Evaluation of the learning curve for reverse shoulder arthroplasty

Abstract

This purpose of this study was to quantify the surgical learning curve and provide guidelines to surgeons interested in teaching and performing reverse shoulder replacement. Sixty-two consecutive primary reverse shoulder replacements performed by a single surgeon were retrospectively reviewed. Using data from consecutive cases, surgical time was plotted against patient case order, and the linear regression slope was calculated. Case length slope analysis demonstrated a significantly negative slope with the first 18 cases, which subsequently leveled thereafter. The number of cases needed to arrive at this flat slope was defined as the proficiency point. The proficiency point was then verified using the measurable variables of baseplate screw number and glenosphere overhang by dividing the series into 2 groups: Group A included patients who had surgery prior to the proficiency point, and group B included patients who had surgery after the proficiency point. In group A, only 33% (6/18) had all 4 glenoid baseplate screws placed as compared to 66% (29/44) in group B (P=.02). Glenosphere overhang increased from a mean of 1.02 mm (±1.29 mm) in group A to 2.58 mm (±1.89 mm) in group B (P=.003). A learning curve of approximately 18 cases was found in this series based on the technical aspects of performing reverse shoulder replacement. This curve is likely even longer for lower volume shoulder surgeons. We recommend specialized training prior to performing this procedure.