Alternative shock index cutoffs for pediatric patients outperform the Shock Index Pediatric Age-adjusted (SIPA) on strength of association with adverse outcomes in pediatric trauma patients

Abstract

BackgroundThe shock index is a tool for evaluating critically ill patients that is defined as the ratio of their heart rate divided by systolic blood pressure. The SI is associated with outcomes in adult trauma patients. The Shock Index Pediatric Age-adjusted was developed as a pediatric-specific tool to account for the physiologic differences of children of varying ages. There is growing interest in Shock Index Pediatric Age-adjusted, which is associated with adverse outcomes in pediatric trauma. We hypothesized that alternative shock index cutoffs based on the Advanced Trauma Life Support or the Pediatric Advanced Life Support vital sign reference ranges would outperform Shock Index Pediatric Age-adjusted. MethodsWe analyzed a retrospective cohort of pediatric trauma patients (age 1 to 16 years old) in the American College of Surgeons Trauma Quality Programs Participant Use File from 2010 to 2018. The primary outcome measure was in-hospital mortality. The Shock Index Pediatric Age-adjusted was compared to an Advanced Trauma Life Support–based and a Pediatric Advanced Life Support–based shock index cutoff system. Our findings were subsequently confirmed with a separate, internal validation data set. ResultsA total of 598,830 Trauma Quality Programs Participant Use File patients were included, 0.9% (n = 5,471) of whom died. For mortality, the Advanced Trauma Life Support–based system yielded the highest positive predictive value (15.8%; 95% confidence interval 15.0%–16.7%) compared with the Pediatric Advanced Life Support–based system (4.3%; 95% confidence interval 4.1%–4.5%). Both the Advanced Trauma Life Support–based and Pediatric Advanced Life Support–based systems achieved higher positive predictive values compared to Shock Index Pediatric Age-adjusted (2.6%; 95% confidence interval 2.5%–2.7%). The negative predictive values were not clinically different. Our findings were validated using a separate internal trauma database, in which the positive predictive value for mortality of the Advanced Trauma Life Support–based system was significantly higher than Shock Index Pediatric Age-adjusted (18.2% [95% confidence interval: 8.2%–32.7%] vs 2.9% [95% confidence interval: 1.6%–5.0%], P < .05). ConclusionAdvanced Trauma Life Support and Pediatric Advanced Life Support–based shock index cutoffs achieved higher positive predictive values and similar negative predictive values compared to Shock Index Pediatric Age-adjusted for adverse outcomes in pediatric trauma.