Which Trauma Score Best Predicts Life-Saving Care in Injured Children? A New Study Has an Answer

When a critically injured child arrives at an emergency department, the first minutes are decisive. Clinicians must rapidly decide whether the patient needs a life-saving intervention—an airway procedure, a chest tube, a blood transfusion, or an emergency operation—long before the full picture of the injuries is clear. For decades, emergency medicine has relied on a family of scoring systems to make that call: anatomical scores that tally damaged body regions, physiological scores that track vital signs and consciousness, and hemodynamic ratios that flag hidden shock. A new retrospective cohort study from a Level I pediatric trauma center in Turkey now offers one of the most direct head-to-head comparisons of these tools in children, and its findings suggest that no single score is enough—the best predictions come from combining them.

The research, conducted at the Pediatric Emergency Department of Hacettepe University between January 2019 and December 2022, enrolled 559 patients aged 0 to 18 years who presented within one hour of injury and were observed for at least eight hours. The cohort had a median age of 91.4 months—roughly seven and a half years—and was nearly two-thirds male, a distribution that reflects the well-documented epidemiology of pediatric blunt trauma. Of the 559 children, 179, or 32 percent, ultimately required at least one life-saving intervention, giving the researchers a substantial group of positive cases against which to test the predictive power of each scoring system.

The study evaluated five widely used instruments. The Abbreviated Injury Scale, or AIS, grades the severity of injury in individual body regions on a six-point scale. The Injury Severity Score, or ISS, is derived from the AIS by summing the squares of the three most severely injured regions, producing a composite anatomical measure. The Pediatric Trauma Score, or PTS, integrates physiological and anatomical elements—airway status, blood pressure, level of consciousness, open wounds, skeletal injury, and weight—into a single number designed specifically for children. The Shock Index, the simple ratio of heart rate to systolic blood pressure, attempts to capture compensated shock through bedside hemodynamics. Finally, the BIG score—Base deficit, International Normalized Ratio, and Glasgow Coma Scale—combines metabolic, coagulation, and neurological markers into a laboratory-anchored risk estimate.

Because several of these scores are mathematically interdependent—the ISS is built directly from AIS values, for example—the researchers faced a statistical trap familiar to anyone modeling correlated predictors. Stuffing all five scores into one regression would produce unstable estimates distorted by multicollinearity. Their solution was a parsimonious multivariable model built from one representative variable per score family: the ISS for the anatomical domain, the PTS for the integrated physiological-anatomical domain, and the Shock Index for the hemodynamic domain. Each variable was entered into a multivariable logistic regression with life-saving intervention as the outcome, allowing the team to isolate the independent contribution of each domain while controlling for the others.

The results were striking. Higher ISS values independently predicted the need for a life-saving intervention, with an adjusted odds ratio of 1.145 per point (95 percent confidence interval 1.091 to 1.202), while lower PTS values were similarly predictive, with an adjusted odds ratio of 0.552 per point (95 percent confidence interval 0.474 to 0.642); both associations were highly significant at p less than 0.001. The Shock Index, by contrast, failed to reach independent significance (p equals 0.134), a finding that will resonate with pediatric emergency physicians who have long suspected that adult-derived hemodynamic thresholds translate poorly to children, whose robust physiological compensation can mask serious blood loss until decompensation is abrupt and late.

Discriminatory power—the ability of a score to separate children who need interventions from those who do not—was quantified with the area under the receiver operating characteristic curve, or AUC. Among individual scores, the Pediatric Trauma Score led the field with an AUC of 0.947, followed closely by the Glasgow Coma Scale at 0.920 and the ISS at 0.918. The BIG score achieved 0.897 and the AIS 0.878, while the Shock Index trailed dramatically at 0.601, barely better than a coin flip. The combined three-variable model, however, outperformed everything: it achieved an AUC of 0.961 (95 percent confidence interval 0.943 to 0.978), and a formal DeLong test confirmed that this improvement over the best individual score, the PTS, was statistically significant (p equals 0.004).

The metabolic story added an important nuance. In a separate adjusted analysis restricted to the subgroup of patients with complete laboratory data, the researchers examined base deficit and lactate—two markers of tissue hypoperfusion measured on arterial or venous blood gases. After adjustment for injury severity and neurological status, base deficit remained an independent metabolic predictor of intervention requirement, whereas lactate did not. This distinction matters operationally: base deficit reflects the cumulative metabolic debt from anaerobic metabolism and is less susceptible to confounding by liver function, clearance kinetics, and the stress response than lactate, which can rise for reasons unrelated to hemorrhage in children. Crucially, the authors note that these metabolic markers were not components of the combined ISS-PTS-Shock Index model, meaning the multiparametric framework they propose can function without waiting for laboratory results.

The clinical implications are considerable. Triage in pediatric trauma is a high-stakes balancing act: over-triage floods trauma teams and imaging suites with children who turn out to be fine, while under-triage sends seriously injured children to wards unequipped for sudden deterioration. A scoring framework that integrates anatomical burden (ISS), child-specific physiological reserve (PTS), and bedside hemodynamics (Shock Index)—with base deficit layered in once laboratory data become available—offers a staged approach that matches the natural rhythm of emergency care. The near-perfect discrimination of the combined model, if replicated prospectively, could support earlier activation of massive transfusion protocols, earlier involvement of pediatric surgeons, and more rational decisions about which children can safely be observed rather than subjected to whole-body computed tomography.

The study’s limitations deserve honest acknowledgment. As a retrospective single-center analysis at a Level I pediatric trauma center, its findings may not generalize to community hospitals or to health systems with different prehospital care patterns, and the requirement that patients present within one hour of injury and be observed for at least eight hours introduces selection considerations. Retrospective designs also cannot fully exclude the possibility that some scoring data were influenced by treatments already underway. The authors themselves frame the multiparametric model as a framework for early in-hospital risk stratification after initial diagnostic evaluation, not as a prehospital triage tool, and they emphasize that the Shock Index’s weak performance argues against relying on vital-sign ratios alone in children.

Still, the central message is likely to influence practice and future research alike: in pediatric trauma, the whole exceeds the sum of its parts. The Pediatric Trauma Score, the Glasgow Coma Scale, and the Injury Severity Score are each strong individual predictors, but a parsimonious model spanning anatomical, physiological, and hemodynamic domains significantly outperforms the best of them. As emergency departments worldwide grapple with rising pediatric injury volumes and pressure on critical care resources, the study offers a data-driven template for identifying, within the first hour, which injured child truly needs the team’s full attention—and which life-saving intervention should begin before the child has even left the resuscitation bay.

Subject of Research: Predictive performance of clinical, physiological, and trauma scoring systems for identifying life-saving intervention needs in pediatric trauma patients

Article Title: Predictive performance of clinical, physiological, and trauma scores for life-saving interventions in pediatric trauma: a retrospective cohort study

Article References: Gungor, E., Birbilen, A. Z., Yildiz, L. A., & Teksam, O. (2026). Predictive performance of clinical, physiological, and trauma scores for life-saving interventions in pediatric trauma: a retrospective cohort study. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07843-x

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07843-x

Keywords: pediatric trauma, life-saving interventions, Injury Severity Score, Pediatric Trauma Score, Shock Index, Glasgow Coma Scale, base deficit, emergency triage, trauma scores, predictive medicine, pediatric emergency medicine, retrospective cohort study