# Evaluating the predictive value of frailty scores for critical care admission and hospital stay in elderly surgical patients: A comparison of the mFI-5 and CCI

## TL;DR

This study compares two frailty scores to see if they can predict hospital stays and critical care needs in elderly surgery patients, but finds they are not effective.

## Contribution

The study evaluates and compares the predictive value of MFI-5 and CCI for critical care admission and hospital length of stay in elderly surgical patients.

## Key findings

- Neither MFI-5 nor CCI predicted critical care admission in elderly surgical patients.
- MFI-5 and CCI scores did not predict extended hospital length of stay (>5 days).
- The study suggests integrating additional surgical and anesthesia factors may improve predictive accuracy.

## Abstract

Frailty is a critical determinant of postoperative outcomes in elderly patients. Several frailty assessment tools, including the Modified Frailty Index (MFI-5) and the Charlson Comorbidity Index (CCI), have been proposed to predict complications, hospital length of stay (LoS), and critical care admission. However, their comparative predictive value across a broad spectrum of non-cardiac surgeries remains unclear. The purpose of this study was to assess the predictive ability of MFI-5 and CCI in predicting critical care admission and length of stay (LoS).

This single-centre retrospective study analysed data from patients over 65 years of age who attended the preoperative assessment clinic at the Mater Misericordiae University Hospital (MMUH), Dublin, between November and December 2023. MFI-5 and CCI scores were calculated, and their ability to predict hospital LoS (>5 days) and critical care admission was assessed using area under the receiver operating characteristic curve (AUROC) analysis.

Data from 100 patients were included. Critical care admission was required for 20 patients, and the average hospital length of LoS was 4.5 days. AUROC analysis demonstrated that neither the MFI-5 nor CCI were predictive of critical care admission or extended LoS in this cohort.

The findings suggest that MFI-5 and CCI alone may not be sufficient to predict critical care admission or hospital LoS in elderly patients undergoing non-cardiac surgery. Given the multifactorial nature of postoperative risk, future models integrating frailty indices with surgical and anaesthesia-specific factors may enhance predictive accuracy, improve risk stratification, and optimize perioperative resource allocation.