EARLY VALUE ASSESSMENT PROGRAMME

The NHS Long Term plan (2019) identified musculoskeletal conditions as one of the key long-term conditions responsible for a substantial amount of poor health in the population. Since 2019, musculoskeletal problems have been among the top 3 reasons for sickness absence in the UK, and in 2022 this equated to 19.5 million workdays (Office for National Statistics report on sickness absence in the UK labour market: 2022). Despite the high volume of orthopaedic procedures being done in the UK, satisfaction with procedures is low. NHS data on patient-reported outcome measures for hip and knee replacement procedures (2021-2022) reported that 64.5% and 77.6% of people rated satisfaction with knee and hip replacements, respectively, as 'excellent' or 'very good'. Commonly cited causes of dissatisfaction include persistent pain, stiffness and unmet expectations of the operation (DeFrance and Scuderi 2022). While not always warranting revision surgery, these post-surgery outcomes may negatively affect health-related quality of life. This can lead to more follow-up appointments and increased prescription of pain medication and cause continued disruption to activities of daily living. The main conceptual benefit of robot-assisted surgery is improving the precision of the implant positioning and consistency of the surgeon's work. Robot-assisted surgery allows the surgeon to position the implant relative to the person's native anatomy and joint alignment, with a higher degree of precision and confidence. Conceptually, this will result in better patient outcomes by improving recovery from surgery, reducing pain and stiffness, and allowing a quicker return to activities of daily living. Robot-assisted surgery may enable more knee replacement procedures to be done as partial replacements rather than total replacements. This could improve recovery because partial replacements have fewer complications, shorter recovery times and shorter lengths of stay, as outlined in the NICE guideline on joint replacement (primary): hip, knee and shoulder.

The committee noted that there is a wider NHS England robot-assisted surgery steering group. The steering group coordinates national strategies for training, procurement and implementation of robot-assisted surgery services, and produces guidance on surveillance of robot-assisted surgery programmes. The committee also noted the British Orthopaedic Association's guidance on robotics in orthopaedics, which provides a tool kit for hospitals when setting up a new musculoskeletal robotic surgical service.

The committee was aware that some technologies included in this assessment are already being used in the NHS. Clinical experts highlighted that the robotic technology adopted by an NHS trust is likely to be influenced by which implants are already in use for conventional surgery. Trusts are likely to favour technologies made by companies with whom there is an established relationship. The committee was also aware of the widespread use in other countries of robotic technologies, including for indications beyond those in the scope of this assessment. The committee agreed that in the future, developments in technology may influence the selection of robotic technologies by expanding the available indications.

Training for the whole surgical team is essential for each robotic technology used in each centre. The NHS England steering group has formed a subcommittee that is in the process of producing guidance for training as part of its guidance on robot-assisted surgery. Experts said that training is technology-specific but that people who are trained in 1 robotic technology may be quicker to learn how to use other robotic technologies. Companies said that training costs are usually included in the cost of the robotic technology, including when the technology is updated. But, experts noted that training time for NHS staff should be included in the economic model, acknowledging that training will need to be repeated for new staff and to maintain competencies.

Most people have joint replacement procedures to reduce the pain and stiffness in their joints, which is usually associated with osteoarthritis. Successful joint replacement surgery helps restore function in the affected joints and allows people to resume daily activities. Clinical experts highlighted that patient dissatisfaction following conventional surgery is quite common, with many people needing ongoing pain medication and other support to help them manage. The experts advised that robot-assisted surgery has potential to improve patient satisfaction. A patient expert spoke of their positive experience of robot-assisted surgery for a knee replacement. The committee agreed that this commentary was helpful to understand the perspectives and experiences of people having robot-assisted surgery. It acknowledged that while this is difficult to capture in research, efforts should be made to better understand patient quality of life after robot-assisted surgery for joint replacement.

The committee noted that the introduction of robot-assisted surgery may increase the safety of orthopaedic surgery for people at higher surgical risk. Some people may be denied conventional joint replacement surgery because of the associated risk of surgery. This includes older people and people with a high body-mass index or multiple comorbidities. Robot-assisted surgery allows enhanced preoperative planning, potentially reducing the risk of complications in people at higher risk. Experts advised that robot-assisted surgery may improve outcomes for people who need different surgical alignments, such as people with bow-leggedness, which is most common in people from Southeast Asian backgrounds. For some people, robot-assisted surgery may not be possible because of the difficulty of attaching sensors to the bone. These groups were identified by the EAG as people with mental or neuromuscular disorders that affect control of the knee joint, and people with insufficient bone quality or mass. It also noted that people with conditions that prevent full articulation of the hip joint may not be able to have robot-assisted surgery for a total hip arthroplasty (THA). The committee acknowledged that while robot-assisted surgery may not be suitable for everyone, conventional surgery will still be available. This means that everyone who needs joint replacement surgery can have appropriate care.

The committee was aware that robotic technologies are expensive and may not be viable in all centres. Experts told the committee that robotic technologies are most commonly obtained through volume-based contracts, whereby NHS trusts commit to a number of procedures each year. This approach to purchasing means that robotic technologies are more likely to be cost-effective in high-volume orthopaedic centres. The committee was also aware that the high cost of the technologies means that robot-assisted surgery is more widely available in the private sector. The committee noted that limiting access to robot-assisted surgery to these hospitals may exacerbate existing inequalities. The committee also noted that robot-assisted surgery may be more beneficial in complex surgical cases. These cases are typically done in lower volume centres, with more prehabilitation and rehabilitation, as well as more advanced planning because of the associated surgical risks. In the future, the NHS England robot-assisted surgery steering group may be influential in moderating access to robot-assisted surgery with a national strategy. The steering group is actively analysing and mapping current robot-assisted surgery provision in England. A key priority will be equitable provision of robot-assisted surgery based on need rather than current configuration.

The committee considered evidence for all 6 technologies from 26 publications and 2 national joint registries. The majority of the evidence was in total knee arthroplasty (TKA), with less evidence identified in partial knee arthroplasty (PKA) and THA. Evidence was prioritised by the EAG per robotic technology, to identify data for all primary outcomes per technology. The study designs of the included evidence ranged from randomised controlled trials to retrospective cohort studies. This represents the wide spectrum in the quality of, and the outcomes in, the evidence base for each robotic technology. The EAG's report summarised the limitations of the evidence. The key considerations were a lack of randomised evidence for THA and large variations in the quality and quantity of evidence across all 3 procedures. The committee was reminded of the uncertainties in the evidence that were considered when forming recommendations. The committee agreed that robot-assisted surgery broadly showed non-inferiority with conventional surgery in primary outcomes. These included length of hospital stay, complications, patient-reported outcome measures (PROMs), utilities and surgical revisions. The committee noted that alignment, which was a secondary outcome, was consistently better with robot-assisted surgery. But the evidence did not suggest that this resulted in better PROMs or clinical outcomes.

The key outcomes considered by the committee were PROMs, which is linked to the unmet need. Clinical experts reiterated that patient dissatisfaction was a key issue experienced in clinical practice. This is often because the procedure does not meet people's expectations because they experience continued pain and stiffness after the procedure, which results in the need for further support. PROMs were reported in a number of different ways, using different scales and follow-up times. This limited the ability of the EAG and committee to draw conclusions. Most PROMs showed no difference between robot-assisted surgery and conventional surgery. When statistically significant differences were seen, the benefit tended to be from robot-assisted surgery. But, the committee noted that many of these differences were below the minimally clinically important difference, which limited the certainty of their clinical significance. The committee acknowledged that there were uncertainties in the PROMs data. It suggested that further evidence generation should focus on reducing these uncertainties through larger studies that will inform future economic modelling.

Revisions were also considered to be a key outcome by the committee, because of the negative effect that revision surgery has on health-related quality of life. Revision data was limited in the published clinical evidence because of small sample sizes, short follow-up, and because they are relatively rare events. So the EAG considered data from the National Joint Registry (NJR) to be the most robust and relevant data to the NHS. The NJR did not contain enough robot-assisted surgery procedures to allow comparisons of revision rates to be made with conventional surgery. Additional work with the NJR that links with the NHS's Hospital Episode Statistics, would allow revision rates to be compared between robotic and conventional surgery with long-term follow up. A clinical expert highlighted that the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) contains separate revision data for robot-assisted surgery and conventional surgery. But this showed no statistically significant difference between surgical methods. Robot-assisted surgery is more established in Australia, so this registry contains more robot-assisted surgery data with longer follow up. The EAG noted that, compared with the NJR data, the population in the AOAOJRR data differed in mean age and American Society of Anesthesiologists risk score. The committee considered evidence from both national registries, but accepted that differences between the UK and Australian healthcare systems limit the generalisability of the data from the AOANJRR. The committee concluded that the AOANJRR data was useful in demonstrating the growth of robot-assisted surgery in other countries but was not generalisable to the NHS. So further revision data should be sought from the NJR to inform future economic modelling.

The learning curve was a primary outcome and was discussed by the committee, which deemed it to be a minor clinical concern. The EAG identified multiple single-arm studies for this outcome. They showed that between 6 and 30 cases are needed to achieve proficiency (although the definition of proficiency varied between studies). This was supported by the clinical experts and the companies. The clinical experts explained that learning is usually supported by the company. The committee agreed that the learning curve may have economic implications, but it is not a concern when considering the clinical effectiveness of robot-assisted surgery.

The EAG suggested that improved ergonomics and surgeon quality of life with robot-assisted surgery could be important considerations for surgeon acceptability. Experts did not agree that this was a key benefit of robot-assisted surgery and no consensus was reached on whether the physical and cognitive burden is increased or reduced. The committee agreed that this benefit was plausible, but that the evidence to support it is limited.

The committee noted that the Mako platform had the most mature evidence for all 3 procedures, with randomised controlled trials for knee procedures and prospective evidence for THA. The other 4 technologies that have a conditional recommendation for use during the evidence generation period had less evidence and it was generally lower quality. But the committee was convinced that the evidence was sufficient to support their use. The committee heard that all 5 technologies that have conditional recommendations for use during the evidence generation period are in use or planned to be in use in the UK and will submit data to the NJR. Experts explained that robotic technologies for orthopaedics typically target knee procedures first, before expanding their indications. For this reason, only 3 technologies in this evaluation are indicated for THAs. Experts advised that if an adopted technology's indication was expanded, it is likely that the centre would use it for the new indication. This was an important consideration in the conditional use during the evidence generation period recommendation for robot-assisted surgery for THA, which had a less mature evidence base. Experts advised that although robotic technologies work in different ways, they all aim to improve the precision of implant positioning. They added that this is still an emerging field within orthopaedic surgery. Experts noted that each robotic technology in the scope has its own specific implants. They explained that all implants must undergo benchmarking through an Orthopaedic Data Evaluation Panel rating to demonstrate safety before they are used in the NHS. This also indicates the safety of the robotic technology. The committee noted that the NJR can tag information related to specific implants and robotic technologies. It agreed that this would allow variations in outcomes related to specific platforms to be identified. The committee concluded that the benefits seen in the evidence for Mako could be similar in the other 4 technologies that have less mature evidence. So it decided to make a conditional recommendation for use during the evidence generation period for the 5 technologies.

The committee were told that the company that produces SkyWalker did not respond to the request for information from NICE, and so the EAG relied on publicly available information. So limited clinical evidence from a single non-UK retrospective cohort study containing 30 participants per arm was identified for the SkyWalker platform. This evidence showed mixed effects of the technology on complications and operating time, with no difference seen in PROMs. Also, no technology costs were available for the SkyWalker platform, so the EAG was unable to produce an economic model. The committee concluded that a recommendation for research only was appropriate for the SkyWalker platform.

The committee agreed that there were evidence gaps in all technologies assessed in this early value assessment. It noted in particular that, for THA, Mako only had limited non-randomised evidence and CORI had no evidence within scope. The committee discussed which outcomes were most important to inform future decision making. Impact on patient quality of life and resource use were prioritised as key areas for further evidence generation. More accurate technology-specific utilities data and more data on resource use could be used to improve the certainty of the economic modelling. The clinical impact of robot-assisted surgery in different subgroups was also identified as an area for further evidence generation, but with lower priority. It was given a lower priority because it does not directly affect the economic modelling or the overall efficacy of robot-assisted surgery across the NHS. But, the committee did agree that robot-assisted surgery may be more beneficial in some groups. For example, people who are at higher risk or people from a Southeast Asian background, in whom bow-leggedness is more common and can result in alignment challenges with conventional surgery. It concluded that further evidence should be generated to inform where robot-assisted surgery should be adopted to provide the greatest benefit to people having orthopaedic procedures. The committee agreed that it is important that future evidence collects information on variables that may confound findings on the effectiveness of robot-assisted surgery. The committee concluded that gathering information on these variables is important for future decision making, especially when assessing data from national joint registries.