Cost-Effectiveness of Total Knee Arthroplasty vs Nonoperative Management in Normal, Overweight, Obese, Severely Obese, Morbidly Obese, and Super-Obese Patients: A Markov Model

We estimated the cost-effectiveness of performing total knee arthroplasty (TKA) vs nonoperative management (NM) among 6 body mass index (BMI) cohorts.

A Markov model was used to compare the cost-utility of TKA and NM in 6 BMI groups (nonobese [BMI 18.5-24.9], overweight [25-29.9], obese [30-34.9], severely obese [35-39.9], morbidly obese [40-49.9], and super-obese [50+] patients) over a 15-year period. Model parameters for transition probability (ie, revision, re-revision, death), utility, and costs were estimated from the literature. Direct medical costs but not indirect societal costs were included in the model. Costs and utilities were discounted 3% annually. The primary outcome was the incremental cost-effectiveness ratio (ICER) of TKA vs NM. One-way and probabilistic sensitivity analyses of the model parameters were performed to determine the robustness of the model.

Over the 15-year period, the ICERs for the TKA vs NM for the different BMI categories were nonobese ($3317/quality-adjusted life years [QALYs]), overweight ($2837/QALY), obese ($2947/QALY), severely obese ($3536/QALY), morbidly obese ($5531/QALY), and super-obese ($11,878/QALY). The higher BMI groups tended to have higher incremental QALYs and also higher incremental costs. The probabilistic sensitivity analysis with an ICER threshold of $30,000/QALY showed that TKA would be cost-effective in 100% of simulations of patients with a BMI<50 and 99.16% of super-obese simulations.

While performing TKA on super-obese patients is more expensive, the substantial improvements in patient outcomes make it cost-effective. Therefore, withholding TKA care based on a BMI would lead to an unjustified loss of health-care access.