All-polyethylene tibial (APT) implants have been available for use in total knee arthroplasty (TKA) for decades. Except for one particular implant design, APT implants have shown equivalent functional outcome and survivorship to metal-backed tibial (MBT) components.1 Two recent systematic reviews have demonstrated no difference in durability or functional outcome between APT and MBT components.1,2 Despite this data, APT components continue to be used uncommonly in the United States. Improved technical ease and the theoretical advantages of modularity are likely responsible for the continued popularity of MBT implants despite the fact that APT implants cost considerably less than their MBT counterparts.
The importance of cost control in TKA is increasing secondary to changing economic realities of healthcare and increasing prevalence of joint replacement. Payers are seeking ways to ensure quality care at more affordable reimbursement rates. Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. Implants account for a substantial portion of hospital costs for knee replacement and have been suggested as an essential part of cost control.3 As such, surgeons in the United States will probably need to factor in value when selecting implants and be required to justify the additional cost of “premium” implants.
Given recent systemic reviews concluding both equivalent effectiveness and survivorship, the APT component would appear to be inherently cost-effective when compared with an MBT design. However, the degree to which this implant is cost-effective has been difficult to quantify. The purpose of this study is to take a novel approach to examine the cost-effectiveness of APT components by determining what theoretical difference in revision rate would make modular MBT implants a more cost-effective intervention using our institutional cost data.
MATERIALS AND METHODS
A Markov decision model was used to evaluate the cost-effectiveness of APT components.4 A Markov decision model is a mathematical framework for modeling decision making in situations where outcomes are partly random and partly under the control of a decision maker. They are powerful tools for determining the best solution from all feasible solutions to a given problem. A decision model was constructed (Figure 1) to depict patients with arthritis of the knee being treated with either APT or MBT implants in a fashion similar to previously published models.5 At each point of a patient’s health status in the 20 years following surgery, they are either considered well after total knee replacement, well after revision surgery, or dead. Patients transition through the decision tree and pass through different states according to the probability of each event occurring, a process that is discussed further below. A utility value, measured in quality-adjusted life years (QALYs), and a cost are assigned to every health state and both primary and revision procedures within the model. The model is designed to determine the maximum failure rate for which the APT is the more cost-effective option.
The model probabilities used for survival and mortality following TKA were adapted from those published previously in the literature.5 A utility value was assigned to each health state. The utility after initial surgery was set to 0.83 and utility after revision was set to 0.6.5 These values were obtained from the Swedish Registry and Tufts Cost-Effectiveness Registry, respectively. We also included a disutility of -0.1 for the first year after surgery and -0.2 for the first year after revision, to account for the disutility of undergoing surgery and the post-surgery recovery. Disutilities represent the negative preference patients have for a particular health state or outcome, such as primary or revision knee arthroplasty.5 It is assumed that there is a higher morbidity associated with revision arthroplasty vs primary arthroplasty and, thus, has a higher disutility value assigned to it.
We assumed the age at the initial surgery to be 65 years. Age-specific mortality rates were taken from the 2007 United States Life Tables published by the Centers for Disease Control and Prevention.6 An additional probability of .007 of dying during the surgery or postoperative from the initial surgery and a probability of .011 from the revision was included.
Costs for the surgery were obtained from the University of Virginia’s billing department. We obtained the average cost for the diagnosis-related group in 2012. The cost of primary knee replacement was $17,578.06 with MBT implants. We subtracted institutional cost savings for the APT that could be achieved to obtain a cost of $16,272.10 for the APT. The cost of revision was $21,650.34 and assumed to be the same regardless of the type of initial surgery. A 3% discount rate was used.
The costs, QALYs, and probabilities were then used to compute cost-effectiveness ratios, or the cost per additional QALY, of the 2 options. Unlike previous models published in the orthopedic literature, we assumed a constant probability of revision for the MBT. We initially assumed a 1.0% probability of failure per year for the MBT implant. We then determined what revision rate for the APT would be necessary to be cost equivalent with the MBT. A sensitivity analysis was performed to examine the impact of varying assumptions regarding the rate of revision.
Continue to: Results...