# Risk analysis index demonstrates superior predictive performance compared to traditional frailty metrics in elderly female patients undergoing inpatient total shoulder arthroplasty

## TL;DR

The Risk Analysis Index better predicts outcomes like discharge and hospital stay length in elderly women undergoing shoulder surgery compared to traditional methods.

## Contribution

The RAI outperforms mFI-5 and GNRI in predicting key outcomes for elderly female TSA patients.

## Key findings

- RAI had higher AUCs for non-home discharge (0.784) and extended length of stay (0.670) compared to mFI-5 and GNRI.
- RAI showed competitive performance across secondary outcomes like mortality and complications.
- Results support RAI as a preferred tool for risk stratification in discharge planning for this population.

## Abstract

While frailty assessment has become integral to preoperative risk stratification, the optimal measurement tool remains unclear for elderly women undergoing total shoulder arthroplasty (TSA). This study compared the predictive performance of the Risk Analysis Index (RAI) against traditional metrics, including the modified frailty index-5 (mFI-5) and Geriatric Nutritional Risk Index (GNRI) in this specific population.

We conducted a retrospective analysis of ACS NSQIP data from 2015–2021, including female patients aged 65–89 undergoing inpatient TSA. RAI incorporates age, functional status, recent weight loss, and physiological markers, including renal failure, congestive heart failure, and dyspnea. The mFI-5 assesses five comorbidities (diabetes, hypertension, COPD, heart failure, functional dependence), while the GNRI evaluates nutritional status using albumin and body weight. The discriminative ability of RAI, mFI-5, and GNRI was assessed using area under the curve (AUC) analysis for multiple 30-day outcomes. Primary outcomes were non-home discharge and extended length of stay (≥ 4 days), selected based on their clinical importance for discharge planning and quality metrics. Secondary outcomes included 30-day mortality, major and minor complications, readmission, and reoperation. Discriminative ability was assessed using area under the curve (AUC) analysis. Internal validation was performed using bootstrap resampling.

Among 11,965 patients analyzed, RAI demonstrated superior predictive performance for primary outcomes with AUCs of 0.784 for non-home discharge and 0.670 for extended length of stay, significantly outperforming mFI-5 (AUCs 0.601 and 0.590, respectively) and GNRI (AUCs 0.544 and 0.543). For secondary outcomes, RAI maintained competitive performance across mortality, complications, readmissions, and reoperations.

The Risk Analysis Index provides superior discrimination for non-home discharge and extended length of stay compared to traditional frailty measures in elderly female TSA patients, with particularly strong predictive performance for discharge disposition, supporting its adoption as the preferred risk stratification tool for discharge planning in this population.

Video Abstract

Video Abstract

The online version contains supplementary material available at 10.1186/s42836-025-00366-3.