Predicting Non-routine Discharge After Scoliosis Surgery: BMI Matters, but WHO Classification Doesn’t Add Value

Introduction: Obesity has been increasingly recognized as a risk factor in postsurgical outcomes, with higher body mass index (BMI) associated with complications and challenges in discharge planning. This study investigates the predictive power of WHO obesity classifications for non-routine discharge following multilevel scoliosis surgery. By analyzing the impact of BMI and comparing model performance with and without WHO classification, we aim to understand its influence on patient discharge disposition.

Methods: We utilized the Nationwide Readmissions Database (NRD) to identify patients who underwent multilevel scoliosis surgery. A total of 8,143 patients were included in the analysis, divided into WHO obesity classifications: non-obese (n=6,825), BMI 30-35 (n=600), BMI 35-40 (n=399), and BMI 40+ (n=319). Mixed-effects models were developed to predict non-routine discharge, defined as discharge destinations other than home (e.g., skilled nursing facilities, home health, rehabilitation). Fixed effects included Elixhauser Comorbidity Index, age, sex, insurance status, NRD discharge weighting, and median income by ZIP code. Random effects captured hospital characteristics. Two models were created: one agnostic to WHO obesity classifications and one incorporating these classifications.

Results: The model agnostic to WHO obesity classification yielded an AUC of 0.868, indicating a strong predictive capability for non-routine discharge. Incorporating WHO obesity classifications increased the AUC to 0.870; however, this difference was not statistically significant (p=0.88), suggesting that while BMI plays a role in predicting discharge outcomes, the specific WHO obesity classification does not add significant predictive value in this context.

Conclusion : This study demonstrates that while elevated BMI is associated with non-routine discharge after multilevel scoliosis surgery, the inclusion of specific WHO obesity classifications does not significantly enhance predictive power. These findings suggest that factors beyond obesity classification, such as comorbidities and hospital characteristics, may be more critical in influencing discharge planning and outcomes in this patient population.