Do We Need to Reconsider How We Gauge Success After Anatomic Total Shoulder Arthroplasty? A Study of Thresholds Optimized for Patient Satisfaction Using the Simple Shoulder Test.

Background: Clinically important improvement after total shoulder arthroplasty is often assessed with shoulder-specific patient-reported outcome measures (PROMs) quantifying reduction in pain and restoration in function. It is unclear if commonly utilized threshold such as minimal clinically important difference (MCID), substantial clinical benefit (SCB), or patient acceptable symptom state (PASS), represent optimal improvement from the patients' perspective. The objectives of this study were to utilize the Simple Shoulder Test to: 1) compare commonly utilized thresholds for change in score and final score to thresholds optimized to patient satisfaction using receiver operative characteristic (ROC) curve analysis, and 2) determine the impact of using different thresholds on reporting of independent predictors of successful outcome in terms of patient satisfaction.

Methods: This study included 406 anatomic total shoulder arthroplasty (aTSA) patients from a longitudinally maintained database with two-year follow-up. Surveys included the Simple Shoulder Test (SST) and a satisfaction questionnaire. SST thresholds were calculated by commonly described techniques for MCID, SCB, %MPI (percentage of maximal possible improvement), and PASS. ROC curves were constructed to determine the optimal threshold of SST change in score (Change-ROC), final SST score (Final-ROC), and %MPI (%MPI-ROC) based on patient satisfaction. Youden's index (J) was calculated to determine each threshold's performance in maximizing sensitivity and specificity. Multivariable analysis was performed to determine predictors of surpassing selected threshold values.

Results: The thresholds with the highest Youden's index correlating best with patient satisfaction were %MPI-ROC (aTSA 61%, J=0.49; RnR 63% J=0.64) and Final-ROC (aTSA 9.5, J=0.48; RnR 9.5 J=0.60). Commonly utilized thresholds including MCID of 2.0 (J=0.21) and SCB of 2.7 (2.7) had the lowest Youden's index of the thresholds studied. Characteristics predictive of success varied substantially based on selected threshold.

Conclusions: Current thresholds commonly utilized to gauge success after aTSA have limited ability to predict success based on patient satisfaction using the Simple Shoulder Test. Given that focus in healthcare value is shifting towards patient satisfaction, optimal thresholds utilized to measure success after shoulder arthroplasty may require reconsideration.