Risk Analysis Index-Based Model for Predicting Failure to Rescue, Non-Routine Discharge and Extended Length of Stay in Surgical Patients

Introduction: Failure to rescue (FTR), or death within 30 days of a major postoperative complication, is a key metric used in evaluating hospital quality. There is a paucity of literature regarding frailty’s impact on FTR, particularly with the risk analysis index (RAI) metric. This study aims to develop and validate a frailty-driven model to predict mortality, non-routine discharge (NRD), and extended length of stay (eLOS) in surgical patients who experienced major complications.

Methods: This analysis utilized 5,793,589 patients in the 2015–2020 ACS-NSQIP database, focusing on 441,435 individuals with major complications. The study assessed FTR, NRD, and eLOS across four RAI tiers (I-IV), incorporating American Society of Anesthesiologists (ASA) classification. Three models were evaluated: Model 1 (RAI), Model 2 (RAI+ASA), and Model 3 (RAI, ASA, emergency surgery status, and the number of complications). Area Under Curve (AUC) analysis was used to evaluate each model’s performance.

Results: FTR was observed in 36,824 cases (8.3%), while 136,685 patients (33.0%) experienced NRD, and 112,393 (26.1%) had hospital stays longer than 8 days. The distribution of RAI categories was as follows: I (8.6%), II (29.7%), III (43.1%), and IV (18.6%). Adjusted odds ratios (aORs) for mortality were 3.299 (RAI II), 7.586 (RAI III), and 16.313 (RAI IV). For NRD, aORs were 3.271 (RAI II), 8.980 (RAI III), and 11.833 (RAI IV); for eLOS, aORs were 1.802 (RAI II), 2.198 (RAI III), and 2.945 (RAI IV). The AUC values for mortality prediction were as follows: Model 1 had an AUC of 0.705 (0.702–0.708), which increased to 0.781 (0.778–0.783) with Model 2, and 0.846 (0.844–0.848) with Model 3. For NRD, the AUC improved from 0.687 (0.685–0.689) with Model 1 to 0.722 (0.720–0.723) with Model 3. In the case of eLOS, AUC increased from 0.573 (0.571–0.574) with Model 1 to 0.707 (0.705–0.708) with Model 3.

Conclusion : Demonstrated by the RAI, frailty is associated with FTR, NRD, and eLOS in surgical patients. The novel combined frailty index demonstrates superior discrimination in identifying high-risk patients, providing valuable insights for improving surgical care.