Preoperative risk stratification of discharge disposition using the RAPT score in brain tumor surgery.

Purpose

Discharge planning following brain tumor resection is often variable, contributing to delays in care coordination after discharge. While structured discharge prediction tools are widely used in orthopedic and spine surgery, no standardized preoperative framework exists in neurosurgical oncology. We conducted an exploratory analysis of the utility of the Risk Assessment and Prediction Tool (RAPT) for preoperative discharge stratification in supratentorial brain tumor resection patients.

Methods

We performed a retrospective cohort study of adults undergoing supratentorial craniotomy for tumor resection at one institution. Preoperative RAPT scores were recorded, and discharge was categorized as home or non-home. Associations between RAPT score and discharge disposition were assessed using logistic regression. Preliminary discrimination was assessed using receiver operating characteristic (ROC) analysis, and all estimates are apparent.

Results

Among 36 patients, higher RAPT scores were associated with increased likelihood of home discharge (OR=1.74; 95% CI: 1.08-2.80). ROC analysis showed preliminary in-sample discriminatory ability (apparent AUC=0.815, 95% CI 0.632-0.999). An optimal RAPT cutoff of 9.5 (operationalized as ≥10 vs ≤9) was associated with an apparent sensitivity of 89.7% and positive predictive value of 92.9% for home discharge. A threshold-based model modestly improved classification accuracy compared to a naïve classifier (86.1% vs 80.6%).

Conclusion

In this exploratory cohort, preoperative RAPT score was associated with discharge disposition and may help identify patients for earlier discharge planning after supratentorial tumor resection. Preliminary performance was driven principally by the identification of patients destined for home discharge. The modest NPV suggests that the characteristics captured by the tool are necessary but not sufficient determinants of disposition in neurosurgical oncology, where postoperative neurological trajectory, which was not measured in this study, may contribute to the unexplained variation. Low scores may prompt earlier care-coordination rather than definitively predict rehabilitation need. These hypothesis-generating findings warrant prospective, multi-institutional validation to determine a defined role in neurosurgical care pathways.