Robotic Arthroplasty: a Clinical and cost Effectiveness Randomised controlled trial. (RACER)

Robotic systems are increasingly used to help surgeons perform knee replacements. They are expensive. We do not know if they provide any patient benefit, or if they are cost effective. In 2017, 109,000 total knee replacements (TKRs) were done in the UK, at a cost of over £550M.(1-4) They substantially improve pain and function in many people. However up to 20% of patients are dissatisfied, 20% report pain as bad or worse than before the operation, and many have ongoing restrictions to activities of daily living.(5-7) Robotic systems aim to improve outcomes by helping the surgeon perform knee replacements with greater precision and less tissue damage than can be achieved with conventional instruments.(8) If robotic assisted surgery is beneficial, the costs could be offset by shorter inpatient stays and fewer complications. Their effectiveness compared with standard care is not known. Despite their high cost and lack of evidence of effectiveness, they are progressively being introduced into the NHS. Our overarching aim is to determine whether robotic-assisted TKR is clinically and cost-effective when compared to TKR using conventional instruments. We will perform a participant and assessor blind, pragmatic multi-centre RCT in 332 people with knee osteoarthritis. We will compare TKR done with the MAKO robot (Stryker, USA; the the most commonly used robotic-arm system worldwide, and until recently the only system available in the UK) to exactly the same type of TKR done with conventional instruments. In addition to the normal TKR incision, two or four small additional incisions in the control group will blind participants and staff to the allocation. We will blind the operation note to prevent bias from staff such as physiotherapists. Everyone will have the same surgical planning before the operation using a CT scan. Everyone will receive the same post-operative rehabilitation programme. Our primary outcome is the Forgotten Joint Score (FJS) at 12 months. This was developed specifically for assessing the results of joint replacement, there are good data on its metrics, and our PPI group approved of its use. We will also collect pain intensity over the first three post-operative days, estimated blood loss, in-patient opioid use, and time to discharge. PROMs at three, six and 12 months and two, five and ten years will include: FJS, Oxford Knee Score, Oxford Activity & Participation Questionnaire, PROMIS pain intensity scale, satisfaction, re-operations, adverse events, and resource use. We will also perform a very low dose CT and long-leg x-ray to check TKR alignment, three months post-operatively. Patient involvement has been, and will be, integral throughout the study; including on the trial management team. An internal pilot study in the first five months will confirm the feasibility of the study. The main analysis will be on an intention-to-treat basis, using a generalised linear model including terms to adjust for allocation group, age, gender, BMI (>35) and primary compartment involved. A prospectively planned economic evaluation will be conducted from a NHS and personal social services perspective.(9) If we find that surgery done with a robot is clinically and cost-effective, it will improve the care of people undergoing knee replacement, the majority of whom do not currently have access to this technology. However, if it is ineffective the study will stop the widespread adoption of an expensive and unnecessary new technology.