Application value of different frailty assessment tools in older patients undergoing major abdominal surgery

Abstract

Background

Multiple frailty assessment tools are available for clinical practice, but the optimal tool remains unclear. This study aimed to compare the diagnostic performance of frail scale (FS), frailty phenotype (FP),11-item modified frailty index (mFI-11), Edmonton Frail Scale (EFS), and Tilburg Frailty Indicator (TFI) for frailty taking the comprehensive geriatric assessment (CGA) as the gold standard, and their ability to predict 30-day postoperative complications and prolonged length of stay (PLOS).

Methods

This study recruited older patients (≥ 65 years) undergoing elective major abdominal surgery. The receiver operating characteristic (ROC) curves, technique for order preference by similarity to ideal solution (TOPSIS), and decision analysis curve (DCA) were used to validate the diagnostic, comprehensive, and predictive performance of 5 tools in frailty, complications, and PLOS.

Results

EFS presented moderate consistency with CGA (Kappa = 0.544, P < 0.001), excellent performance in diagnosing frailty (area under the ROC curve (AUC) = 0.881, P < 0.001), and high clinical net benefit within the risk threshold ranging from 0.8 % to 57.44 %. Although EFS had the largest AUC for predicting complications (AUC = 0.612) and PLOS (AUC = 0.642) and showed high clinical net benefit, its predictive performance was poor (AUC < 0.7). The TOPSIS indicated that EFS required optimization in multiple aspects (closeness coefficient (Ci) < 0.8).

Conclusion

EFS has excellent diagnostic performance and clinical net benefit for frailty. However, further research is required to identify optimal tools or combine EFS with additional indicators to enhance its comprehensive and predictive performance for complications and PLOS.