Effectiveness and cost-effectiveness of Reverse Shoulder Arthroplasty versus Hemiarthroplasty versus Non-surgical care for acute 3 and 4 part fractures of the proximal humerus in older adults - The PROFHER-2 Randomised Clinical Trial

DESIGN: Multi-centre, open-label, 3-arm parallel group, pragmatic, randomised controlled trial with internal pilot to determine if Reverse Shoulder Arthroplasty (RSA) is more clinically & cost effective than Hemiarthroplasty (HA) in treating complex fractures of the proximal humerus; & to determine if either surgery is more effective than non-surgical care (NS) SETTING: Secondary care NHS hospitals TARGET POPULATION Inclusion: Consenting patients, 65 years of age or older, with a displaced multifragmentary (more than 2 parts) fracture presenting within 3 weeks of injury Exclusion: Patients not fit for anaesthesia; pathological (other than osteoporotic) fractures; lack of mental capacity to understand the trial or instructions for rehabilitation OUTCOME MEASURES Primary outcome is difference in mean OSS between RSA & HA 2 years after surgery. Difference in OSS between surgical (RSA or HA) & non surgical treatment will also be determined. Secondary outcomes include differences in EQ-5D scores; pain & range of motion. Data on perioperative care (including details of anaesthesia), physiotherapy, complications of surgery, further procedures & adverse events will be collected. SAMPLE SIZE: To detect a 5-point difference in OSS between surgical arms & a 6-point difference between either surgical arm & NS with 90% power, a total of 380 participants are required, assuming a SD of 12, 5% 2-sided statistical significance & 15% attrition at 2 years. RANDOMISATION: Randomisation will be imbalanced with a ratio of 2:2:1, favouring recruitment to RSA & HA arms. It will be done by York Trials Unit's secure, web-based, randomisation service ensuring adequate allocation concealment. PROJECT PLAN: The study will take 72 months, recruiting over 36 months at around 35 sites. Participant follow-up will be at 6 months, 1 year and 2 years. There will be a 12-month internal pilot phase from start of recruitment, by the end of which we aim set up 50% of centres, recruiting on average one participant per centre every 2 months to confirm recruitment assumptions & progress the trial accordingly. An extension phase for remote follow-up to 5 years will be considered based on findings of the main study. This position has been strengthened by the support for this application provided by the National Joint Registry. STATISTICAL ANALYSIS: Primary analysis will assess OSS scores by longitudinal regression analysis up to 2 years follow-up for three group comparisons: RSA vs HA, RSA vs NS & HA vs NS, adjusting for relevant baseline characteristics. Surgeons will be added as a random effect to account for individual differences. Potential learning effects for surgical techniques will be explored. Secondary outcomes will be analysed by longitudinal regressions appropriate for each type of outcome variable. HEALTH ECONOMICS: Costs & EQ-5D index score will be collected to conduct a cost-utility analysis up to 2 years to identify the most efficient provision of future care within the NHS. We will estimate incremental costs & QALYs & Net Monetary Benefit by means of regression methods. Probabilistic sensitivity analysis will be conducted to test the robustness of the results. We will extrapolate long-term cost-effectiveness beyond the trial if deemed appropriate. PROJECT TEAM: We have assembled an expert multidisciplinary team of clinicians, patient representatives, methodologists, health economists, & statisticians with vast experience in running such studies