2017

Aims: Clinical decision rules (CDRs) can be applied in Emergency Departments (EDs) to optimise the use of computed tomography (CT) in children with head trauma. The national Emergency X-Radiography utilization Study II (NEXUS II) CDR, as amended for children, has not been externally validated in a large paediatric cohort.  The objective of the study was to conduct a multicentre external validation of NEXUS II CDR in children.
Methods: We performed a prospective observational study of patients < 18 years presenting with head trauma of any severity to 10 Australian/New Zealand EDs.  In a planned secondary analysis we assessed the performance of the NEXUS II CDR for its diagnostic accuracy (with 95% confidence intervals CI) of clinical important intracranial injury (ICI) as identified in CT scans performed in ED.
Results: Of 20,137 total patients, we excluded 28 with suspected penetrating injury. Median age was 4.2 years. CTs were obtained in ED for 1962 (9.8%), of whom 377 (19.2%) had a clinically important ICI as defined by NEXUS II. 74 (19.6%) patients underwent neurosurgery.  Sensitivity for clinically important ICI based on the NEXUS II CDR was 373/377 (98.9%; 97.3%-99.7%) and specificity 156/1585 (9.8%; 8.4%-11.4%). Positive and negative predictive values were respectively 373/1802 (20.7%; 18.8%-22.6%) and 156/160 (97.5%; 93.7%-99.3%). Of the 18,147 children without CT 49.5% had at least one NEXUS II risk criterion.
Conclusion: NEXUS II had very high sensitivity when analysed with a focus on head injured patient who had a CT performed, similar to the derivation study. With half of unimaged patients positive for NEXUS II risk criteria the use of this rule has the potential to increase the number of CTs.

Aims: Clinical decision rules (CDRs) can assist in determining the need for computed tomography (CT) in children with head injuries (HIs). We assessed the accuracy of 3 high quality CDRs (PECARN, CATCH and CHALICE) in a large prospective cohort of head injured children. In addition to rule accuracy, however, among a number of factors physician accuracy is also important when determining whether a particular rule should be implemented.  The objective of the study was to assess the accuracy of physician practice in detecting clinically important traumatic brain injuries.
Methods: Prospective observational study of children <18 years with HIs of any severity at 10 mainly tertiary Australian/New Zealand centres. We extracted a cohort of children with mild HIs (GCS 13-15, presenting <24h) and assessed physician accuracy for the standardised outcome of clinically important traumatic brain injury (ciTBI) and compared this with the accuracy of PECARN, CATCH and CHALICE. Physician accuracy was defined as CT obtained during the initial ED visit.
Results: Of 20,137 children, 18,913 had a mild HI as defined. Of these 1,578 (8.3% = actual CT rate) received a CT scan during the ED visit, 160 (0.8%) had ciTBI and 24 (0.1%) underwent neurosurgery. Physician practice for detecting ciTBI based on CT performed had a sensitivity of 157/160 ((98.1% (94.6% – 99.6%) and a specificity of 17,332/18,753 (92.4% (92.0% – 92.8%)). Sensitivity of PECARN <2 years was 42/42 (100.0%, 91.6% to 100.0%), PECARN >2 years 117/118 (99.2%; 95.4% to 100.0%), CATCH (high/medium risk) 147/160 (91.9%; 86.5% to 95.6%) and CHALICE 148/160 (92.5%; 87.3% to 96.1%). Projected CT rates for PECARN <2/>2 years was 8.0%/9.4% (high risk only) to 41.4%/48.5% (high and intermediate risk, considering the unlikely scenario that all patients in the intermediate risk group receive a CT scan), CATCH 30.2% (medium and high risk) and CHALICE 22.0%.
Conclusions: Physician accuracy was high. The application of PECARN, CATCH or CHALICE CDRs in this setting has the potential to increase the CT rate with limited potential to increase the accuracy of detecting ciTBI.

Aims:Mild traumatic brain injury (TBI) in children is a major public health issue.  Yet there is a wide variation in the way “mild” TBI is defined in literature and guidelines. To date, no study has prospectively detailed the proportion of children presenting with mild TBI to emergency departments (EDs) that are identified by the various definitions.  The objective of the study was to apply published definitions of mild TBI to a large prospectively collected data set of head injuries and to determine the proportions of mild head injuries included by various definitions.
Methods:Prospective observational study of children with HIs of any severity at 10 Australian/New Zealand centres. We applied 18 different definitions of mild TBI, identified through a systematic review of the literature, to children aged 3 to 16 years. We assessed the number and percentage of cases the definitions applied using described inclusion and exclusion criteria.
Results:Of 20,137 children with HI of any severity, 11,907 were aged 3 to 16 years. Mean age was 8.2 years, 32% were female. 61.9% were fall related. Cranial CT rate was 12.7% and neurosurgery rate was 0.5%.  Adjustments were made to the definitions to apply to the data set: none in 7, minor in 9, substantial in 2. Percentages of the cohort covered by the definitions of “mild” TBI ranged from 2.4% (284) to 98.7% (11,756) of the cohort. The median percentage of definitions among 18 definitions investigated was 21.7% (2,589).
Conclusions:When applying different definitions of mild TBI to a single data set including all severities, a wide range of cases are included. Clinicians and researchers need to be aware of this important variability when attempting to apply the published literature to children presenting to EDs with TBI in the Australian and New Zealand setting.

Background: While the majority of head injuries in children are non intentional, there is limited information on intentional injuries outside abusive head trauma.
Objective: To investigate intentional head injuries in terms of demographics, epidemiology, and severity.
Methods: Planned secondary analysis of prospective multicentre cohort study of children aged <18 years across 10 centres in Australia and new Zealand.  Victorian state epidemiology codes (intent, activity, place, mechanism) were used to prospectively code the injuries. Data were descriptively analysed.
Results: Intentional injuries were found in 441 of 20,137 (2.2%) head injured children.  Injuries were caused by peers (166, 37.6%), by caregiver (103, 23.4%), by sibling (47, 10.7%), due to attack by stranger (35, 7.9%), attack by person with unknown relation to patient (19, 4.3%), intentional self-harm (7, 1.6%), other cause or undetermined (64, 28.2%).  Children were <2 years old in 75.7% injuries caused by caregivers and 8.3% in other causes.  Overall, 71.0% of victims were male.  Admission rates varied from 77.7% for assault by caregiver, 37.1% attack by stranger, 22.3% by peer and 8.5% by sibling.  Peer assaults were related to sports in 69 cases (41.6%), with highest frequency in contact sports, rugby (40.6%) and Australian football (26.1%).
Conclusion: Intentional head injuries are infrequent in children. The most frequent cause is peer assault, often during sports, and injuries caused by caregivers.

Background: Abusive head trauma (AHT) is associated with high morbidity and mortality.  It may be difficult to detect in the emergency department (ED).
Objective: To determine how cases of suspected AHT differ from non abusive head injuries in the ED setting.
Methods: Planned secondary analysis of prospective multicentre cohort study of children aged <18 years across 10 centres in Australia and new Zealand with head injuries.  We identified cases of suspected AHT when ED clinicians identified such suspicion on a clinical report form or based on research assistant assigned epidemiology codes.  We compared suspected AHT cases and non AHT cases using risk ratios with 95% CIs.  We assessed the rate of clinically important traumatic brain injuries (ciTBI: death; neurosurgery; intubation >1 day, admission >2 days with abnormal CT scan).
Results: AHT was suspected in 103 of 20,137 (0.5%) head injured children.  Mean age was 2.4 years (SD 4.0). GCS was <12 in 9 (8.7%), 12 (11.7%) presented with seizures, 71 (68.9%) underwent CT head, of which 49 (69%) were abnormal.  80 (77.7%) children were admitted.  Neurosurgery was performed in 3 and 2 children died.  23 (22.3%) had ciTBI.  RR (95% CI) for AHT vs non AHT were: LOC 1.4 (0.7-2.7), vomiting 1.6 (1.202.2), scalp haematoma 2.3 (2.0-2.6), GCS <12 8.5 (4.5-16.0), abnormal neuroimaging 16.4 (13.2-20.4), neurosurgery 7.4 (2.4-22.9), mortality 29.9 (6.8-130.9), ciTBI 17.4 (6.8-25.5).
Conclusion: In the ED clinical presentation in children with suspected AHT differ from non AHT cases.  Suspected cases of AHT are at increased risk of abnormal CT scans, ciTBI and death.

Background: Missing cases of abusive head trauma (AHT) can have severe consequences for affected children. Therefore a 4-variable AHT clinical prediction rule (CPR) for use in the Paediatric Intensive Care Unit (PICU) has been derived and validated for children <3 years by the Pediatric Brain Injury Research Network (PediBIRN).
Objective: We set out to externally validate the CPR applying the tool as designed (PICU only) as well as using broader inclusion criteria (all admitted children with HI).
Methods: In a planned secondary analysis of a prospective multicentre study of paediatric head injuries (HI) of any severity, we extracted patients at 3 tertiary paediatric centres who were either 1) prospectively identified by clinician suspicion for NAI, 2) had epidemiology codes for NAI assigned by a research assistant or 3) were part of a high-risk group (age < 3 years, admitted, abnormal neuroimaging). At one centre we also extracted HI cases from the forensic data base. Based on medical records we assigned all potential AHT cases as AHT positive, negative or indeterminate based on multidisciplinary review and applied PediBIRN criteria 1) as published (PICU only) and 2) using broader inclusion criteria (any admission).
Results: Medical records of 157 possible AHT cases were reviewed. 83.4% were aged less than 2 years, 59.2% were male, 75.8% had an abnormal CT scan and 3.2% died due to head injury. Out of these, 121 had cranial or intracranial injuries confirmed by neuroimaging and were categorized as AHT positive in 30 cases (24.8%), AHT indeterminate in 18 cases (14.9%) and AHT negative in 73 cases (60.3%) based on the decision of a multidisciplinary child protection team. 26 cases met the original inclusion criteria and using the tool as designed, it detected 12 out of 12 AHT positive cases (sensitivity = 100.0% (70.0% û 100.0%)) and correctly categorized 1 out of 8 AHT negative cases (specificity = 12.5% (0.6% û 53.3%)). Using broader inclusion criteria of admitted patients, 117 cases were included and the CPR demonstrated a sensitivity of 27/28 (96.4% (80.0% û 99.8%)) and a specificity of 31/71 (43.7% (32.1% û 55.9%)).
Conclusion: The 4-variable AHT CPR derived and validated by the PediBIRN investigators demonstrated high sensitivity using the tool as designed for PICU patients as well as in a broader group of admitted patients.

Background: Clinical decision rules (CDRs) can be applied in Emergency Departments (EDs) to optimise the use of computed tomography (CT) in children with head trauma. The National Emergency X-Radiography Utilization Study II (NEXUS II) CDR, as amended for children, has not been externally validated in a large pediatric cohort
Objective: To conduct a multicentre external validation of NEXUS II CDR in children.
Methods: We performed a prospective observational study of patients < 18 years presenting with head trauma of any severity to 10 Australian/New Zealand EDs. In a planned secondary analysis we assessed the performance of the NEXUS II CDR for its diagnostic accuracy (with 95% confidence intervals CI) of clinical important intracranial injury (ICI) as identified in CT scans performed in ED.
Results: Of 20,137 total patients, we excluded 28 with suspected penetrating injury. Median age was 4.2 years. CTs were obtained in ED for 1,962 (9.8%), of whom 377 (19.2%) had a clinically important ICI as defined by NEXUS II. 74 (19.6%) patients underwent neurosurgery. Sensitivity for clinically important ICI based on the NEXUS II CDR was 373/377 (98.9%; 97.3%-99.7%) and specificity 156/1585 (9.8%; 8.4%-11.4%). Positive and negative predictive values were respectively 373/1802 (20.7%; 18.8%-22.6%) and 156/160 (97.5%; 93.7%-99.3%). Of the 18,147 children without CT 49.5% had at least one NEXUS II risk criterion.
Conclusion: NEXUS II had very high sensitivity when analysed with a focus on head injured patients who had a CT performed, similar to the derivation study. With half of unimaged patients positive for NEXUS II risk criteria the use of this rule has the potential to increase the number of CTs.

Background: Mild traumatic brain injuries (TBI) in children are a major public health issue. Yet, there is a wide variation in the way ômildö TBI is defined in literature and guidelines. To date, no study has prospectively detailed the proportion of children presenting with mild TBI to emergency departments (EDs) that are identified by the various definitions.
Objective: To apply published definitions of mild TBI to a large prospectively collected data set of head injuries and to determine the proportions of mild head injuries included by various definitions.
Methods: Prospective observational study of children with HIs of any severity at 10 Australian/New Zealand centres. We applied 18 different definitions of mild TBI, identified through a systematic review of the literature, to children aged 3 to 16 years. We assessed the number and percentage of cases the definitions applied using described inclusion and exclusion criteria
Results: Of 20,137 children with HI of any severity, 11,907 were aged 3 to 16 years. Mean age was 8.2 years, 32% were female. 61.9% were fall related. Cranial CT rate was 12.7% and neurosurgery rate was 0.5%. Adjustment were made to the definitions to apply to the data set: none in 7, minor in 9, substantial in 2. Percentages of the cohort covered by the definitions of ômildö TBI ranged from 2.4% (284) to 98.7% (11,756) of the cohort. The median percentage of head injuries defined as mild among 18 definitions investigated was 21.7% (2,589).
Conclusion: When applying different definitions of mild TBI to a single data set including all severities, a wide range of cases were included. Clinicians and researchers need to be aware of this important variability when attempting to apply the published literature to children presenting to EDs with TBI.

Background: Clinical decision rules (CDRs) can assist in determining the need for computed tomography (CT) in children with head injuries (HIs). We assessed the accuracy of 3 high quality CDRs (PECARN, CATCH and CHALICE) in a large prospective cohort of head injured children. In addition to rule accuracy, however, among a number of factors physician accuracy is also important when determining whether a particular rule should be implemented.
Objective: To assess the accuracy of physician practice in detecting clinically important traumatic brain injuries.
Methods: Prospective observational study of children <18 years with HIs of any severity at 10 mainly tertiary Australian/New Zealand centres. We extracted a cohort of children with mild HIs (GCS 13-15, presenting <24h) and assessed physician accuracy for the standardised outcome of clinically important traumatic brain injury (ciTBI) and compared this with the accuracy of PECARN, CATCH and CHALICE. Physician accuracy was defined as CT obtained during the initial ED visit.
Results: Of 20,137 children, 18,913 had a mild HI as defined with a median age of 4.1 years. Of these 1,578 (8.3% = actual CT rate) received a CT scan during the ED visit, 160 (0.8%) had ciTBI and 24 (0.1%) underwent neurosurgery. Physician practice for detecting ciTBI based on CT performed had a sensitivity of 157/160 ((98.1% (94.6% – 99.6%) and a specificity of 17,332/18,753 (92.4% (92.0% – 92.8%)). Sensitivity of PECARN <2 years was 42/42 (100.0%, 91.6% to 100.0%), PECARN >2 years 117/118 (99.2%; 95.4% to 100.0%), CATCH (high/medium risk) 147/160 (91.9%; 86.5% to 95.6%) and CHALICE 148/160 (92.5%; 87.3% to 96.1%). Projected CT rates for PECARN <2/>2 years was 8.0%/9.4% (high risk only) to 41.4%/48.5% (high and intermediate risk, considering the unlikely scenario that all patients in the intermediate risk group receive a CT scan), CATCH 30.2% (medium and high risk) and CHALICE 22.0%.
Conclusion: Physician accuracy was high. The application of PECARN, CATCH or CHALICE CDRs in this setting has the potential to increase the CT rate with limited potential to increase the accuracy of detecting ciTBI.

Background: The majority of ill or injured children requiring emergency care present outside of the tertiary setting, yet most high quality research such as randomised control trials are conducted in tertiary paediatric emergency departments (EDs). Understanding how to effectively apply research findings to the nontertiary setting is important in achieving effective equitable health care improvements for all children, regardless of where they present. Knowledge translation (KT) research aims to gain an understanding of why there are practice variations, and then design, implement and evaluate strategies to increase the uptake of research evidence.

PREDICT (Paediatric Research in Emergency Departments International Collaborative) is a peer recognised research network and has contributed substantially to the evidence base in paediatric emergency medicine (PEM) in important clinical areas such as bronchiolitis and head injury. Bronchiolitis and head injury have been chosen as topics due to their high frequency of presentation to large tertiary EDs as well as small rural or metropolitan EDs.

Evidence based clinical guidelines and clinical decision rules provide the basis for best practice for these conditions respectively and PREDICT has been at the forefront of research in these areas. Despite the existence of high quality research evidence in PEM, practice variation in clinical management of common and important health conditions is well reported.3 Bronchiolitis is the most common reason for hospital admission in infants less than 12 months and head injury is among the most common ED presentations for children across the age range, yet even these common presentations have considerable variation in care. For example, the use of medications for management of bronchiolitis has been shown to vary widely across different settings, from 27% to 49% in a study of seven different sites in Australia and New Zealand despite good evidence of lack of benefit for these medications. Further, the use of Head Injury Clinical Decision Rules, such as the Pediatric Emergency Care Applied Research Network (PECARN) and Children’s Head Injury Algorithm for the Prediction of Important Clinical Events (CHALICE) head injury rules, to determine need for cranial computer tomography (CT) scans in children is also an area of significant practice variation.

There is limited evidence for the most effective methods to translate knowledge into practice in PEM. A systemic review of the literature identified studies of varied design, including cluster-controlled trials, and interrupted time series and before and after intervention studies. Knowledge translation interventions were predominantly aimed at the treating clinician and changes in clinical practice were variable.

Similarly, studies of KT intervention strategies appropriate to the rural and remote clinical practice setting are limited. The Cochrane Effective Practice and Organisation of Care (EPOC) group undertake systematic reviews of educational and organisational interventions designed to improve health professional practice and the organisation of health care services. Specific interventions tailored to provide professional support to health professionals working in rural and remote areas has been identified as a priority topic area by the Australian satellite EPOC group but this is yet to be addressed.

The PREDICT network is investigating the effectiveness of tailored KT strategies to increase the uptake of research findings in tertiary and non-tertiary ED settings. The KT research follows a systematic, stepped approach as described in the knowledge to action framework.10 One of the first steps is to adapt new knowledge from research or evidence based guidelines to the local context and this should take into account paediatric expertise, geographic isolation and accessibility to education and equipment. Perceived barriers and enablers to evidence based care will be explored and used to develop tailored KT strategies. The effectiveness of these strategies will be evaluated with a focus on maintaining long-term sustainability.

Methods:Two research projects will be presented that potentially represent models for engagement of rural clinicians. The primary outcomes of each study are specific clinical measures at sites, reflecting the impact of intervention strategies, however the process of engaging non-tertiary clinicians in KT research is the unifying objective for both studies. The knowledge to action framework10 components will be employed with emphasis on the action cycle.

Bronchiolitis KT research: A cluster randomised control trial (cRCT) will investigate the effectiveness of tailored KT strategies for implementing a universal Australasian acute care guideline, the Australasian Bronchiolitis Guideline. This evidence based guideline for ED and general ward management of bronchiolitis was developed in 2016 using a methodological approach including Grading of Recommendations Assessment, Development and Evaluation (GRADE) and National Health and Medical Research Council (NHMRC) Evaluation of Evidence processes. The guideline development committee consisted of twenty individuals from six Australian states and territories and New Zealand representing tertiary, metropolitan and regional medical and nursing specialists from EDs, general paediatrics and respiratory medicine. The Population Intervention Comparator Outcomes and Time of Interest (PICOt) questions that formed the basis of the systematic literature search were developed by expert consensus. The final document includes a useable clinical interface for bedside use and a descriptive summary of evidence base and evidence tables for each key statement.

The cRCT study design incorporates stratification of tertiary and secondary providers of paediatric care from tertiary children’s hospitals, to metropolitan centres and regional hospitals with general paediatric wards that are referral centres for surrounding rural towns. Site inclusion will be determined by the identification of nursing and medical champions and the capacity to audit medical records. A sample size of 135 per annum ED presentations with a diagnosis of bronchiolitis was predetermined to achieve appropriate statistical power. Site recruitment will occur through established hospital networks in Australia and New Zealand. Sites will be randomised to control or intervention group, with the later group implementing the guideline with supportive educational and behavioural change strategies. Following the 2017 bronchiolitis season, audit data on the use of the five interventions for management of uncomplicated bronchiolitis for which there is high quality evidence of no clinical 3 benefit (use of chest x-ray; use of salbutamol; use of adrenaline; use of glucocorticoids; use of antibiotics) will be collected retrospectively from both intervention and control sites for 2014 to 2017.

Head injury KT research Australian Paediatric Head Injury Rules Study . PREDICT has recently completed a multicenter observational study of over 20,000 paediatric ED presentations for head injury across Australia and New Zealand. The study examines three existing clinical decision rules for the management of paediatric head injuries (PECARN, CHALICE and Canadian Assessment of Tomography for Childhood Head Injury (CATCH)) when applied to the Australasian population. The main outcome measures were head CT rates and detection of clinically significant intracranial injuries.

An audit of practice variation in cranial CT rates for acute paediatric head injury across 30 EDs will compare regional and rural sites to tertiary sites. This study will also include qualitative telephone interviews to identify information needs of doctors and nurses, to inform the content and methods to deliver KT strategies to improve appropriateness of cranial CTs in children with mild head injuries.

Results:
Bronchiolitis trial: The recruitment of sites for participation in the cRCT for bronchiolitis has been completed, with 26 sites across Australia and New Zealand enrolled, 7 tertiary and 19 secondary sites, of which four are outside major metropolitan centres. Engagement in the recruitment process was high with 10 additional sites, predominantly regional centres, expressing interest but unable to participate due to inadequate numbers of bronchiolitis presentations to their EDs. The relevance of the study to clinical practice in their own settings and the facilitated process to participate in the study, through a site coordinator visit and supportive measures such as “ Train the Trainer day” for intervention sites, were appealing to sites that usually have barriers to participation in research. The study required endorsement from both the paediatric department and ED at each site. A few sites were unable to participate due to only one of the departments being engaged.

Parallel to site recruitment, 20 semi-structured interviews were undertaken by investigators with clinicians to explore factors perceived to influence the uptake of five key evidence-based recommendations from the Australasian Bronchiolitis Guideline. Both nurses and medical officers (senior and junior) were interviewed from four hospitals, one regional and one tertiary metropolitan hospital in both Australia and New Zealand, with staff from the paediatric department and ED being interviewed. Sixteen interviews were face-to-face, with four completed by phone. The resulting interviews were transcribed and analysed using NVivo11 (QSR software) into a theoretical domains framework model. An expert panel then developed suggested interventions based on mapping of behaviour change techniques.

The 26 sites have been randomised 1:1 into intervention or control sites. Intervention sites will complete an audit of the management of 40 bronchiolitis patients from 2016. The results of the audit will identify areas for improvement, gain buy-in from sites and enable the process of tailoring interventions to site-specific requirements. Each intervention site will choose four clinical champions to attend a “train-the-trainer” day in February 2017. These champions will be responsible for delivering the study interventions to their staff prior to the 2017 bronchiolitis season.

Cranial Computed Tomography for head injury audit: The project is in the recruitment phase with 31 sites confirmed across Australia and New Zealand (9 tertiary/11 major urban/11 rural/regional). A national ethics application is in progress. All sites will audit 100 head injury presentations. Audit and analysis is planned for completion prior to end 2017. Clinicians will be invited to participate in qualitative interviews after site recruitment has been finalised.

Conclusion: An understanding of the most effective KT strategies to implement research findings in paediatric emergency care within regional and rural health settings is fundamental to improving the care of all children presenting with acute illness or injury. Design of translational research projects to include rural and regional centres is integral in achieving this.

Background: Paediatric cervical spine injury (CSI) is rare but can have devastating consequences. Many more children are assessed in the Emergency Department(ED) for possible CSI. While some of these children receive imaging, others are “clinically cleared” without any radiological investigation being undertaken.Very little information exists in the literature on this complete group, including in an Australian context. This paper examines the clinical characteristics of the first 300 eligible children of the larger prospective observational CRIC study.
Objective: To examine the clinical characteristics of children presenting with possible CSI to a single Queensland centre.
Methods: All children under 16 years presenting to ED with possible CSI were eligible for inclusion. At least one of the following was present i.Immobilised for possible CSI  ii.neck pain in the context of trauma or iii.otherwise considered at risk  by the ED team.
Results:  278 children were included for analysis. Recruitment occurred from mid September 2015-mid January 2016.  Median age was 9.8 years; over 25% were aged under 6.  63.4% were male.  61.9% were assessed after falls; 21.2% after motor vehicle related accidents. 32% were related to specific sports. On initial examination, 89% were GCS 15, 2.2% arrived intubated, 40% had posterior neck tenderness, 3.9% had focal neurology and 6.5%, torticollis.
Conclusion: This cohort includes a significant proportion of younger children, identifying that while this group is poorly described in the literature, they are regularly assessed in practice for possible CSI. Further research is required to determine if observed patterns persist outside of a spring/summer cohort.

Background: Paediatric cervical spine injury (CSI) is rare, but can have devastating consequences. Many more children are assessed for possible CSI.  Some of these children are “clinically cleared” in the Emergency Departmen(ED) with no imaging undertaken, others receive imaging to exclude radiologically apparent injury. Little information exists on the former group, and imaging patterns among children assessed for possible CSI are poorly understood. This paper examines the first 300 eligible children of the larger prospective observational CRIC study.
Objective: To determine the rates of imaging and “clinical clearance” among children presenting with possible CSI to a single Queensland centre.
Methods: All children aged under 16 years presenting to ED with possible CSI were eligible for inclusion.  At least one of the following was present i.Immobilised for possible CSI  ii.neck pain in the context of trauma or iii.otherwise considered at risk  by the ED team.
Results: 278 children were included for analysis. Median age was 9.8 years. 45% received imaging for possible CSI in the ED- 34.1% had plain films, 15.8% CT scans, and 4.3% MRI scans. Plain film imaging rates increased with age, whereas CT rates remained relatively constant (15-17%).  92.8% of children had their cervical spine “cleared” in ED.
Conclusion: A significant number of children are clinically cleared (i.e. without imaging)of cervical spine injury, highlighting the importance of further research into paediatric appropriate clinical decision rules in the assessment and imaging of children with possible CSI. More extensive exploration into patterns of imaging, including multi-centre, would also be beneficial.

Background: Bronchiolitis represents the most common cause for non-elective hospital admission in infants. Nasal High Flow (NHF) therapy has achieved high uptake of use in infants with bronchiolitis despite limited high-­‐quality evidence. The efficacy and safety of NHF therapy in infants with bronchiolitis outside intensive care remains unknown.

Methods: Open-labelled randomized controlled trial using delayed consent, occurring in 17 emergency departments and general paediatric wards of tertiary paediatric and secondary hospitals in Australia and New Zealand, comparing NHF therapy (2L/kg/min) vs. standard oxygen therapy (SOT) via nasal cannula (0-2L/min) in infants ,<12 months admitted with bronchiolitis and hypoxia (SpO₂<92%/94%, threshold dependent on hospital guideline).  Primary outcome was treatment failure during hospital admission requiring escalation of respiratory support and/or intensive care admission.  Escalation of therapy occurred if ≥3 out of 4 criteria were met: persistent tachycardia, tachypnea, hypoxemia, and/or hospital early warning tool activated.  Secondary outcomes were length of oxygen therapy (LoO₂T) and serious adverse events.

Results: 1,476 patients were randomized over 3 years (Figure 1).  Baseline characteristics were similar between groups (Table 1). Mean (SD) age in SOT was 6.1 (3.4) months and NHF therapy 5.8 (3.5) months.  Escalation of care was required in 89/745 (12%) of infants on NHF therapy vs. 167/731 (23%) of infants on SOT (risk difference 10.9%, 95% CI 7.1-14.7, p<0.001) (Table 2).  Median LoO₂T for NHF therapy was 1.24 (IQR 1.81) days, and SOT was 1.23 (IQR 1.82) (p=0.218).  Other than one (0.1%) pneumothorax in each study arm there were no serious adverse events.

Conclusions: In infants with bronchiolitis with an oxygen requirement NHF therapy had a signficantly lower treatment failure rate than SOT with a number to treat is 9.  NHF therapy appears safe in a large data set when delivered to infants with bronchiolitis in the emergency department and general paediatric ward.

Nasal High Flow (NHF) has rapidly become the new respiratory support modality in Neonatal care and in the Adult Intensive Care.  High grade evidence obtained from randomized controlled trials have shown that NHF has its clinical validity similar to bubble CPAP.  Adult Trials are showing a similar but yet not fully convincing trend.  The uptake of NHF in paediatrics, particularly in bronchiolitis has been high despite any high grade evidence.  Good quality physiological studies have shown that the work of breathing can be reduced using NHF in bronchiolitis and large case/cohort studies describing substantial change in ventilator practice once NHF has been introduced.  The use of NHF outside the paediatric intensive care unit, in general ward settings is controversial.  Safety and quality data are indicating that NHF can be applied in less well monitored settings, particularly in view to reduce intensive care admissions.  Worldwide PICU admissions due to bronchiolitis are increasing, and the demand on health care cost reduction is increasing.  Our own data obtained from a large RCT investigating the efficacy of NHF compared to standard care in general paediatric wards has shown a high reliability of the use of NHF with a good safety and quality profile.

Aims: To describe the prevalence of clinically significant abdominal pathology (CSAP), patterns of imaging and pathology tests, management and outcomes of children presenting with abdominal pain to multiple Paediatric Research in Emergency Departments International Collaborative (PREDICT) EDs in Australia.
Methods: A prospective multi-centre observational cohort study of children presenting to four PREDICT EDs with abdominal pain was performed. Children aged 0-16 presenting with ‘abdominal pain’ recorded at triage were recruited for a one-month period at each site. The primary outcome was to describe the prevalence of CSAP, defined as any abdominal surgery, any abdominal interventional procedure, or admission to hospital for over 48 h for investigation of abdominal pain. Age, gender, length of stay, type of surgery or interventional procedure, type of imaging and pathology tests performed and analgesia use were recorded.
Results: We studied 555 cases with a median age of 9 years (IQR 6-12). Eighty-two patients (14.8%) had (Table Presented) CSAP, of which 41 (7.4%) had appendicitis (Table 1). Out of 555 cases, 348 (62.7%) were discharged directly from ED and 207 (37.3%) were admitted to hospital. Two hundred and fifty-five patients (45.9%) had pathology tests and 55 (9.9%) had imaging tests during their ED attendance (Table 2).
Conclusion: The prevalence of CSAP and appendicitis in our study were 14.8 and 7.4% respectively. Fewer than half of patients received blood tests and 9.9% received imaging tests during their ED attendance.

Aims: To assess the diagnostic accuracy based on clinical gestalt of ED clinicians according to clinician seniority in diagnosing paediatric acute appendicitis across multiple Paediatric Research in Emergency Departments International Collaborative
(PREDICT) ED in Australia.
Methods: A prospective multi-centre observational cohort study of children presenting to four PREDICT EDs with abdominal pain was performed. Children aged 0-16 presenting with “abdominal pain” recorded at triage were recruited over a one-month period at each site. Clinicians completed a questionnaire stating their seniority and estimated likelihood of their patient having acute appendicitis, which was compared with the patient’s final diagnosis determined by histopathology of a surgical specimen, or hospital discharge summary supplemented by telephone follow-up for those not undergoing an operation. The primary outcome was presence or absence of appendicitis.
Results: 555 patients presented during the study period, with 381 having clinician questionnaires completed (139 junior, 196 intermediate, 46 senior clinicians). The Area Under Curve (AUC) of all ED clinician’s diagnosis of acute appendicitis was 0.839, and AUC stratified to clinician seniority (junior, intermediate and senior) were 0.911, 0.797 and 0.793 respectively. The sensitivity and specificity of clinician gestalt were 81% and 77% respectively, with a positive predictive value of 24% and negative predictive value of 98%.
Conclusion: Clinical gestalt in all levels of seniority of ED clinicians in diagnosing acute appendicitis is comparable to validated clinical scoring systems with a decreasing AUC associated with increasing clinician seniority. (Table Presented).