What predicts attainment of substantial clinical benefit in neck pain-related disability 24 months after surgery for cervical spondylotic myelopathy? A QOD CSM study.

OBJECTIVE

Although surgery for cervical spondylotic myelopathy (CSM) has traditionally focused on halting neurological deterioration from myelopathy, many patients also report long-term improvements in neck pain and neck pain-related disability-outcomes not typically considered primary goals of intervention. This study aimed to identify preoperative predictors for attainment of substantial clinical benefit (SCB) in neck pain-related disability after CSM surgery.

METHODS

The authors retrospectively analyzed prospectively collected data from the 14 highest accruing Spine CORe™ sites in the Quality Outcomes Database CSM dataset. Logistic regression, XGBoost, and random forest models were trained to predict SCB attainment, defined as ≥ 20-point improvement in Neck Disability Index (NDI) score at 24 months postoperatively. Input features included demographic, clinical, and surgical variables in addition to baseline patient-reported outcomes. Model performance was evaluated via area under the receiver operating characteristic curve (AUROC) analysis.

RESULTS

Of 1085 prospectively enrolled patients, 904 met the inclusion criteria for this analysis. The cohort had a mean age of 60.9 years, 48.3% were female, and 76.1% were White. The mean NDI score improved from 38.0 at baseline to 21.1 at 24 months, with 52.7% of patients achieving minimal clinically important difference and 44.7% achieving SCB. Significant positive predictors of SCB included higher baseline neck pain intensity (OR 1.99) and greater interference with recreation (OR 1.77), work (OR 1.57), lifting (OR 1.44), and driving (OR 1.36), as well as anterior surgical approach (OR 1.67) and higher socioeconomic status (SES) index (OR 1.23). Significant negative predictors included fusion stopping at C7 (OR 0.77) and Medicare, Medicaid, or government/Veterans Affairs (VA) insurance (OR 0.79). Logistic regression, XGBoost, and random forest models showed good discriminative performance (AUROC 0.78-0.82).

CONCLUSIONS

In a large prospective cohort of operative CSM patients, higher baseline neck pain intensity; greater interference with recreation, work, lifting, and driving; anterior surgical approach; and higher SES were associated with greater likelihood of achieving SCB for neck pain-related disability. Fusion stopping at C7 and Medicare, Medicaid, or government/VA insurance were negative predictors. This is the first study to identify factors predicting SCB in NDI score after CSM surgery, challenging traditional views on the effectiveness of surgery for relief from neck pain-related disability in this population and demonstrating that a subset of patients can achieve substantial, sustained relief. These results highlight the potential for data-driven approaches to enhance patient selection and optimize surgical decision-making in the treatment of CSM.