Articles Performance of the Paediatric Trauma Score on survival prediction of injured children at a major trauma centre: A retrospective Colombian cohort, 2011−2019 Ana De los Ríos-Perez,a,b,c* Alberto García,b,c,d,e Laura Cuello,b,c Sara Rebolledo,b,c and Andres Fandi~ no-Losada d,e a noma de Barcelona, Barcelona, Program in Methodology of Biomedical Research and Public Health, Universitat Auto Spain b Fundacio n Valle del Lili University Hospital, Cali, Colombia c Faculty of Health Sciences, Universidad Icesi, Cali, Colombia d Faculty of Health, Universidad del Valle, Cali, Colombia e Cisalva Institute, Faculty of Health, Universidad del Valle, Cali, Colombia Summary Background Despite improvements in children’s health due to a reduction in infections, trauma continues to cause The Lancet Regional many deaths among adolescents. Strategies to mitigate morbidity and mortality from trauma include severity scores Health - Americas to classify and refer patients to the appropriate hospitals to provide better management; however, these strategies 2022;13: 100312 Published online 29 June have not been assessed in Colombian children. This study aimed to describe the characteristics and outcomes of 2022 injured children and evaluate the performance of the Pediatric Trauma Score (PTS) in predicting survival at a major https://doi.org/10.1016/j. trauma centre in a Colombian city. lana.2022.100312 Methods This was a retrospective cohort study of children aged <18 years who were treated for injuries at a hospital in Colombia. The primary outcome was 30-day mortality. A simple logistic regression model was used with PTS as the predictor variable and vital status at discharge as the outcome variable. PTS performance was assessed by dis- crimination using the area under the receiver-operating characteristic (AUROC) curve and by calibration using the Hosmer-Lemeshow (HL) goodness-of-fit test. Findings A total of 1047 children were admitted. The median age was 12 years (interquartile range [IQR]=5−15); 73¢7% were male, and 66¢1% had blunt trauma. The most frequent cause of injury was traffic accident (31¢5%) fol- lowed by assaults (29%). Mortality was 5¢9%; 61¢3% of these deaths occurred in adolescents between 15 and 17 years of age and 71% of deaths in this age group were due to injuries from a firearm. The PTS had a median of 7 (IQR=5 −9), an AUROC of 0¢93, and good calibration (HL=7¢97, p = 0¢33). Interpretation The highest proportion of trauma and death occurred among adolescents. Interpersonal violence was the most frequent cause of death in this age group. The PTS showed good predictive power for survival, with excel- lent discrimination and good calibration. Funding None. Copyright Ó 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/) Keywords: Injuries; Paediatric trauma; Trauma severity indices; Trauma score; Pediatric trauma score; Mortality; Survival prediction Introduction According to the Global Burden of Disease Study, the health of the population worldwide has improved; how- ever, this improvement has been modest in children *Corresponding author. older than 10 years, in whom injuries continue to be the E-mail address: delosrios2@yahoo.com leading cause of death and disability.1,2 Injuries cause (A. De los Rıos-Perez). more years of potential life to be lost before 18 years www.thelancet.com Vol 13 September, 2022 1 Articles than sudden infant death, cancer and infectious dis- eases combined.3 Unintentional injuries are the second Research in context leading cause of death after 5 years of age, and inten- Evidence before this study tional injuries are the third leading cause after 15 years.1 In Colombia, violence is the leading cause of death after This study was conducted in 2019. Before its initiation, a 5 years of age.4 These high injury rates represent a severe PubMed search was conducted using the search terms public health problem requiring multisectoral collabora- ("Pediatrics"[Mesh] OR pediatric* OR child*) AND tive efforts, and are an evident opportunity for preventa- (trauma OR injur*) AND ("survival prediction") without language or year of publication restrictions. No study tive injury intervention. However, when prevention fails, has assessed the prediction of survival in injured chil- the next opportunity for intervention is to provide opti- dren or the performance of PTS in their care in Latin mal care for the injured to impact their survival favour- America. Few studies on paediatric trauma were found ably with the fewest sequelae possible. Strategies such as from Colombia; most descriptive and used a combina- using trauma scores have been proposed; such scores tion of adult and paediatric subjects, and none applied have existed for more than 40 years and have shown vari- trauma scores in children. We repeated the search in able performance results. Nonetheless, each scoring sys- July 2021, and again found no publications in Latin tem has particular benefits that must be considered, America. along with the ease of application and availability of the resources necessary to calculate the scores.5 Trauma Added value of this study scores have prognostic utility, standardise the language Research on paediatric trauma in Colombia is scarce, lim- used by different professionals (prehospital healthcare iting the availability of data to enable the proposal of workers, emergency physicians, intensivists, surgeons, intervention strategies that mitigate the incidence of traumatologists, epidemiologists, and researchers), allow trauma and its devastating consequences. This study interhospital comparisons, and improve resource plan- contributes to scientific evidence in Colombia and Latin America and is one of the region's largest cohorts of pae- ning by providing information about patient severity. diatric trauma studies. Unlike most studies reported The Pediatric Trauma Score (PTS), designed for children, worldwide, this study showed a significant degree of fire- is an available trauma scoring tool. The American Col- arm violence among adolescents; firearm violence was lege of Surgeons Committee on Trauma (ACS COT) in the leading cause of death. Therefore, firearm violence is Advanced Trauma Life Support has highlighted the PTS a relevant public health problem that requires intersec- as a helpful tool for the initial approach to paediatric toral interventions. Results also showed that a high pro- trauma.6 Its performance has not been assessed in portion of children were referred from other hospitals Colombia, a region with trauma characteristics and care with insufficient resources and training, suggesting an systems that differ from the United States, where the opportunity for intervention in prehospital care in both score was developed. The present study is the first in the initial approach of the injured child and their transfer from the scene to the appropriate hospital for continued Colombia and Latin America to evaluate the ability of the health care. Most studies on paediatric trauma in this PTS to predict survival in injured children. This study region have been descriptive. Our study is the first in also aimed to describe the trauma characteristics and out- Colombia and Latin America to assess survival prediction comes of the assessed children. in children with trauma using the PTS. Although interna- tional authorities have historically recommended the use of PTS in the injured, it is not routinely used in our health- care settings; consequently, children are not transferred Methods from trauma sites with objective criteria such as severity Study design and setting classification. Rather, criteria upon transfer are subjective This was an analytical observational study of a retrospec- or administrative, delaying the definitive care of more seriously injured patients. tive cohort of children admitted between January 2011 and May 2019 to the emergency room of Fundaci on Implications of all the available evidence Valle del Lili University Hospital in Cali, Colombia. Ini- tially, the study period was proposed until December Our results showing that high levels of violence are the 2018; however, the period was extended until May 2019 leading cause of death among adolescents will facilitate to achieve an optimal sample size. the proposal of multidisciplinary and intersectoral inter- vention strategies of a preventive nature. The findings The study hospital has 523 beds; 205 are in the inten- support the systematic use of PTS in the care of injured sive care unit. The hospital treats highly complex patholo- children, thereby optimising their transfer to the appro- gies, including severe trauma, and is equivalent to a priate hospital. This prevents wasting of the resources of major trauma centre. Colombia is a country in northern large centres in the care of trivial injuries or delaying the South America, with 48,258,494 inhabitants.7 According care of seriously injured children in centres with insuffi- to the World Bank classification, the country is categor- cient resources, directly affecting patient outcomes. ised as a medium-high income country.8 Cali is a city in the southwestern region with 2,496,442 inhabitants.9 2 www.thelancet.com Vol 13 September, 2022 Articles The study protocol was approved by the hospital's lead to an erroneous classification of individuals Biomedical Research Ethics Committee under registra- (Figure 1). tion number 295−2018. Data sources and measurement Participants The clinical registries of the included patients were Participants included children under 18 years of age reviewed; diagnostic coding was verified, sociodemo- who presented to the emergency room with a trauma graphic information, the mechanism and cause of diagnosis (International Classification of Diseases, 10th trauma, and the type and severity of injuries were edition, discharge codes between S00 and T149) and obtained, and radiology reports were reviewed. Trauma required hospitalisation consisting of a stay longer than time was defined as the time elapsed between injury and 6 h or death before that time. Patients were excluded if hospital admission. This information was collected from they had burns, were involved in a drowning accident, the patient’s clinical history which included the date and were diagnosed as having a foreign body, or were intoxi- time of the injury. Clinical variables such as heart cated. Further exclusions included if more than 24 h rate, respiratory rate, blood pressure, Glasgow Coma had elapsed since the trauma, if they were transferred to Scale score, and the variables necessary to calculate another hospital before day 30 of hospitalisation due to the PTS, among others, were obtained. The primary the impossibility of verifying their vital status at that outcome variable was mortality at discharge, mea- time, or if they were transferred from provinces other sured on day 30 of hospitalisation or earlier if the than Valle del Cauca or Cauca. In addition, patients patient died or was discharged home. We considered were excluded if they had haematological, oncological, surgery, blood transfusion, and intensive care as sec- metabolic, or osteopathic comorbidities that could influ- ondary outcome variables. The predictor variable was ence the approach or outcome, or if they had undergone the PTS. The injury severity score (ISS), a score surgical interventions at another institution due to the widely used by trauma researchers, was calculated impossibility of classifying their initial lesions; there- and compared between living and dead patients and fore, trauma scores could not be calculated which could included in the analysis.10 Figure 1. Participant flow chart. *Haematological, oncological, metabolic, or osteopathic comorbidities. yOther causes: unknown time since trauma, escape, transfer from provinces other than Valle del Cauca or Cauca. www.thelancet.com Vol 13 September, 2022 3 Articles The direct cause of death was determined from death n = 1047 certificates and verified by reviewing the patients’ clinical registries. Median age, years (p25 to p75) 12 (5−15) Age groups 0−4 years 221 (21¢1%) Pediatric trauma score 5−9 years 229 (21¢9%) The PTS was described in 1987 by Tepas et al.11 Its 10−14 years 249 (23¢8%) 15−17 years 348 (33¢2%) application in children combines physiological and ana- Sex tomical variables—weight, airway status, systolic blood Male 772 (73¢7%) pressure, central nervous system status, fractures, and Female 275 (26¢3%) wounds—to classify the severity of the injury. The tool Payor status yields a score between -6 and +12, with lower scores Private insurance 692 (66%) indicating greater severity. A score ≤8 suggests that the Subsidised 355 (34%) patient should be transferred to a trauma centre because Origin (province) of a significantly increased risk of mortality.12 Valle del Cauca 775 (74%) Cauca 272 (26%) Transportationa Injury severity score Ambulance 791 (76¢1%) Private/not ambulance 248 (23¢9%) The ISS was first described in 1974 as the sum of the Transferred from another hospitalb squares of the three highest Abbreviated Injury Scale Yes 725 (70%) (AIS) scores between different body areas.13 A score of No 311 (30%) 16 or more has traditionally been established as a severe Mechanism of injury injury.10 The AIS, necessary to calculate the ISS, was Blunt 692 (66¢1%) described in 1971, and its score ranges from 0 to 6, with Penetrating 355 (33¢9%) 3 being a severe injury and 6 being the worst injury.14 Cause of injury Traffic accident 330 (31¢5%) Assault 304 (29%) Statistical methods Firearm 224 Categorical variables are presented as frequencies and Stabbing 59 Hit by/against something 13 proportions and were compared between survivors and War explosive 8 deceased using either a chi-square or Fisher's exact test, Fall 297 (28¢4%) as appropriate. Continuous variables are presented as Other 110 (10¢5%) the mean and standard deviation or median and inter- Unknown 6 (0¢6%) quartile range (IQR) according to the normality of their Age-adjusted hypotension distribution. Variables were compared between the Yes 74 (7%) deceased and survivors using the Student’s t-test or Wil- No 973 (93%) coxon-Mann-Whitney sum of ranges, as appropriate. Glasgow Coma Scalec 13−15 790 (78%) The performance of PTS in predicting survival was eval- 9−12 79 (7¢8%) uated using a simple logistic regression model, consid- 3−8 144 (14¢2%) ering PTS as a predictor variable and vital status at PTS, median (p25 to p75) 7 (5−9) discharge (alive or dead) as a dichotomous outcome vari- ISS, median (p25 to p75) 9 (5−17) able, with the application of statistical tools such as dis- ICU crimination and calibration. Discrimination using the Yes 521 (49¢8%) area under the receiver-operating characteristic No 526 (50¢2%) (AUROC) curve accurately demonstrates the ability of Blood transfusion Yes 265 (25¢3%) the score to predict survival, with a value of 1 indicating No 782 (74¢7%) perfect discrimination and a value of 0¢5, a predictive Surgery capacity that is not superior to chance. Calibration was Yes 649 (62%) assessed using the Hosmer-Lemeshow goodness-of-fit No 398 (38%) test, which measures the agreement between the Hospital stay, median (p25 to p75) 3¢5 (1¢7−6¢9) observed probability of survival and that predicted by Mortality 62 (5¢9%) the model, indicating p > 0¢05 as a good calibration.15 We hypothesised a good discriminative capacity with Table 1: Sociodemographic and clinical characteristics of the children. an ROC curve greater than 0¢9 and a good fit between Data are n (%) or median (p25 to p75). PTS=pediatric trauma score. observed and expected mortality with a goodness of fit of ISS=injury severity score. ICU=intensive care unit. p > 0¢05. The sample size was calculated using a propor- a Data were available for 1039 patients. b Data were available for 1036 patients. tional-comparison formula. The a priori estimated sam- c Data were available for 1013 patients. ple size was 1043 children, which was estimated to detect 4 www.thelancet.com Vol 13 September, 2022 Articles a difference of 2% between observed and expected mor- common causes of injury were traffic accidents and tality (two-sided 5% significance, power 90%). assaults, accounting for 31¢5% and 29%, respectively; The analyses were performed using StataÒ version the former was more frequent in children under 10 years 16.1 (StataCorp, College Station, Texas 77845, USA). of age, and the latter was more frequent after 10 years of age. The median PTS was 7 (IQR, 5−9). The mortality Role of the funding source rate was 5¢9% (Table 1); mortality occurred in the first There was no funding source for this study. All authors 24 h in 67% of cases, and the median was 15 h (IQR 3¢8 had full access to the study data and final responsibility −49) (Figure 2) without deaths after day 20 of hospital- for the decision to submit for publication. isation. Cranioencephalic trauma was the most frequent cause of death (Table 2). Sixty-one and three tenths per- cent of the deaths were due to intentional trauma, with Results firearms caused 95% of these deaths. Table 2 shows the Between January 2011 and May 2019, 1047 children cause of death according to the anatomical location and aged <18 years were admitted to the emergency room intentionality of injury. and fit the criteria to be included in the study. Among Regarding the secondary outcome variables, 62% those excluded, 29 (2¢7%) were transferred to hospitals underwent surgery, 49¢8% were admitted to the ICU, of a lower level of complexity before the 30th day of hos- and 25¢3% received a blood transfusion. pitalisation. These patients were excluded because their For bivariate analysis, the patients were classified vital status could not be verified at discharge. Reasons according to their vital status at discharge (Table 3). for transferring a patient before 30 days of hospitalisa- Eighty-four percent of the deceased patients were tion were related to administrative issues affecting the between 10 and 17 years of age, and 71% of them died completion of treatment. Thus, 29 patients, considered due to an assault. Most deceased patients were adoles- a low proportion, were transferred with no 30-day sur- cent males with a lower ability to pay, penetrating vival data. However, these patients likely were alive and trauma, admitted in a coma (Glasgow median, 3), with a thus would have a minimal effect on survival estimates. median PTS of 0, and a median ISS of 31 (severe The median age was 12 years (IQR, 5−15 years), and trauma). Almost 25% were admitted to the hospital in a 73¢7% were male. Colombia's subsidised health insur- state of decompensated shock (hypotension adjusted for ance system covered 34% of the total, and the remaining age), indicating a state of extreme severity, with death 66% was covered by the contributory health insurance occurring within the first hour after admission system (private insurance). The injury mechanisms (median, 60 min; IQR, 45−94 min). In this group were blunt (66¢1%) and penetrating (33¢9%). The most of dead children, 51¢6% underwent surgery, 53¢2% Figure 2. Mortality timing. Mortality was measured on day 30 of hospitalisation or earlier if the patient died or was discharged home. Data were available for 815 patients. www.thelancet.com Vol 13 September, 2022 5 Articles Head Chest Abdomen Thoracoabdominal Unintentional 20 1 1 1 Intentional 32 1 2 3 Indeterminate 1 Table 2: Cause of death according to intentionality and anatomical site of the injury. Data are n. Alive (n = 985) Dead (n = 62) p value Median age, years (p25 to p75) 11 (5−15) 15 (13−17) <0¢0001 Age groups <0¢0001 0−4 years 215 (21¢8%) 6 (9¢7%) 5−9 years 225 (22¢8%) 4 (6¢4%) 10−14 years 235 (23¢9%) 14 (22¢6%) 15−17 years 310 (31¢5%) 38 (61¢3%) Sex 0.03 Male 719 (73%) 53 (85¢5%) Female 266 (27%) 9 (14¢5%) Payor status <0¢0001 Private insurance 667 (67¢7%) 25 (40¢3%) Subsidised 318 (32¢3%) 37 (59¢7%) Origin 0¢24 Valle 733 (74¢4%) 42 (67¢7%) Cauca 252 (25¢6%) 20 (32¢3%) Transferred from another hospitala 672 (69%) 53 (87%) Mechanism of injury <0¢0001 Blunt 667 (67¢7%) 25 (40¢3%) Penetrating 318 (32¢3%) 37 (59¢7%) Age-adjusted hypotension 59 (6%) 15 (24¢2%) <0¢0001 Glasgow Coma Scale, median (p25 to p75)b 15 (14−15) 3 (3−6) <0¢0001 PTS, median (p25 to p75) 8 (5−9) 0 (0−1) <0¢0001 ISS, median (p25 to p75) 9 (4−16) 31 (25−37) <0¢0001 ICU <0¢0001 Yes 470 (47¢7%) 51 (82¢3%) No 515 (52¢3%) 11 (17¢7%) Blood transfusion <0¢0001 Yes 232 (23¢6%) 33 (53¢2%) No 753 (76¢4%) 29 (46¢8%) 0¢08 Surgery Yes 617 (62¢6%) 32 (51¢6%) No 368 (37¢4%) 30 (48¢4%) <0¢0001 Table 3: Bivariate analysis: characteristics of patients by vital status at discharge. Data are n (%) or median (p25 to p75). PTS=pediatric trauma score. ISS=injury severity score. ICU=intensive care unit. a Data were available for 1036 patients. b Data were available for 1013 patients. received a blood transfusion, and 82¢3% were admit- goodness of fit of 7¢97 (p = 0¢33; Table 4), indicating ted to the ICU since the remaining percentage died excellent discrimination and good calibration, respec- before ICU admission. tively, and a performance that was maintained for the We found 689 (66%) patients with a PTS ≤8 and different injury mechanisms (blunt or penetrating) 358 with a PTS >8. The performance of the PTS showed (Table 4). Given the classic PTS cut-off point of 8 as the an AUROC of 0¢93 (95% CI 0¢89 to 0¢96; p < 0¢0001) value from which mortality increases significantly, we (Table 4; appendix p 2) and Hosmer−Lemeshow evaluated the performance of the PTS by categorising 6 www.thelancet.com Vol 13 September, 2022 Articles OR p value 95% CI AUROC HL Prob > X2 PTS Complete cohort (n = 1047) 1¢8 <0¢0001 1¢61−2¢02 0¢93 7¢97 0¢33 Blunt trauma (n = 692) 1¢68 <0¢0001 1¢44−1¢95 0¢91 3¢19 0¢87 Penetrating trauma (n = 355) 1¢9 <0¢0001 1¢59−2¢28 0¢93 8¢72 0¢12 Table 4: Performance of Pediatric Trauma Score by the mechanism of injury. OR=odds ratio, CI=confidence interval, AUROC=area under the receiver-operating characteristic curve, HL=Hosmer−Lemeshow. its score, observing an AUROC of 0¢92 (95% CI 0¢89 arrival included coma and median PTS and ISS scores −0¢94) in those with PTS ≤8 and 0¢74 (95% CI 0¢71 of 0 and 31, respectively, values with a high mortality −0¢76) in those with a PTS >8. We “rescued” the pre- rate.12,24 Furthermore, a proportion of them arrived dictive capacity in children with a worse score for whom hypotensive, which in children is considered a terminal referral to the appropriate hospital is a priority because situation with increased mortality25; therefore, it is cru- of their high probability of dying. An additional analysis cial to receive the required care in the shortest time pos- was done to evaluate the performance of the PTS sible. These critical patients should be transferred according to age and the presence of severe head injury directly from the injury site to a major trauma centre (Appendix p 3). because the risk of imminent death is high without ade- quate care and timely intervention.26,27 The proportion of violent trauma evidenced in this Discussion study represents a serious public health problem that This study presents data on the frequency, severity, and must be addressed with investment and combined characteristics of trauma in a paediatric sample in efforts at the local, international and intersectoral lev- Colombia. Traffic accidents and assaults were the most els.28 Such efforts include those proposed by the World frequent causes of injury. The results regarding traffic Health Organization (WHO) in collaboration with the accidents are in accordance with existing evidence from United States Centers for Disease Control and Preven- other researchers worldwide16−20 and within the tion (US CDC), the Pan American Health Organization region.21,22 These injuries are potentially preventable (PAHO), and the United Nations Children's Fund (UNI- with the implementation or optimisation of adequate, CEF), among other organisations, with the implementa- exhaustive, and strong prevention measures linked to a tion of the INSPIRE strategies for preventing violence solid and robust surveillance system. Notably, reported against children.29 Although the prevention of injuries deaths in our study were predominantly intentional and is an ideal concept, injuries do occur despite the best caused by firearms, which contrasts with studies report- preventive efforts, and physicians must be prepared to ing that deaths due to trauma were mostly uninten- treat them accordingly. One strategy with demonstrated tional.1 We also observed that most of the deceased value is the use of trauma scores30 that objectively clas- children had severe cranioencephalic trauma and that sify patients and facilitate routing to an appropriate death occurred quickly, similar to studies that analysed trauma centre; such scores have been used in paediat- trauma mortality.23 The proportion of deceased who did rics since the 1980s. However, their performance in not receive surgery or require admission to the ICU was survival prediction has not been evaluated in Latin attributed to the extreme severity of their admission sta- America. The PTS was designed in a developed country tus, with rapid progression to brain death and death. (United States), with social characteristics different Although mortality from injuries has decreased world- from those of our country, as well as an organised road wide, injuries remain among the leading causes of infrastructure, a more specialised, structured and coor- death,1 and many young lives continue to be lost. In dinated health and trauma care system, and trauma Colombia, violence is the leading cause of death after characteristics different from those of our environment. 5 years of age4; this is a serious and critical social prob- Additionally, the PTS has been implemented for several lem and of major concern considering that these victims decades, during which time there have been progressive are children. These deaths are “the tip of the iceberg” as technological developments and the creation of particu- more injured children survive but require hospital care lar protocols for caring for injured patients. It was or are left with disabilities.16 The considerable percent- important to assess the effectiveness of the PTS in our age of hospitalised patients who were transferred from environment. We found that the PTS was considerably hospitals with a lower level of complexity, mainly those lower in deceased patients. The critical value at which who died, is a noteworthy finding of our study. We do infant mortality has traditionally been considered to not have information on the treatment received in the increase was reported by Tepas as 8.12 However, in our originating hospitals; however, the patient status upon study, we evaluated the cut-off point in the ROC curve www.thelancet.com Vol 13 September, 2022 7 Articles with greater sensitivity and specificity, finding this the use of safety measures in traffic accidents. The study threshold at 4, very different than the threshold pro- was monocentric, which limits the generalisability of posed by Tepas.12 A study by Orliaguet et al.31 had simi- the results. Including other geographical contexts lar results to ours. Like our study, their study used ROC implies collaborative inter-institutional efforts, but gives curves, one of the statistical tools with greater diagnostic us privileged information, which will be interesting to accuracy in this type of analysis.32,33 We evaluated the consider in the future. Regarding the strengths of our capacity of the scoring tool to predict survival, finding study, the sample size was appropriate compared to an excellent performance in both discrimination and other studies on paediatric trauma in the region and calibration, which has also been demonstrated by included a significant proportion of patients with violent researchers in other countries.17,31,34−36 Other studies, trauma, which is low in most paediatric studies. such as that of Yoon et al.37 reported that although the In conclusion, most deaths in this study were caused performance of the PTS was good, it was lower than by intentional trauma (violence), and firearms were that of other scores such as the BIG score (composed of used in almost all deaths. Cranioencephalic trauma was the base deficit [B], International normalized ratio [I], the leading cause of death. PTS showed an excellent Glasgow Coma Scale [G]). However, those scores ability to predict survival, regardless of the mechanism require laboratory tests, resulting in additional time of injury. Prospective studies are needed; ideally, these waiting for results or a complex calculation in some con- would include the analysis of regional trauma records to texts. Furthermore, it may not be available at all levels of allow epidemiological surveillance at the hospital level. care or in the prehospital setting. There are several rea- Such studies would complement the surveillance car- sons for the good performance of the PTS found in our ried out by government entities, and thus reduce the study. First, PTS combines anatomical and physiologi- knowledge gap and enhance planning strategies that cal factors, including weight, among its variables, recog- would impact the health and well-being of children. nising that the severity of injuries and responses differ depending on age or body size.5,38 Second, PTS evalu- ates the central nervous system, which is of great rele- Contributors vance because with multisystem involvement as often ADlR-P, AG and AF-L contributed to the conceptualisa- occurs with severe trauma, the level of consciousness is tion, methodology, statistical analysis, and interpreta- one of the main signs of hypoperfusion and haemody- tion of data and administered the project. ADlR-P did namic instability that occurs even before hypotension. the data curation and wrote the original draft. AG and Third, as an element of quality control, the PTS makes AF-L provided academic resources, supervised the proj- it possible to identify deaths with less severe injuries, in ect, and validated the data. LC and SR contributed to which handling errors could have occurred. However, data collection and wrote different draft versions. All the current study did not address the characteristics of authors had full access to all of the data, critically the treatment received as this was not part of the stated reviewed and approved the manuscript's final version, objectives. Added to these strengths is the ease of calcu- and had final responsibility for the decision to submit lation of the PTS in both prehospital and in-hospital set- for publication. tings by medical or non-medical health personnel. In addition to its good results, it is reasonable to systemati- cally include PTS in the care for injured children, as the Data sharing statement American College of Surgeons recommends in its According to the Helsinki Declaration, the datasets gen- Advanced Trauma Life Support (ATLS) manual.6 The erated and analysed during the current study are not PTS quantifies the severity of trauma, supports efficient publicly available to protect the rights of research partic- resource allocation and use for the care of injured chil- ipants. However, these are available from the corre- dren and the dissemination of data and experiences in sponding author upon reasonable request. hospitals, and generates scientific evidence that objec- tively helps in public health decision-making. With its use, we, as healthcare providers in our country, could Declaration of interests contribute to and align ourselves with the Global Strat- We declare no competing interests. egy for the Health of Women, Children, and Adoles- cents, which seeks, among other objectives, to reduce preventable mortality and improve health for children Acknowledgements and adolescents by 2030.39 We thank the Clinical Research Center of Fundaci on This study had the following limitations. The study Valle del Lili University Hospital for providing support was retrospective with inherent limitations. We did not during this study. Ana Marıa de los Rios is a PhD candi- have information about prehospital care, the treatment date in the Doctorate Program on Biomedical Research received before arrival at our hospital, or the specific cir- and Public Health, Universitat Aut onoma de Barcelona, cumstances under which the trauma occurred, such as Barcelona, Spain. 8 www.thelancet.com Vol 13 September, 2022 Articles Supplementary materials Journal Europeen des Urgences. 2005;18(3):131–139. https://doi.org/ Supplementary material associated with this article can 10.1016/S0993-9857(05)82480-X. 19 Nwanna-Nzewunwa O, Ngamby MK, Cox J, et al. Epidemiology be found in the online version at doi:10.1016/j. and cost of pediatric injury in Yaounde, Cameroon: a prospective lana.2022.100312. study. Eur J Trauma Emerg Surg. 2020;46(6):1403–1412. https:// doi.org/10.1007/s00068-019-01104-6. 20 Diamond IR, Parkin PC, Wales PW, et al. Preventable pediatric trauma deaths in Ontario: a comparative population-based References study. J Trauma. 2009;66(4):1189–1194. https://doi.org/ 1 Institute for Health Metrics and Evaluation (IHME). GBD Com- 10.1097/TA.0b013e31819adbb3. discussion 1194−5. pare. Seattle, WA: IHME, University of Washington; 2015. http:// 21 Botelho F, Truche P, Mooney DP, et al. Pediatric trauma primary vizhub.healthdata.org/gbd-compare. Accessed 22 April 2021. survey performance among surgical and non-surgical pediatric pro- 2 Vos T, Lim SS, Abbafati C, et al. Global burden of 369 diseases and viders in a Brazilian trauma center. Trauma Surg Acute Care Open. injuries in 204 countries and territories, 1990-2019: a systematic 2020;5(1):e000451. https://doi.org/10.1136/tsaco-2020-000451. analysis for the Global Burden of Disease Study 2019. Lancet. 22 Cleves D, Gomez C, Davalos DM, et al. Pediatric trauma at a gen- 2020;396:1204–1222. https://doi.org/10.1016/S0140-6736(20) eral hospital in Cali, Colombia. J Pediatr Surg. 2016;51(8):1341– 30925-9. 1345. https://doi.org/10.1016/j.jpedsurg.2016.01.008. 3 Shook J, Chun T, Conners G, et al. Management of pediatric trauma. 23 Gunst M, Ghaemmaghami V, Gruszecki A, Urban J, Frankel Pediatrics. 2016;138(2):e20161569. https://doi.org/10.1542/peds.2016- H, Shafi S. Changing epidemiology of trauma deaths leads to a 1569. bimodal distribution. Proc (Bayl Univ Med Cent). 2010;23 4 Global Health Estimates 2020: Deaths by Cause, Age, Sex, by Country (4):349–354. https://doi.org/10.1080/ and by Region, 2000-2019. Geneva: World Health Organization; 08998280.2010.11928649. 2020. https://www.who.int/data/gho/data/themes/mortality-and- 24 Ali B, Fortun M, Belzunegui T, Reyero D, Castro M. Escalas para global-health-estimates/ghe-leading-causes-of-death. Accessed 23 prediccion de resultados tras traumatismo grave. An Sist Sanit April 2021. Navar. 2017;40(1):103–118. https://dx.doi.org/10.23938/assn.0001. 5 St-Louis E, Seguin J, Roizblatt D, Deckelbaum DL, Baird R, Razek T. 25 Alberto EC, McKenna E, Amberson MJ, et al. Metrics of shock in pedi- Systematic review and need assessment of pediatric trauma outcome atric trauma patients: a systematic search and review. Injury. 2021;52 benchmarking tools for low-resource settings. Pediatr Surg Int. 2017;33 (10):3166–3172. https://doi.org/10.1016/j.injury.2021.06.014. (3):299–309. https://doi.org/10.1007/s00383-016-4024-9. 26 Dharap SB, Kamath S, Kumar V. Does prehospital time affect sur- 6 American College of Surgeons. Advanced Trauma Life Support. 10th vival of major trauma patients where there is no prehospital care? J ed. Chicago: American College of Surgeons, Committee on Postgrad Med. 2017;63(3):169–175. https://doi.org/10.4103/0022- Trauma; 2018. 3859.201417. 7 DANE. Censo Nacional de Poblaci on y Vivienda-2018. 2018. Bogota, 27 American College of Surgeons Committee on Trauma. Prehospital Colombia. https://www.dane.gov.co/index.php/en/estadisticas- trauma care. In: Rotondo MF, Cribari C, Smith RS, eds. Resources por-tema/demografia-y-poblacion/censo-nacional-de-poblacion-y- for the Optimal Care of the Injured Patient. Chicago, IL: American vivenda-2018/cuantos-somos. Accessed 17 April 2021. College of Surgeons; 2014:23–29. 8 The World Bank. Data for Colombia, Upper middle income. 28 Robertson LS. Human factors. Injury Epidemiology. 4th ed. Morris- Washington DC: World Bank; 2021. https://data.worldbank.org/? ville, NC, USA: Lulu Books; 2015. locations=CO-XT. Accessed 17 April 2021. 29 Global Status Report on Preventing Violence Against Children. 9 DANE. Proyecciones de Poblacion departamentales y municipales Geneva: World Health Organization; 2020. License: CC BY-NC-SA por area 2005-2020. Bogota, Colombia; 2010. https://www.dane. 3.0 IGO. gov.co/index.php?option=com_content&view=article&id=853&Ite 30 St-Louis E, Deckelbaum DL, Baird R, Razek T. Optimizing the mid=28&phpMyAdmin=3om27vamm65hhkhrtgc8rrn2g4. assessment of pediatric injury severity in low-resource settings: Accessed 17 April 2021. consensus generation through a modified Delphi analysis. Injury. 10 Brown JB, Gestring ML, Leeper CM, et al. The value of the injury 2017;48(6):1115–1119. https://doi.org/10.1016/j.injury.2017.03.013. severity score in pediatric trauma: time for a new definition of 31 Orliaguet GA, Meyer PG, Blanot S, et al. Predictive factors of outcome severe injury? J Trauma Acute Care Surg. 2017;82(6):995–1001. in severely traumatized children. Anesth Analg. 1998;87(3):537–542. https://doi.org/10.1097/TA.0000000000001440. https://doi.org/10.1097/00000539-199809000-00006. 11 Tepas JJ, Mollitt DL, Talbert JL, Bryant M. The pediatric trauma 32 Swets JA. Measuring the accuracy of diagnostic systems. Science. score as a predictor of injury severity in the injured child. J Pediatr 1988;240(4857):1285–1293. https://doi.org/10.1126/science.3287615. Surg. 1987;22(1):14–18. https://doi.org/10.1016/s0022-3468(87) 33 Hajian-Tilaki K. Receiver Operating Characteristic (ROC) curve analy- 80006-4. sis for medical diagnostic test evaluation. Caspian J Intern Med. 2013;4 12 Tepas JJ, Ramenofsky ML, Mollitt DL, Gans BM, DiScala C. The (2):627–635. PMID: 24009950; PMCID: PMC3755824. pediatric trauma score as a predictor of injury severity: an objective 34 Breaux Jr CW, Smith G, Georgeson KE. The first two years’ experi- assessment. J Trauma. 1988;28(4):425–429. https://doi.org/ ence with major trauma at a pediatric trauma center. J Trauma. 10.1097/00005373-198804000-00001. 1990;30(1):37–43. https://doi.org/10.1097/00005373-199001000- 13 Baker SP, O'Neill B, Haddon Jr W, Long WB. The injury severity 00006. score: a method for describing patients with multiple injuries and 35 Lironi A, Zawadynski S, La Scala G, Thevenod C, Le Coultre C. evaluating emergency care. J Trauma. 1974;14(3):187–196. PMID: [Value of the Pediatric Trauma Score in routine hospital practice 4814394. −apropos of a prospective one-year trial]. Swiss Surg. 1999;5 14 Rating the severity of tissue damage. I. The abbreviated (6):271–275. https://doi.org/10.1024/1023-9332.5.6.271. scale. JAMA. 1971;215(2):277–280. https://doi.org/10.1001/ 36 Grinkeviciute DE, Kevalas R, Saferis V, Matukevicius A, Ragaisis V, jama.1971.03180150059012. Tamasauskas A. Predictive value of scoring system in severe pediat- 15 Hosmer DW, Lemeshow S, Sturdivant RX. Applied Logistic Regres- ric head injury. Medicina (Kaunas). 2007;43(11):861–869. PMID: sion. 3rd ed. Hoboken: Wiley; 2013. 18084143. 16 Committee on Trauma, American College of Surgeons. National 37 Yoon TJ, Ko Y, Lee J, Huh Y, Kim JH. Performance of the BIG trauma data bank: Annual/Pediatric Report 2016. Chicago, IL. score in predicting mortality in normotensive children with https://www.facs.org/media/d3ufvsmy/ntdb-pediatric-annual- trauma. Pediatr Emerg Care. 2021;37(12):e1582–e1588. https://doi. report-2016.pdf. Accessed 13 May 2021. org/10.1097/PEC.0000000000002122. 17 Yousefzadeh-Chabok S, Kazemnejad-Leili E, Kouchakinejad-Eram- 38 Eastern Association for the Surgery of Trauma. Triage of the sadati L, et al. Comparing pediatric trauma, Glasgow Coma Scale Trauma Patient. Chicago, IL; 2010. https://www.east.org/educa and Injury Severity scores for mortality prediction in traumatic tion-career-development/practice-management-guidelines/details/ children. Ulus Travma Acil Cerrahi Derg. 2016;22(4):328–332. triage-of-the-trauma-patient. Accessed 13 May 2021. https://doi.org/10.5505/tjtes.2015.83930. 39 Kuruvilla S, Bustreo F, Kuo T, et al. The global strategy for wom- 18 Assez N, Hubert H, Boddaert AC, Goldstein P. Score predictif et en’s, children’s and adolescents’ health (2016−2030): a road- traumatismes graves de l’enfant: Inter^ets et limites du Pediatric map based on evidence and country experience. Bull World Trauma Score (PTS) dans l’evaluation du devenir a un an des Health Organ. 2016;94(5):398–400. https://doi.org/10.2471/ enfants traumatises. Etude retrospective d’un SAMU Regional. BLT.16.170431. Accessed 7 December 2021. www.thelancet.com Vol 13 September, 2022 9