Correlation of multiple patient-reported outcome measures across follow-up in patients undergoing primary shoulder arthroplasty

Abstract

Multiple validated outcome scores are used to assess patients undergoing shoulder arthroplasty. The purpose of this study was to determine whether a correlation exists between 3 commonly used patient-reported outcome (PRO) measures in this population: Shoulder Pain and Disability Index (SPADI), American Shoulder and Elbow Surgeons (ASES) Assessment Form, and Simple Shoulder Test (SST). We performed a retrospective review of a shoulder arthroplasty database that routinely collects SPADI, ASES, and SST scores at each visit prospectively. Patients undergoing primary shoulder arthroplasty were identified. Assessments of correlation coefficients (Pearson correlation coefficient for ASES and SPADI scores and Spearman correlation coefficient for SST score) between each combination of PROs were performed overall and at each time point (preoperativelyand 3, 6, 12, and 24 months postoperatively) to determine the level of association between PROs. In total, 848 shoulder arthroplasty procedures were performed in 754 patients with 2796 unique clinical encounters. Preoperative correlations among PROs were moderate to strong (range, 0.66-0.77) but had the lowest correlation among all comparisons. Postoperative correlations were strong for all PRO comparisons (range, 0.73-0.94). Postoperative PRO correlations continued to strengthen over longer follow-up, with all values exceeding 0.78 at 2 years postoperatively. Conversion equations between PROs were calculated based on these highly correlated data. After primary shoulder arthroplasty, there exists a high degree of correlation among all 3 studied PROs. Correlations were stronger postoperatively and improved with longer follow-up. Surgeons may use this information to minimize the number of questionnaires patients answer at each follow-up time point, and the conversion equations can be used for study comparison in meta-analyses.