Evaluating Frailty Screening Tools for Predicting Morbidity and Mortality for Patients undergoing Spine Surgery

Introduction: The Risk Analysis Index (RAI) is gaining use for assessing surgical frailty, although its predictive power for mortality and inpatient outcomes in spine surgery is not fully defined. This study aimed to evaluate the predictive accuracy of RAI, the modified frailty index-5 (mFI-5), and age on outcomes, including extended length of stay, 30-day readmission, complications, and mortality in spine surgery patients.

Methods: A retrospective cohort study was performed utilizing the 2016-2021 ACS NSQIP database. Adult spine surgery patients were identified via CPT codes, encompassing trauma, degenerative, tumor, and infectious causes. Receiver operating characteristic (ROC) and multivariable analyses were applied to compare RAI, mFI-5, and age in relation to extended length of stay (LOS), adverse events (AE), 30-day readmission, and mortality risk.

Results: Among 352,794 patients studied, 53.1% were male, and 78.9% were non-Hispanic White. RAI categorized 0.4% as frail and 0.1% as very frail, compared to mFI-5's identification of 17.1% frail and 2.0% very frail patients. In predicting extended LOS, RAI and mFI-5 exhibited modest capabilities (AUC 0.6055 and 0.5825); RAI was superior to age (p < 0.001). For 30-day readmission, mFI-5 outperformed RAI (p < 0.001), while no significant difference was found between RAI and age. Regarding 30-day complications, mFI-5 and age surpassed RAI (p < 0.001). RAI, however, was superior for 30-day mortality prediction (AUC: 0.8257 vs. mFI-5 at 0.6551 and age at 0.6834, p< 0.001). Multivariable analysis revealed RAI as a significant predictor of extended LOS for frail [aOR: 10.33, 95% CI: 8.32–12.82] and very frail [aOR: 12.82, 95% CI: 7.38–22.25]. RAI was also a significant predictor of 30-day readmission for frail [aOR: 1.85, 95% CI: 1.51–2.27] and very frail [aOR: 2.60, 95% CI: 1.70–3.98]. RAI did not significantly predict 30-day AEs. However, RAI significantly predicted 30-day mortality for frail [aOR: 8.64, 95% CI: 6.82–10.95] for frail and very frail [aOR: 11.04, 95% CI: 6.87–17.75] patients.

Conclusion : Our study demonstrates the utility of RAI in predicting morbidity and mortality in patients undergoing spine surgery. Despite the development of indices to identify at-risk patients, more robust tools are required for enhanced risk stratification, aiming to optimize patient care and healthcare resources.