Oral Posters: MIS
Presented by: N. Anand - View Audio/Video Presentation (Members Only)
N. Anand(1), J. Cohen(1), R. Cohen(1), B. Khandehroo(1), S. Kahwaty(1), E. Baron(1)
(1) Cedars Sinai Medical Center, Spine Center, Los Angeles, CA, United States
Introduction: A key limitation of MCID is the large variability in reported values using respective anchor and distribution-based values. We conducted this study to determine the optimal calculation strategy for MCID based on Oswestry Disability Index (ODI) in patients undergoing CMIS correction of adult spinal deformity (ASD).
Methods: This is a single center study from a prospectively collected database of all patients who underwent CMIS correction for ASD (Cobb angle > 20 degrees or SVA > 50 mm or PI/LL mismatch > 10). Out of 124 patients with 3 or more levels operated Only 81 patients with a documented 2-year ODI were included for this study. Pre-operative and latest ODI scores were collected. MCID was determined using established anchor-based and distribution-based methods as well as a combined anchor-based minimal detectable change (MDC) approach. A linear regression model was constructed comparing baseline ODI to delta-ODI.
Results: The study consisted of 81 patients (56 females) with mean age 65.1 (range 18-84, SD 12.6), BMI 28.1 (range 15.6-50.9, SD 6.99), follow-up 64 months (range 24 -105, SD 26) and levels operated 5.6 (range 3-16, SD 2.7). MCID values ranged from 5.0 to 19.3 and 3.45 to 15.1, consistent with a 386% and 438% variability using respective anchor and distribution-based methods. The combined anchor-based MDC method yielded an MCID value of 15.1. Regression analysis of baseline ODI to delta-ODI showed a statistically significant positive correlation with R square value of .128.
Conclusion: We report a large variability in MCID values in patients undergoing CMIS correction, 386% and 438% using respective anchor and distribution-based methods. Our study argues for a combined anchor-based MDC approach, which produced an MCID of 15.1 for CMIS patients. The method remedies limitations of anchor and distribution-based methods, follows a standardized calculation methodology, does not require a steep learning curve and conveniently produces a single-score threshold.