Abstract
Introduction
We investigate the incremental cost-effectiveness of ASD surgery relative to nonsurgical treatment by classification of sagittal modifier using the SRS–Schwab classification.
Methods
This is a retrospective study of consecutive surgical and nonsurgical patients with ASD. Surgical data were from a single center observational data set, whereas nonsurgical data were from a prospective multicenter database. Hospital costs were collected from administrative data and QALYs were calculated from the SF-6D, each discounted at 3.5% annually. We analyzed the incremental cost-effectiveness ratios (ICERs) based on classification of sagittal modifier at baseline. The three modifier types considered were mild (SVA < 4 cm, PT < 20 degrees, PI–LL within 10 degrees), moderate (at least 1 of SVA 4–9.5 cm, PT 20–30 degrees; PI–LL 1020 degrees, and no severe modifiers); and severe (at least 1 of SVA > 9.5 cm, PT > 30 degrees, PI–LL > 20 degrees). The difference in QALYs between the surgical and nonsurgical cohorts at 2 years was projected through 5-year follow-up. 95% confidence intervals were calculated using nonparametric bootstrap methods.
Results
Two-year follow-up was available for 278 patients (168 surgical and 110 nonsurgical), predominantly female (n = 245, 88%) with average age of 52 years (range, 18–81 years). Total surgical costs averaged $121,857, including readmissions. The 2-year ICER was least cost-effective for patients with severe sagittal modifiers at $424,978 and most cost-effective for patients with mild modifiers at $206,730. Projected through 5 years, the ICER among patients with severe sagittal modifiers was $151,689, reducing to $73,790 for patients with mild modifiers.
Conclusion
On the basis of the World Health Organization's suggested upper threshold for cost-effectiveness (three times GDP, or $140,000), ASD surgery appears cost-effective after a 5-year period for mild and moderate sagittal modifier types. Although average 5-year ICERs for severe modifier types are above this threshold, the threshold is included within the confidence interval, indicating the 5-year ICER is not statistically significantly higher than the threshold value. The ICER is expected to further decrease over extended follow-up.