Comparisons of surgeon and patient prediction of 1-year outcomes following rotator cuff surgery

Background Preoperative patient expectations are thought to be predictors of outcomes in rotator cuff repair (RCR) surgery; however, these expectations lack an objective component. Surgeons are thought to possess the ability to more accurately predict postoperative patient outcomes compared to patients themselves. Surgeon's predictions of outcome could potentially serve to set realistic patient expectations of surgery and thus optimize surgical outcomes. The objective of this study was to describe patient and surgeon predictions of outcomes in those undergoing RCR surgery and determine the degree of agreement between these predictions and actual one-year postoperative outcomes. Methods Data for this study were collected in a healthcare registry as standard of care for all patients undergoing RCR from January 1 to December 31, 2022 at a single surgical center. Surgeries were conducted by fellowship-trained upper extremity surgeons. The primary outcome was the SANE (Single-Assessment Numeric Evaluation) score which required a written response by the patient to the question, "How would you rate your affected shoulder today as a percentage of normal (0% to 100% with 100% being normal)?". Preoperatively, patients completed their current SANE score and the SANE score they predicted to have at one-year postoperatively. The surgeon was asked to predict the patient's one-year SANE score immediately following completion of the surgery. A repeated-measures ANOVA compared surgeon-predicted, patient-predicted, and actual mean one-year postoperative SANE. Tukey's post hoc test was used to adjust for pairwise comparisons. The differences between actual and surgeon predicted SANE and actual and patient predicted SANE were calculated. "Accurate prediction" was defined a priori as a difference being within +/-5% of the actual SANE. Significance level was set to p <0.05. Results Sixty-nine patients were included in this study with a mean age of 60.9 (SD=6.9) years and 77% (N=53) were male. Surgeon-predicted one-year postoperative SANE was 77.6% (Min=60%; Max=95%; SD=6.8%), patient-predicted was 88.7% (Min=50%; Max=100%; SD=11.1%), and actual SANE was 80.6% (Min=8%; Max=100%; SD=19%) (p<0.001). Mean patient-predicted scores were significantly higher than both mean surgeon-predicted (p<0.001) and mean actual postoperative SANE (p=0.002). There was no statistical difference between mean surgeon-predicted and mean actual postoperative SANE. Conclusion On average, surgeons were better able to predict outcomes compared to patients. Patients tended to overestimate their outcomes, while surgeons tended to underestimate patient outcomes. These findings raise the importance of preoperative patient counselling to set more realistic expectations that would potentially improve their achieved outcomes.