Background: Effective prehospital triage is critical to ensure that severely injured patients are rapidly identified and transported to appropriate trauma centres. We synthesised evidence on the predictive performance of prehospital trauma triage tools and their relationship with key outcomes. Methods: Using a predefined corpus of 9 original studies, we conducted a narrative systematic review according to PRISMA principles. Eligible original studies evaluated triage tools (GAP, MGAP, NEWS2, TRISS, RTS, Shock Index, mREMS) using patient-important outcomes; review/guideline articles contextualised under-/over-triage, geriatric triage, and tool heterogeneity. Results: Across diverse settings, several tools showed strong discrimination for mortality and early resuscitative needs. AUROCs were high for MGAP (0.971 and 0.949 in different cohorts), GAP (0.949–0.935), mREMS (0.967), and TRISS (0.934). NEWS2 performed best for very early mortality but less well for longer horizons. Shock Index (prehospital and ΔSI) predicted transfusion, ICU admission, operative need, and short-term mortality signals in selected cohorts. However, under-triage remained common—particularly in older adults—despite guideline updates, and thresholds varied widely between systems. Conclusions: Prehospital triage tools can accurately stratify risk and inform destination decisions, but performance is tool- and population-dependent. Persistent under-triage among older adults and heterogeneity in thresholds support adopting age-attuned criteria, continuous monitoring for NEWS2-based approaches, and rigorous system-level evaluation against under-/over-triage targets.