Purpose: Failure to rescue (FTR), the mortality rate after post-operative complications, is a quality metric contributing to hospital variation in mortality. Patients can ‘spiral’ into complication cascades, challenging hospitals’ capacity to rescue patients from death. This study evaluated how the number and type of complications influence FTR rates following gastrointestinal (GI) and hepatopancreatobiliary (HPB) cancer surgery and how complication rates, clusters and FTR rates vary across hospitals stratified by FTR quartiles. Methodology: A retrospective population-based cohort study was conducted using linked national registry data from NZCR (New Zealand Cancer Registry) and NMDS (National Minimum Dataset). Patients were grouped by the number of complications (1 to 5+), and FTR rates were calculated. Latent class analysis identified common complication clusters. Complication rate, cluster proportion and FTR rates were compared across FTR quartiles. Results: 31,199 patients across 20 hospitals were included from 2005-2020, with 14464 experiencing ≥1 complication. FTR rates increased concordantly with the number of complications (p < 0.001). Hospitals in higher quartiles had significantly higher FTR rates across all complication groups, despite similar complication rates (p < 0.001). Five common clusters of post-operative complications were identified: Surgical Site Infections, Bleeding, Cardiorespiratory, Serious Surgical, and Severe Respiratory. Cluster proportions were comparable across hospitals, but FTR rates were consistently higher in higher quartile hospitals (p < 0.001). Conclusion: Hospitals vary in their capacity to rescue, independent of complication burden or type. System-level factors likely influence this variation, highlighting the need for targeted hospital interventions to recognise and respond to patients in a complication spiral and a part of clusters.