Comparison of minor head trauma management in the emergency departments of a United States and Italian Children’s hospital

Head injury is a leading cause of emergency department (ED) visits for children in hospitals worldwide, and traumatic brain injury (TBI) is a leading cause of pediatric death and disability. When a pediatric patient presents with head trauma to the ED, the clinician has several strategies to use in the management of the patient, primarily observation or cranial computed tomography (CT) evaluation [1‐3]. If there is concern for a potential clinically important traumatic brain injury (ciTBI), which is defined as traumatic brain injury requiring a neurosurgical procedure, hospitalization for > 2 nights or resulting in death, CT is the study of choice for evaluation as it is highly sensitive for detecting intracranial injury [4, 5].

CT use had been steadily increasing over the last decade and then plateaued [6‐8]. Although cranial CT is the gold standard to identify cranial and intracranial lesions, it also exposes patients to ionizing radiation, which is associated with an increased risk of fatal malignancies [9, 10]. The effective use of cranial CT for diagnosing TBI must, therefore, be balanced with the potential malignancy risk to the patient and the cost of the exam.

To guide the clinician in the clinical decision-making in the evaluation of pediatric head trauma, several guidelines have been developed. Some of these include: the U.S. Pediatric Emergency Care Applied Research Network (PECARN), [11] the United Kingdom’s Children’s Head Injury Algorithm for the Prediction of Important Clinical Events (CHALICE), [12] and the Canadian Assessment of Tomography for Childhood Head Injury (CATCH) [13]. In Italy, the Italian Society of Pediatric Emergency Medicine (SIMEUP) and the Italian Society of Pediatrics (SIP) have drafted a national set of guidelines for the management of pediatric head injury [14]. As international consensus on pediatric head trauma management has not been reached, these guidelines are utilized by different hospitals and trauma care facilities internationally to aid in managing pediatric head trauma patients [15‐19]. This may be due to differences in mechanisms of injury and severity of head trauma presenting to different hospitals, as well as different clinical preferences and hospital resources [19].

The objective of the present study is to compare the clinical management of head trauma in two pediatric hospitals, in the U.S and Italy. By comparing the management of head trauma between these two institutions, it may be possible to elucidate whether certain management guidelines are more effective than others and to develop a framework for international recommendations for the evaluation and management of pediatric head trauma.

During the one-year study period, there were 1966 patients in the total study sample who were evaluated for minor head trauma: 1783 children in Boston and 183 children in Trieste. In both hospitals, the majority of patients with head trauma were male. In Trieste, the median age was higher (7.0 years) compared to Boston (5.0 years) (p = 0.007) (Table 5). Neurologic symptoms, as locally defined, were present in 61.2% of Boston’s patients (1092/1783) and 6% of Trieste’s patients (11/183) at the time of presentation to the ED (p < 0.001). Cranial CT was performed on 309 patients in Boston (17.3%) and 12 patients in Trieste (6.6%) (p < 0.001). Among those with cranial CT in Boston, 158 (51%) had the CT performed at the referring institution, prior to coming to Boston Children’s Hospital. Among those seen initially in Boston (not transferred from a referring institution), only 151 (10%) had a CT evaluation for their head trauma. In Boston 154 patients were hospitalized (8.6%), and in Trieste 102 patients were hospitalized (55.7%) (p < 0.001).

We analyzed separately subjects with and without neurologic symptoms after minor head trauma. Inclusion criteria for neurologic symptoms were headache, vomiting, and lethargy. There were 1092 (61.2%) patients in Boston with neurologic symptoms and 11 (6.0%) in Trieste, according to local definition of neurologic symptoms. The median ages respectively were 9.0 years (IQR 3.0–14.0) and 12.0 years (IQR 5.0–16.0) (p = 0.22). In this group a CT scan was performed in 23.5% of Boston patients (257/1092) and in 45.5% of Trieste patients (5/11) (RR 0.52, 95% CI 0.27, 1.00) (Table 6).

There were 691 patients with no neurologic symptoms in Boston and 172 in Trieste. The median ages respectively were 2.0 years (IQR 0.0–5.0) and 6.5 years (IQR 2.0–12.0) (p < 0.001). CT scan was performed in 52 subjects without neurological symptoms in Boston (7.5%) vs 7 (4.1%) in Trieste (RR 1.85, 95% CI 0.86, 4.00). Characteristics (age, sex, presence of neurological symptoms) of patients who had CT scans in the two institutions are described in Table 6.

This study comparing management of minor head trauma between a U.S. and Italian children’s hospital revealed substantial international differences, including the way patients are triaged for care between the two hospitals, the definitions in the triage level and severity of head trauma, and the use of CT for patients between the two institutions. As a result each institution then managed their respective minor head trauma populations differently. Overall, the use of cranial CT was higher in Boston compared to Trieste. However, more children were hospitalized for their head injury in Trieste compared to Boston.

The patient populations with head trauma as well as the assignment of triage levels and definitions of head injury severity varied considerably between the two hospitals, resulting in dissimilar populations presenting to the two hospitals and with differences in the management of these children. In Boston, children presenting with any severity of head trauma are evaluated and managed, including those requiring neurosurgery or intensive care unit (ICU) level of care. Whereas in Trieste, any patient at risk for a more severe head injury is transferred to another facility that has neurosurgical services available. The distribution of triage levels demonstrates a higher proportion of children with lower severity of symptoms presenting in Boston compared to Trieste. This most likely accounts for the significantly higher prevalence of neurological symptoms in the Boston children, compared to those in Trieste.

In Boston, the Pediatric Emergency Care Applied Research Network (PECARN) guidelines are employed to aid clinicians in deciding whether to order a head CT or observe the patient [11]. The Evidence Based Guidelines (EBG) algorithm© [20] (summarized in Table 3) summarizes these guidelines and recommends CT scan only if a patient has 3 or more PECARN risk factors. Children with two or fewer PECARN risk factors are recommended for observation for the development of any signs or symptoms of intracranial injury. In Trieste, an adapted PECARN guideline is employed to aid clinicians in their management decisions for minor head injury patients (Table 4), in accordance with national guidelines regarding pediatric head injury management [14].

The guidelines at both institutions provide clinicians with similar advice regarding observation for low-risk patients and CT use for intermediate-risk patients. Overall, there was a higher rate of CT use in Boston compared to Trieste. The difference in CT use between Boston and Trieste is partially attributable to Boston being the major referral facility for pediatric emergency care in the region. It is interesting to note that 51% of the CT scans for the Boston patients were done at referring hospitals prior to the patient arriving to the Boston Children’s Hospital ED. As Boston Children’ Hospital is one of the largest pediatric trauma centers in the region, many children with traumatic injuries are referred to this institution for more definitive management and care. For many of these patients, if they had been evaluated at Boston Children’s Hospital first, the PECARN guidelines may have recommended observation over CT scan, as only 10% of children who were evaluated initially at Boston Children’s Hospital had a CT performed. The difference in CT use between Boston and Trieste was not statistically significant when stratified by neurologic symptoms, although the sample size from Trieste was limited.

In contrast, the rate of admission for observation was significantly higher in Trieste compared to Boston, and these admissions were not for intensive neurosurgical monitoring or procedures. This demonstrates the differences in management preferences for minor pediatric head trauma between the institutions, which highlights some of the international differences in head trauma management. These global differences in CT rates are likely due to triage differences, institutional differences, and variations in the use of different decision rules. Even with a disparity in CT rates between countries, the effective impact on patient safety and outcomes is minimal. However, there are arguments to be made for the reduction of extraneous CT use, which include the radiation risk to patients [9, 10] and the financial cost incurred [22].

Several head injury decision rules have been developed: PECARN, [11] CATCH, [13] and CHALICE [12]. The most widely used of these is the PECARN rule, which is utilized in modified forms in Boston and Trieste. There have been studies conducted to compare the relative accuracies of these decision rules in various international settings, including Australia [18, 23, 24] and New Zealand [1]. In addition, there are also studies demonstrating that these decision rules can be successfully implemented in various international settings, including: the PECARN rule in Padova, Italy, [21] Nantes, France, [17] Tehran, Iran, [19] Boston, U.S., [25] and Tokyo, Japan [16].

In contrast, a cogent study to evaluate the suitability of implementing one of these rules in Denmark found that the limited availability of CT in Scandinavia would make it inappropriate to implement CHALICE or PECARN guidelines [5]. Thus, clinical decision rules for the management of head trauma may not be universally applicable to all institutions due to differences in institutional resources and patient population with varying mechanisms and severities of injury.

Our study has some limitations. As this was a retrospective database study, it is possible some patients were missed. However, a comprehensive list of ICD-9-CM codes was used to identify patients with head trauma from the electronic medical record in Boston. Some individual medical records may not have included all of the data elements included in the study. Also, between the two hospitals, some data elements were not consistently recorded, which limited some of the analyses. There is also likely some variation in the definition of some of the variables, including the neurologic symptoms, which may account for some of the differences in the population. The sample size between the two hospitals is also quite different with a small proportion of children having a cranial CT in Trieste, which limits interpretation of some of the analyses.

In this study of pediatric minor head trauma patients in Boston and Trieste, we demonstrated variation in the triage of patients, definitions of triage and head trauma severity levels, CT use, and hospital admission. Both hospitals successfully implemented some form of the PECARN head trauma clinical decision rule as their guideline. Institution-specific and population-specific differences are important to consider when implementing these type of guidelines to provide optimal care with appropriate testing.