Page 45 - ISAKOS 2021 Newsletter Volume 1

CURRENT CONCEPTS
Cognitive biases that act as barriers include confirmation bias (interpretation of data that confirms pre-existing beliefs), anchoring (overweighting first piece of information that we receive), neglect of probability, availability bias (overweighting evidence that we have personally observed), and conservatism or decision inertia.
TKR options in which the potential or theoretical design advantages of an option are outweighed by its actual disadvantages represent a “developmental dead-end.” Conversely, the continued use of an option despite evidence of no benefit at a higher cost, or actual higher risk, represents an “evidence-utilization incongruity.” Unfortunately, the lack of a universally accepted definition of practice-changing evidence, as detailed above, makes these concepts more theoretical than practical.
Is Revision Risk a Viable Outcome Measure?
When considering the variety of TKR prosthesis options that are available to surgeons, it must be recognized that these options often interact in terms of revision risk, and hence, cannot be examined in isolation. For example, AOANJRR data indicates that cementless fixation and patellar non- resurfacing are associated with a much greater risk of revision when used in Posterior Stabilized TKR than when used in Cruciate Retaining / Minimally Stabilized TKR. As revision TKR is an expensive procedure, with greater risks and worse functional outcomes than primary TKR, measures to reduce the rate of revision with optimum prosthesis selection are vital, particularly when no identifiable clinical advantage exists between alternative design options.
As discussed, registry data represent comprehensive and pragmatic information on the mean prosthesis revision rate, but direct prosthesis comparisons are at a risk of confounding bias, despite revision risk being adjusted for age and gender. Recognition of this confounding bias may represent a possible explanation of why, in a recent survey of members of the knee societies of Australia, New Zealand, Britain, and South Africa, 20% of surgeons responded that they disregard registry data when choosing their preferences. It is difficult to reconcile some surgeons’ stated disregard for revision risk as an attribute as it is not independent of other attributes. As an illustration, patients without infection who have adequate fixation, acceptable pain, functional range of motion, and stable implants will have lower revision rates than those with poor flexion, stiffness, infection, and loosening.
Doctor, I’m Not Satisfied
Patient dissatisfaction due to both prosthesis and non- prosthesis-related factors remains a challenging issue in a subgroup of patients, estimated to represent between 1% and 17% of those who undergo TKR. Prosthesis-related factors that have been linked to patient dissatisfaction include ongoing pain, poor flexion, poor function, instability, infection, and component loosening.
While patient satisfaction rates are increasingly being investigated by both clinical investigators and health-care funders, the relationship between revision rates and patient dissatisfaction rates remains complex3, correlating with unmet expectations, ongoing pain, and complications. To perform a revision, the surgeon obtains consent from the patient to replace all or some of the components of the prosthesis to correct a valid identified underlying problem. Assuming that a surgeon-assessed correctable problem exists and the patient is able and willing to undergo a revision, this consent to proceed with revision represents a reasonable proxy for a patient’s dissatisfaction exceeding the threshold at which it warrants another procedure with its associated risks, costs, and discomfort (Fig. 1). It should be noted from Figure 1 that patients who undergo revision represent a subset of those who are dissatisfied, as those who do not have a surgeon-assessed correctable problem and those who have a surgeon-assessed correctable problem but are unable or unwilling to have a revision are not represented. It is likely that these dissatisfied patients who do not undergo revision will be evenly distributed across all groups examined. Moreover, the lack of a surgeon-assessed correctable problem does not mean that no problem exists with the implant, just that one has not been identified. Finally, some patients who require a revision may not be dissatisfied, such as those requiring revision for late infection, fracture, or late loosening of a previously well-functioning implant.
01 Flowchart of Interaction between Dissatisfaction and Revision
With registry data, the reasons for revision are provided by the revising surgeon, who may or may not be the primary surgeon. In order of decreasing frequency, the listed reasons for revision in the AOANJRR are loosening-lysis, infection, patellofemoral pain, pain and instability, patellar erosion, arthrofibrosis, fracture, malalignment, tibial insert wear, metal- related pathology, incorrect sizing, and “other”. Given that a large majority of these reasons are altered by the primary surgeon’s implant choices, surgeons disregard the registry revision risk of their preferred prosthesis at their patient’s peril.
ISAKOS NEWSLETTER 2021: VOLUME I 43