The purpose of this study is to estimate the cost-effectiveness of performing total hip arthroplasty (THA) versus nonoperative management (NM) in non-obese (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. We constructed a state-transition Markov model to compare the cost-utility of THA and NM in the six above-mentioned BMI groups over a 15-year time period. Model parameters for transition probability (i.e. risk of revision, re-revision, death), utility, and costs (inflation adjusted to 2017 US dollars) were estimated from the literature. Direct medical costs of managing hip arthritis were accounted in the model. Indirect societal costs were not included. A 3% annual discount rate was used for costs and utilities. The primary outcome was the incremental cost-effectiveness ratio (ICER) of THA versus NM. One-way and Monte Carlo probabilistic sensitivity analysis of the model parameters were performed to determine the robustness of the model.Introduction
Methods
The purpose of this study is to compare 90-day costs and outcomes for primary total hip arthroplasty (THA) patients between a non-obese (BMI 18.5–24.9) versus overweight (25–29.9), obese (30–34.9), severely-obese (35–39.9), morbidly-obese (40–44.9), and super-obese (45+) cohorts. We conducted a retrospective review of a prospective institutional database of primary THA patients from 2006–2013, including patients with a minimum of three-year follow-up. Thirty-three super-obese patients were identified, and the other five cohorts were randomly selected in a 2:1 ratio (total n = 363). Demographics, 90-day outcomes (costs, reoperations, and readmissions), and outcomes after three years (revisions and change scores for SF12, HHS, and WOMAC) were collected. Costs were determined using unit costs from our institutional administrative data for all in-hospital resource utilization. Comparisons between the non-obese and other groups were made with Kruskal-Wallis tests for non-normal data and chi-square and Fisher's exact test for categorical data. The 90-day costs in the morbidly-obese ($13,134 ± 7,250 mean ± standard deviation, p <0.01) and super-obese ($15,604 ± 6,783, p <0.01) cohorts were statistically significantly greater than the non-obese cohorts ($10,315 ± 1,848). Only the super-obese cohort had statistically greater 90-day reoperation and readmission rates than the non-obese cohort (18.2% vs 0%, p <0.01 and 21.2% vs 4.5%, p=0.02, respectively). In addition, reoperations and septic revisions after 3 years were greater in the super-obese cohort compared to the non-obese cohort 21.2% vs 3.0% (p = 0.01), and 18.2% vs 1.5% (p= 0.01), respectively. There were no other statistical differences between the other cohorts with the non-obese cohort at 90-days or after 3 years. Improvements in SF12, HHS, and WOMAC were comparable in all cohorts. The 90-day costs of a primary total hip arthroplasty for morbidly-obese (BMI 40–44.9) and super-obese (BMI>45) are significantly greater than for non-obese patients, yet these patients have comparable improvements in outcome scores. Health care policies, when based purely on the economic impact of health care delivery, may place morbidly-obese and super-obese patients at risk of losing arthroplasty care, thereby denying them access to the comparable quality of life improvements.
