From Comprehensive Geriatric Assessment to Rapid Bedside Screening: Comparing MPI and BRASS in Predicting Hospital Outcomes in Older Adults. The S. Giovanni-Addolorata Hospital-SIGOT GRACE Study

Methods: We analyzed data from 896 patients aged ≥ 65 years admitted to the acute care unit of S. Giovanni Addolorata Hospital (Rome, Italy) as part of the GRACE study (2022–2024). The MPI and BRASS scores were collected within 48 and 24 hours of admission, respectively. Associations of each tool with in-hospital mortality, institutionalization, and length of stay (LOS) were assessed using logistic regression analyses. The predictive performance of the two tools was subsequently compared through receiver operating characteristic (ROC) curve analysis. Results: Higher MPI and BRASS risk categories were associated with greater mortality, longer LOS, and higher rates of institutionalization (p < 0.0001 for all). Compared to low-risk patients, high-risk MPI patients were associated with in-hospital mortality (OR 21.28; 95% CI 2.76 164.5; <0.0001), institutionalization (OR 3.30; 95% CI 1.68–6.52; p = 0.004), and non-home discharge (OR 5.93; 95% CI 3.10–11.3; p < 0.0001). Similarly, patients with high-risk BRASS category were associated with higher odds of mortality (OR 4.15; 95% CI 1.37–12.58; p = 0.005), institutionalization (OR 3.09; 95% CI 1.77–5.37; p = 0.0003) and non-home discharge (OR 3,78; 95% CI 2.26–6.73; p < 0.0001) compared to the low risk group. ROC curves showed similar performance of both tools for mortality (AUC 0.785 vs 0.729, p = 0.13), institutionalization (AUC 0.615 vs 0.623, p = 0.73), and non-home discharge (AUC 0.702 vs 0.692, p = 0.54). Conclusions: MPI showed a higher AUC for mortality compared to BRASS although the difference did not reach statistical significance. For discharge-related outcomes, BRASS performed comparably to MPI. Given its simplicity and minimal resource requirements, BRASS may serve as a rapid, cost-effective screening tool to complement – or, when appropriate, precede - full CGA-based MPI evaluation.