Trauma is a leading cause of mortality among infants and children over the age of one [1]. It is also a significant cause of physical disabilities that may pose a long-term impact on quality-of-life for both children and their families [2,3]. To improve outcomes, regionalization of care for severe pediatric trauma was developed in several western countries, such as the United States, France, and Austria [4–6]. Inter-hospital transfer (IHT) is one of the crucial components in the success of the regionalized trauma systems. A recent meta-analysis of IHT effects in trauma patients revealed an inconsistency in pooled estimated mortality rates between transfer and direct admission groups [7].

Successful development of regional pediatric trauma care systems has necessitated a more meticulous investigation into the factors associated with IHT of pediatric trauma patients. Previous studies have reported that trauma severity or medical insurance status were predictive factors of transfer to a trauma center [8,9]. Whereas in pediatric trauma patients, age has been reported as a key determinant in the decision to transfer [10]. Most previous studies however, have focused only on patients who undergo transfer from non-trauma center [8–10]. Few studies have focused specifically on transfer from hospitals with a high capacity, despite the secondary triage in such hospitals was crucial factors to regionalization strategies.

Having identified these gaps in both research and practice within emergency departments (EDs), the aim of the present study was to investigate factors associated with IHT among pediatric trauma patients who were initially transported to ED with a high capacity for general emergency care (regional emergency centers, RECs).

Of the weighted total of 73,841 injured pediatric patients in the EMS-IPP sample database, 12,176 patients were admitted to, or transferred from hospitals in South Korea. We limited the sample to patients who were transported to the REC. After excluding known information on ISS cases, 2,031 patients were evaluated (Fig. 1).

Table 1 summarizes patient characteristics by age group. Patients aged 15 to 19 years old constituted 49.5% of the total cases. As age increased, the severity of injury rate tended to increase. The proportion of severe injuries was less than 10% for the 0- to 4-year-old group and 5- to 9-year-old group, but 27.2% for the 10- to 14-year-old and 53.5% for the 15- to 19-year-old group. In contrast, among patients under 14 years old, more than half were transferred to another hospital, and among the 15- to 19-year-old group, 34% were transferred.

Table 2 shows patient characteristics by ED disposition. Overall, 47.3% were transferred to other hospitals. The transfer group were younger and had more head injuries than the admission group. There was no statistically significant difference in injury severity between the admission and transfer groups (P=0.34).

After adjusting for potential covariates, age group had a significant effect on transfer. When compared to the 15- to 19-year-old group, the younger age groups had higher odds of transfer (0- to 4-year-old group, aOR 7.65 [95% CI, 1.24 to 47.38]; 5- to 9-year-old group, aOR 14.48 [95% CI, 2.08 to 100.55]; 10- to 14-year-old group, aOR 5.03 [95% CI, 10.5 to 26.79]) (Table 3).

Our study revealed that patient age was independently associated with IHT in moderately to severely injure pediatric trauma patients who were initially transported to a REC. There was no association between injury severity and transfer.

Interestingly, the tendency for transfer in young pediatric trauma patients with severe injuries was revealed in a previous study conducted in non-trauma centers [10]. However, despite that the subjects of this study were pediatric patients in a REC, the effects of age on transfer were similar to those in non-trauma centers. It suggests that a comprehensive review of the management capacity of RECs pediatric patients with severe trauma is necessary.

In order to develop a pediatric trauma system in the future, the current pattern of transfer practice in pediatric trauma patients should be considered. The results of this study are also applicable in guiding development and improvements in the field of pediatric trauma triage and transport protocols of EMS. Age is an important factor in the management of pediatric trauma patients, and the management of younger pediatric trauma patients requires sophisticated skills [15]. The ideal cutoff age for treatment in pediatric trauma centers is controversial. Studies that have evaluated outcome, including mortality among adolescents treated at adult or pediatric trauma centers, have shown conflicting results [16,17]. This cutoff age varies in different trauma systems based on local preferences and practice patterns. Institutional policies on the age limit for treatment in pediatric trauma centers ranged from 12 to 18 years [18–21]. Aside from age, patients with torso injuries were likely to transfer to another hospital. This association between injured body region and transfer decision was similar to that of a previous study conducted within a non-trauma center [9].

In this study, we used nationwide EMS data sampled from a database. There were notable strengths in the database used in this study. First, the EMS-IPP sample database was highly representative of injured pediatric patients transported by EMS. It consists of mandatory registries for all patients accessed by EMS. Taking into consideration the time and cost of data collection, e.g., collecting information on in-hospital mortality, we applied a multi-stage cluster sampling method. We attempted not to compromise the representation of different regions and levels of ED in the data. Second, it is challenging to link EMS data and hospital information, which includes the severity of injury or outcomes. The Act on the Protection of Personal Information of Korea prohibits the use of registration numbers for data linkage. In this study, we probabilistically matched pre-hospital EMS data with in-hospital medical records. This method was successful in previous studies and its reliability has been established for out-of-hospital cardiac arrest and severe trauma [22–25]. Third, the medical record reviewers who participated in the EMS-IPP sample database were trained and had experience on similar projects, such as medical record review and AIS coding in severely injured EMS patients [24,25]. We also operated a data quality management committee system for the reviewers to minimize information bias caused by ISS coding problems.

Several limitations were identified in the present study. First, our analysis did not include patient outcomes such as mortality or disability. This was due to the fact that, although our database included outcomes (such as in-hospital mortality and disability) upon discharge of admitted patients, it did not include outcomes for transferred patients. Second, our database did not include information on the hospitals to which patients were transferred, such as the level of ED or region. Third, we did not evaluate the reasons for the transfer. One example was patient or guardian preference with respect to economic considerations or distance from hometown. Another example is the temporary resource depletion in the REC, i.e., a lack of intensive care units, for which information was not available through a review of medical records. Finally, as this was an observational study, there is a possibility of residual confounding effects.

In summary, our findings suggest that younger pediatric patients with moderate to severe trauma are more likely to be transferred to another hospital despite being initially transported to a REC.