Preoperative patient-reported outcome score thresholds predict the likelihood of reaching MCID with surgical correction of adult spinal deformity.

Abstract

Preoperative (pre-op) identification of patients likely to achieve a clinically meaningful improvement following surgery for adult spinal deformity (ASD) is critical, especially given the substantial cost and comorbidity associated with surgery. Even though pain is a known indication for surgical ASD correction, we are not aware of established thresholds for baseline pain and function to guide which patients have a higher likelihood of improvement with corrective surgery. We aimed to establish pre-op patient-reported outcome measure (PROM) thresholds to identify patients likely to improve by at least one minimum clinically important difference (MCID) with surgery for ASD. This is a retrospective cohort study using prospectively collected data. We reviewed 172 adult patients' charts who underwent corrective surgery for spinal deformity. Included measures were the Visual Analog Scale for pain (VAS), Oswestry Disability Index (ODI), and Scoliosis Research Society-22 (SRS-22). Our primary outcome of interest was improvement by at least one MCID on the ODI and SRS-22 at 2years after surgery. As part of usual care, the VAS, ODI, and SRS-22 were collected pre-op and re-administered at 1, 2, and 5years after surgery. MCIDs were calculated using a distribution-based method. Determining significant predictors of MCID at two years was accomplished by Firth bias corrected logistic regression models. Significance of predictors was determined by Profile Likelihood Chi-square. We performed a Youden analysis to determine thresholds for the strongest pre-op predictors. At year two, 118 patients (83%) reached MCID for the SRS and 127 (75%) for the ODI. Lower pre-op SRS overall, lower pre-op SRS pain, and higher pre-op SRS function predicted a higher likelihood of reaching MCID on the overall SRS (p < 0.05). Higher pre-op ODI, lower SRS pain and self-image, and higher SRS overall predicted a higher likelihood of reaching MCID on the ODI (p < 0.05). An ODI threshold of 29 predicted reaching MCID with a sensitivity of 0.89 and a specificity of 0.64 (AUC = 0.7813). An SRS threshold of 3.89 predicted reaching MCID with a sensitivity of 0.93 and specificity of 0.68 (AUC = 0.8024). We identified useful thresholds for ODI and SRS-22 with acceptable predictive ability for improvement with surgery for ASD. Pre-op ODI, SRS, and multiple SRS subscores are predictive of meaningful improvement on the ODI and/or SRS at 2years following corrective surgery for spinal deformity. These results highlight the usefulness of PROMs in pre-op shared decision-making.