Patient-reported outcomes measurement information system depression psychometrically underperforms compared to legacy measures and is poorly associated with postoperative functional outcomes in shoulder arthroplasty patients

BackgroundThe purpose of this study was to investigate the associations between each mental health patient-reported outcome measure with postoperative functional outcomes following shoulder arthroplasty, and to compare psychometric properties of patient-reported outcomes measurement information system depression to the legacy (VR-12 Mental) patient-reported outcome measure.MethodsPatients who underwent primary shoulder arthroplasty from July 2018 to February 2019 were retrospectively reviewed. Patient-reported outcomes measurement information system depression and VR-12 Mental were administered preoperatively; American Shoulder and Elbow Surgeons and Single Assessment Numeric Evaluation were administered at 6-month and 1-year postoperatively. Rasch partial credit modeling analysis was used to compare psychometric properties of legacy versus patient-reported outcomes measurement information system instruments in assessing mental health.ResultsNinety-three patients who underwent total shoulder arthroplasty (n = 52), reverse total shoulder arthroplasty (n = 39), or hemiarthroplasty (n = 2) were included. Preoperative VR-12 Mental scores were moderately associated with American Shoulder and Elbow Surgeons at 6-months (coefficient: 0.52, P = 0.026) and 1-year (coefficient: 0.65, P = 0.002), while preoperative patient-reported outcomes measurement information system depression scores were not. Patient-reported outcomes measurement information system depression demonstrated significant floor effects (16%); VR-12 Mental demonstrated minimal floor and ceiling effects (1.1% for both). VR-12 Mental demonstrated broader coverage of mental outlook on Rasch modeling than patient-reported outcomes measurement information system depression and had adequate model fit after one round of reiterative item elimination.DiscussionPatient-reported outcomes measurement information system depression was poorly associated with postoperative American Shoulder and Elbow Surgeons scores, demonstrated significant floor effects, and had limited coverage of mental health on Rasch modeling with reiterative elimination.Level of EvidenceIV

Automated summary

The study found that VR-12 Mental scores were moderately associated with American Shoulder and Elbow Surgeons scores at 6 months and 1 year postoperatively, while PROMIS depression scores were not. VR-12 Mental had broader coverage of mental outlook and better model fit in Rasch modeling compared to PROMIS depression.