumulative sum (CUSUM) chart was then used to evaluate learning curve. Diagnostic accuracy and demographic data between every twentieth TLUSG were compared. 640 TLUSG examinations had been performed by 8 residents. 95.1 % of VCs could be assessed by SRs. The CUSUM curve showed a rising pattern (learning phase) until 7th TLUSG and then flattened. The curve declined continuously after 42nd TLUSG (after reaching a plateau). Rates of assessable VCs were comparable in every twentieth cases performed. It took a longer time to complete TLUSG in 1st-20th than 21st-40th examinations. (45 vs. 32s, p = 0.001). Although statistically not significant, proportion of false-negative results was higher in 21st-40th (2.5 %) than 1st-20th (0.6 %), 41st-60th (0.7 %), and 61st-80th (0.7 %) TLUSG performed. After a short formal training, surgeons could master skill in TLUSG after seven examination and assess vocal cord function consistently and accurately after 40 TLUSG. © 2015 Société Internationale de Chirurgie.