CORR Insights®: Can Technology Assistance be Cost Effective in TKA? A Simulation-based Analysis of a Risk-prioritized, Practice-specific Framework

Abstract

Where Are We Now? We are in an exciting phase in the technological evolution of total knee arthroplasty (TKA). A number of technology-assisted TKA (TA-TKA) approaches using computer navigation, patient-specific instrumentation, or robotics have been proposed to improve the implant position and alignment and, ostensibly, the clinical results of TKA. However, although some studies have found that using TA-TKA improves the precision by which the planned coronal alignment is achieved [4, 5, 8], this has not resulted in a decrease in the revision rate or an improvement in function in the long term compared with conventional TKA. Because it is expensive and requires more intraoperative time than conventional TKA, there are concerns of time and cost efficiency with the use of TA-TKA. There are also other risks, such as the uncertainty and potential harms associated with novelty itself [3, 7]. Proponents of TA-TKA have suggested that with increased use and surgical experience, these technologies would reduce the risk of revision and mitigate the concerns of cost and time efficiency [11]. However, considering that TKA has low revision rates with or without technology assistance, a clinical trial evaluating the efficacy of TA-TKA compared with that of conventional TKA to reduce revision rates would need a sample size of several thousands of patients. Such a trial would be a mammoth task requiring substantial financial and human resources dedicated to ascertaining the efficacy of a modality that so far has shown to be ineffective for improving clinical outcomes. What if there was a way to simulate a large-sample study using modern advancements based on the available evidence, using baseline variables and published clinical outcomes of patients who had TA-TKA and those with conventional TKA? An earlier computer simulation study by Hickey et al. [2] did exactly that; they wanted to predict whether TA-TKA, owing to its precision in implant placement, would reduce the revision risk and be cost effective. Although such simulation studies have been done before with differing results regarding the cost-effectiveness of TA-TKA [10, 11], Hickey et al.’s study [2] was convincing. It found that studies enrolling all-comers would have to be impracticably large to find any benefit to the added technology. This fits well with what we have seen in the trials I mentioned; perhaps there would be improvement in alignment, but no differences in revision favoring the expensive new technology in unselected populations. But what would happen if we focused our attention on patients who are at a higher risk of undergoing premature revision? I’m referring to factors such as younger age, higher BMI, and perhaps men. This is what Hickey et al. [3] explored in an exciting follow-up study, published in this month’s Clinical Orthopaedics and Related Research®. Using computer simulation, the authors evaluated data of 20,000 TKAs by randomly assigning high-risk and low-risk variables. Differences in patient-specific reductions in the revision risk and change in quality-adjusted life were predicted for 5000 simulated operations per patient using the reported precisions of the three TA-TKA techniques (navigation, patient-specific instrumentation, and robotic), conventional TKA technique, and an ideal TKA with neutral coronal alignment. The authors did not find enough potential benefit in the low-risk groups to make a further inquiry worth considering, but their simulations suggested that in an elevated-risk population, a practice size of 100 patients per year who would meet their criteria as being at a higher risk of revision could experience cost-effectiveness with navigation and robotic TKA, with use percentages of 6% and 20%, respectively. Thus, it certainly seems like a potentially helpful intervention. However, as a simulation, one would not want to draw strong clinical inferences or use this to implement expensive new technology in practice. Rather, these findings