Long-Term Cost-Effectiveness of Adult Spinal Deformity Surgery

Abstract

Introduction There is a lack of long-term outcomes in the existing cost-effectiveness literature. This study examines the long-term cost-effectiveness of surgical treatment of ASD with a Markov model populated by estimates from literature and observed data on hospital costs and outcomes. Using estimates from literature and observed data on costs and health-related quality-of-life (HRQOL) outcomes, we simulate costs and outcomes for surgical and nonsurgical treatment to estimate incremental cost-effectiveness ratios (ICERs) after 10- and 20-year follow-up. Material and Methods We develop a robust Markov model with parameter uncertainty, incorporating costs, and outcome distributions from published literature and observed patient data. Summary statistics on mortality rates, revision rates, and nonsurgical costs are taken from the literature, whereas statistics for surgical costs and QALYs are taken from a single-center, retrospective administrative data set and a multicenter prospective data set. Costs of surgery are measured as total hospital costs from administrative records. QALYs were calculated from the SF-6D. We projected costs and QALYs for the surgical and nonsurgical populations allowing varied readmission, cost, and QALY outcomes over 10- and 20-year follow-up. Costs for surgical and nonsurgical patients are intended to reflect the full spectrum of services incurred for surgical or nonsurgical care (e.g., physical therapy, pain management, etc.). Costs and QALYs were in 2010 dollars and discounted at 3.5%. Our results are based on 1,000 simulated patients, each proceeding through 6-month cycles over a 10- and 20-year follow-up periods. Confidence intervals are formed by repeating the simulations for 1,000 iterations. Results Discounted surgical costs averaged $170,523 over 10-year follow-up, with average discounted QALYs of 6.3. Meanwhile, nonsurgical costs averaged $40,046 over 10-year follow-up with average QALYs of 5.4. This resulted in an average ICER from surgical treatment of $144,000 (95% CI: $121,500–$177,500) per QALY gained. Over 20-year follow-up, surgical costs averaged $208,616 and QALYs averaged 9.8, whereas nonsurgical costs averaged $63,467 and QALYs averaged 8.6. Over 20-year follow-up, the ICER decreased (became more cost-effective) to $115,000 (95% CI: $88,850–$156,200) per QALY gained. Conclusion Our results illustrate the potential for surgical treatment for ASD to be cost-effective over extended follow-up compared with nonsurgical treatment. These findings argue in favor of longer follow-up in ASD studies for accurate cost-effectiveness comparison. Future research should also pursue measurement of indirect costs/benefits resulting from changes in absenteeism or productivity.