Use of pre-operative SRS-22r profiles to predict clinically meaningful improvement after AIS surgery: a detailed multivariable MCID analysis in 189 patients.

Abstract

Retrospective study. The aim of this study is to evaluate the predictive ability of pre-operative Scoliosis Research Society-22 revised questionnaire (SRS-22r) domains for achieving Minimum clinically important difference (MCID) at 2 years after adolescent idiopathic scoliosis (AIS) surgery. Understanding which AIS will achieve clinically meaningful postoperative benefit remains challenging, particularly when counseling families about expected changes in health-related quality of life. MCID thresholds for SRS-22r domains have been described, but it is unclear how baseline profiles can be leveraged to anticipate meaningful improvement at 2 years after surgery. A retrospective analysis of AIS patients undergoing posterior spinal fusion was performed. Pre-operative, 1-year, and 2-year SRS-22r domain scores (Function, Pain, Self-Image, Mental Health), VAS pain, and ODI were collected. MCID was defined as a ≥ 1-point improvement from baseline to 2 years. Univariate and multivariable logistic regression models were constructed, and model performance was assessed using ROC-AUC and calibration curves. ROC-based threshold optimization using Youden's J statistic, and fivefold cross-validation was used to derive clinically interpretable pre-operative cut-points. In total, 189 patients were analyzed for the study. All SRS-22r domains improved significantly at 2 years, with the greatest absolute gain in Self-Image 64.9% achieved MCID, followed by Pain 46.5%, Function 37.7%, and Mental Health 36.0%. Lower pre-operative domain scores strongly predicted MCID (OR range 0.03-0.13) and pronounced ceiling effects were observed among patients with high baseline scores. Multivariable models demonstrated excellent to outstanding discrimination via AUC: Function (0.95), Mental Health (0.91), Pain (0.87), and Self-Image (0.83), with good calibration across risk strata. ROC-based optimal pre-operative thresholds were Function ≤ 3.2, Pain ≤ 3.4, Self-Image ≤ 3.2, and Mental Health ≤ 3.8, and showed stable performance on cross-validation. Pre-operative SRS-22r domain scores are robust predictors of clinically meaningful improvement following AIS surgery. Self-Image demonstrates the greatest postoperative responsiveness, and lower baseline PROM severity consistently confers higher likelihood of achieving MCID, whereas patients starting with near-ceiling scores show limited numerical gain. The combination of domain-specific logistic regression models, ROC-derived cut-points, and well-calibrated probability estimates supports the development of practical PROM-based prediction tools to enhance pre-operative counseling, expectation management, and shared decision-making in AIS care. III.