2019

Aims: There are currently no standardised guidelines for treating cellulitis in children, specifically whether oral or intravenous (IV) antibiotics should be used. This potentially results in over treatment leading to unnecessary hospital admission or under treatment resulting in further complications. This study aimed to derive and validate a cellulitis risk assessment scoring system to guide which patients require IV antibiotics.

Methods: A prospective cohort study of children presenting to the Emergency Department aged 6 months-18 years diagnosed with cellulitis, from January 2014-August 2017 for the derivation cohort and over a 14-month period from June 2015-August 2017 for the validation cohort. Demographics and clinical features of cellulitis were collected, and patients were divided into two groups based on route of antibiotic at 24 hours (the pre-determined gold standard). A comparison of demographics and clinical features were made between the two groups.  Clinicians were surveyed on which features they used to decide whether to start IV antibiotics. Combinations of differentiating features were plotted on receiver operating characteristic curves.

Results: There were 285 children in the derivation cohort to create the Melbourne ASSET (Area, Systemic features, Swelling, Eye, Tenderness) score out of 7. The area under the curve (AUC) was 0.86 [95% confidence interval (CI) 0.83-0.91]. Using a cutoff score of 4 to start IV antibiotics gave the highest correct classification of 80% of patients (sensitivity 60%, specificity of 93%). This score was validated in 251 children and maintained a robust AUC of 0.83 (95% CI 0.78-0.89), correctly classifying 76% of patients (sensitivity 85%, specificity 63%).

Conclusion: The Melbourne ASSET score is the first risk assessment scoring system for childhood cellulitis proposed to aid clinicians in deciding whether to treat with IV or oral antibiotics. It is simple, easy to use, applicable and reliable.

Aims: Management of children with minor blunt head trauma often includes a period of observation to determine the need for cranial computed tomography (CT). Objective:
We explored the relationship between planned observation and cranial CT use among children <18 years old with minor blunt head trauma.

Methods: This was a planned sub-analysis of a prospective observational study at 10 pediatric emergency departments (EDs) in Australia and New Zealand. Children <18 years old with Glasgow Coma Scale (GCS) scores ≥14 presenting within 24 hours of blunt head trauma were included. After the initial assessment, clinicians documented responses to two questions that were not mutually exclusive: if they planned to observe the patient, and if the planned to obtain neuroimaging. Based on the responses, the planned observation group included those with planned observation and no immediate plan for cranial CT. The no planned observation group included those with either planned immediate cranial CT or no planned observation. For patient outcomes we used the PECARN definition of a clinically important traumatic brain injury (ciTBI): either intubation for more than 24 hours, ≥2 day hospitalization, requiring neurosurgical intervention or death following head trauma. We estimated the association of cranial CT use with planned observation using a generalized linear model with mixed effects, adjusting for time from injury, patient characteristics, PECARN TBI risk group and hospital clustering.

Results: Of 20,137 children enrolled, 18,471 (92%) presented within 24 hours of sustaining blunt head trauma, had GCS scores ≥14 and documented plans for observation and neuroimaging. The cranial CT rate in the total cohort was 8.6%, and 0.8% had ciTBI. The planned observation group had 4,945 (27%) children compared to 13,526 (73%) in the no planned observation group. The cranial CT rate was lower in those with planned observation (4.2%) than those in the no planned observation group (10.1%), (rate difference 5.7%, [95% CI: 5.0 – 6.5%]). Similarly, the ciTBI rate in the planned observation group was lower (0.4% vs. 0.9%; rate difference 0.5%, [95%CI: 0.3-0.8%]). There was no difference in missed ciTBI rates between the groups with 1 patient (0.1% of patients with ED return visit) in each group. After adjusting for PECARN TBI risk groups, time from injury, patient characteristics, and hospital cluster effects, patients with planned observation had significantly lower cranial CT rates (adjusted odds ratio 0.2, [95% CI: 0.15-0.21]).

Conclusion: Even in a setting with low overall cranial CT rates in children with minor blunt head trauma, planned observation was associated with decreased cranial CT use.

Aims: Infants and young children frequently experience painful procedures in the emergency department, the management of which is contingent on accurate assessment. The most suitable tool for this purpose is unclear. The aim of this study was to compare the reliability, validity, feasibility and utility of three observational pain scales to quantify procedural pain in infants and young children.

Methods: A convenience sample of twenty-six clinicians used the Face, Legs, Activity, Cry & Consolability (FLACC) scale, the Modified Behavioural Pain Scale (MBPS) and the Visual Analogue Scale applied by an observer (VASobs) to segments of video from 100 children aged six to 42 months undergoing a procedure.

Results: Inter-rater reliability was poorest for VASobs pain scores (ICC – 0.55). VASobs pain scores were lower than FLACC and MBS scores during the procedure but MBPS scores were higher during non-painful phases (p < 0.001). The FLACC scale provided the best sensitivity (94.9%) and specificity (72.5%) for the lowest cut-off score (pain score 2). Correlations between FLACC and VASobs (distress) were strongest (r = 0.89). The FLACC scale resulted in more incomplete scores (p < 0.000) than the other scales. Finally, reviewers liked the VASobs (pain) most, considered it quickest and easiest to apply but all scales were considered of limited use for procedural pain assessment.

Conclusion: This study supported the reliability and sensitivity of the FLACC and MBPS. There were practical concerns for application of the FLACC scale and the MBPS and doubt about the capacity of both scales to differentiate between pain- and non-pain related distress exist. The VASobs, although practical, was less reliable than either the FLACC scale or the MBPS. The results of this study demonstrated that the FLACC scale may be best suited for procedural pain assessment.

Aims: Computed tomography of the brain (CTB) for paediatric head injury is used at low rates at tertiary paediatric Emergency Departments (EDs) in Australia and New Zealand. However, most paediatric patients are seen in mixed, non-tertiary EDs. International studies have found large variation in CTB rates across hospital types. We aimed to assess variation in CTB use for paediatric head injury in Australia and New Zealand across tertiary, urban/suburban and regional/rural EDs.

Methods: A retrospective observational study of medical and neuroimaging records of presentations to 30 tertiary, urban/suburban and regional/rural EDs in Australia and New Zealand in 2016.  Case inclusion criteria; 1) Primary ED diagnosis of head injury; 2)Age <16 years. Data extraction was undertaken on 100 sequential eligible cases per site for head injury severity by Glasgow Coma Scale (GCS) scores, CTB rate and clinical management.

Results: 2472 eligible presentations from 25 of 32 EDs recruited have been analysed, 9 tertiary (n=900), 9 urban/suburban (n=872) and 7 regional/rural EDs (n=700). Proportion of children <2 years; tertiary 54.6%; urban/suburban 42.9%; regional/rural 36%, p<.001. Proportion of children presenting with a known GCS of 15/14/13/12-9/3-8; tertiary 95.2%/2.4%/0.4%/0.6%/0.22%, urban/suburban 93.7%/4.6%/0.57%/0.8%/0.1%, regional/rural 93.6%/2.57%/0.3%/0.71%/0.1%, p=0.001. CTB imaging rates were; tertiary 8.2%, urban/suburban 7.2%, regional/rural 4.14%, p=0.004. Median length of stay was; tertiary 2.6 hours, urban/suburban 2.7 hours, regional/rural 1.7 hours, p=<0.001. Full data are expected to be analysed by January 2019, when severity weighted analysis will occur.

Conclusion: Neuroimaging rates for paediatric head injury in Australia and New Zealand are not higher in mixed urban/suburban or regional/rural EDs when compared with tertiary paediatric EDs.  Lower imaging rates do not seem to be offset by increased length of observation.  Assessing variation in CTB rates across hospital types is an important consideration in strategies to improve care.

Aims: Bronchiolitis is the most common reason for infants to be hospitalised following presentation to emergency departments (EDs). Management is supportive with high-level evidence of no efficacy for salbutamol, glucocorticoids, chest x-rays, antibiotics, or adrenaline. Despite all international guidelines recommending against the use of these therapies, significant practice variation exists, with the use of inappropriate therapy a worldwide problem. Knowledge translation (KT) interventions that are tailored to the factors that influence practice can improve care; however there is little high-level evidence in acute paediatrics. The primary objective was to establish whether tailored, theory informed KT interventions, compared to passive guideline dissemination, reduces the use of salbutamol, glucocorticoids, chest x-rays, antibiotics, and adrenaline, in infants. 

Methods: Tailored KT interventions were developed, following qualitative interviews using the Theoretical Domains Framework, to target key identified factors influencing bronchiolitis management. We then compared the tailored KT interventions versus passive dissemination in a cluster randomised controlled trial of 26 hospitals in Australia and New Zealand during the 2017 Australasian bronchiolitis season. The primary outcome was compliance with the Australasian Bronchiolitis Guideline during the first 24 hours of care (no use of salbutamol, glucocorticoids, chest x-rays, antibiotics, or adrenaline). Secondary outcomes included length of stay, median medication doses, health costs and process evaluation.

Results: Data was collected on 3,727 infants. Compliance with the Australasian Bronchiolitis Guideline was 86.1% (95%CI 82.6-89.7%) in the intervention sites and 72.0% (95%CI 65.3-78.8%) in the control sites, risk difference 14.1% (95%CI 6.5%-21.7%), p<0.0001.

Conclusion: The use of tailored KT interventions substantially reduces the use of inappropriate therapies in the management of infants with bronchiolitis. As bronchiolitis is the most common reason for infants to be admitted to hospital this study has important implications for future management of bronchiolitis and for KT in general in EDs.

Aims: Nasal high-flow oxygen therapy is increasingly used for respiratory failure in infants with bronchiolitis.  It is unclear if children receiving high flow can be safely fed enterally.

Methods: We performed a preplanned secondary analysis of a multi-center, randomised controlled trial of 1,472 infants <aged ,12 months with bronchiolitis and an oxygen requirement.  Children were assigned to treatment with either high-flow or standard-oxygen therapy with optional rescue high-flow.  For a subgroup we assessed how children on high-flow therapy were hydrated and fed; intravenously (IV), via bolus or continuous naso-gastric tube (NGT) or orally.  Secondary outcomes included adverse events while receiving enteral feeding.

Results: 505 patients (mean age 5.8 months, female 186 [36.8%]) on high-flow via primary study assignment (n=408) or as rescue therapy (n=97) were included.  While on high-flow, 15 of 505 (3%) received only IV fluids, 360 (71.3%) received only enteral feeds, and 93 (18.4%) received both IV and enteral feeds.  453 infants were enterally fed at some stage during their treatment on high-flow.  Of these, 80 (15.8%) received NGT bolus, 217 (43%) NGT continuous, 118 (23.4%) both bolus and continuous, 32 (6.3%) received only oral feeds and 171 (33.9%) a mix of NGT and oral feeds.  None of the patients receiving oral or NGT feeding on high-flow sustained pulmonary aspiration (0%; 95% CI 0% to 0.8%); one patient had a pneumothorax (0.2%; 95% CI 0% to 1.2%).

Conclusion: In children with bronchiolitis treated with HF, enteral feeding using either oral or NGT route was safe.

Aims: Using intravenous (IV) antibiotics to treat urinary tract infection(UTI)/pyelonephritis places a significant burden on patients and health care resources. Despite a recent Cochrane review, guidelines do not clearly state the criteria for IV antibiotics in children with UTI/pyelonephritis. We aimed to derive and validate a clinical score that incorporates clinical features and patient complexity to guide the decision on the route of antibiotics.

Methods: This was an observational study (May’16-March’18) of all children (3m-18y) diagnosed in ED with UTI/pyelonephritis and subsequently confirmed on urine culture. To derive the score, half the cohort was used (aged 12m-12y) who met criteria from a recent systematic review. Patients were defined those with and without a ‘true’ need for IV antibiotics using the pre-determined gold standard based on the Cochrane review and route of antibiotic used at 24 hours. Features which were significantly different between the groups were identified. Combinations of these differentiating features were used to generate receiver operating characteristics(ROC) curves. The score was validated on the second half of the cohort and additional cohorts.

Results: 1,240 patients had a confirmed UTI/pyelonephritis: 831 (67%) aged 12m-12y, 276 (22%) aged 3-12m and 133(11%) aged 12-18y. Of those aged 12m-12y, 335 (40%) met the definition for inclusion, 167 were used to derive the RUPERT score (Rigors, Urological abnormality, Pyrexia(≥38◦C), Emesis, Tachycardia, Recurrent(>2) UTI – one point each – maximum 6), area under curve(AUC) of 0.85. A score of ≥3 to commence IV antibiotics resulted in correct classification of 80% of patients(sensitivity 77%, specificity 81%). In the validation cohort the AUC was 0.8, those not meeting full systematic review criteria AUC 0.82, and those age 12-18y AUC 0.86. It was not reliable in those age 3-12m AUC 0.58.

Conclusion: The RUPERT score can aid clinicians in deciding an appropriate route of antibiotics for UTI/pyelonephritis in children aged 12m-18y.

Aims: Computed tomography of the brain (CTB) for paediatric head injury is used at low rates at tertiary paediatric Emergency Departments (EDs) in Australia and New Zealand. However, most paediatric patients are seen in mixed, non-tertiary EDs. International studies have found large variation in CTB rates across hospital types. We aimed to assess variation in CTB use for paediatric head injury in Australia and New Zealand across tertiary, urban/suburban and regional/rural EDs.

Methods: A retrospective observational study of medical and neuroimaging records of presentations to 30 tertiary, urban/suburban and regional/rural EDs in Australia and New Zealand in 2016.  Case inclusion criteria; 1) Primary ED diagnosis of head injury; 2)Age <16 years. Data extraction was undertaken on 100 sequential eligible cases per site for head injury severity by Glasgow Coma Scale (GCS) scores, CTB rate and clinical management.

Results: 2472 eligible presentations from 25 of 32 EDs recruited have been analysed, 9 tertiary (n=900), 9 urban/suburban (n=872) and 7 regional/rural EDs (n=700). Proportion of children <2 years; tertiary 54.6%; urban/suburban 42.9%; regional/rural 36%, p<.001. Proportion of children presenting with a known GCS of 15/14/13/12-9/3-8; tertiary 95.2%/2.4%/0.4%/0.6%/0.22%, urban/suburban 93.7%/4.6%/0.57%/0.8%/0.1%, regional/rural 93.6%/2.57%/0.3%/0.71%/0.1%, p=0.001. CTB imaging rates were; tertiary 8.2%, urban/suburban 7.2%, regional/rural 4.14%, p=0.004. Median length of stay was; tertiary 2.6 hours, urban/suburban 2.7 hours, regional/rural 1.7 hours, p=<0.001. Full data are expected to be analysed by January 2019, when severity weighted analysis will occur.

Conclusion: Neuroimaging rates for paediatric head injury in Australia and New Zealand are not higher in mixed urban/suburban or regional/rural EDs when compared with tertiary paediatric EDs.  Lower imaging rates do not seem to be offset by increased length of observation.  Assessing variation in CTB rates across hospital types is an important consideration in strategies to improve care.

Aims: We set out to to determine how many possible injuries are clinically cleared, what imaging is used on the remainder and the incidence and characteristics of confirmed paediatric cervical spine injuries.

Methods: We undertook a retrospective electronic medical record review of children <18 years with potential cervical spine injuries at a large tertiary paediatric trauma centre in Victoria, Australia over a 12-month period (annual census 87 000). For possible injuries we extracted key epidemiologic, imaging and short-term outcome data.

Results: During the study period, a total of 617 patients with potential neck injuries were seen in the ED(617/87000 0. 7%). The median age was 11 years. The most common mechanisms of injury were falls (41%), motor vehicle injuries (28%) and sports-related injuries(24%). 414/617 (67%) underwent neck imaging(345/414 83% plain radiograph, 100/414 24% computed tomography, 7/414 1.6% magnetic resonance imaging). 23/617 (4.1%) had radiologically documented cervical spine injuries. Two required operative interventions for their neck injuries.

Conclusion: While two-thirds of children with potential cervical spine injuries undergo radiological evaluation,actual injuries are rare (<4%). These data suggest that there is a potential for improved targeting of cervical spine imaging for trauma. The development of a clinical decision tool may help reduce neck radiography.

Aims: To determine clinical characteristics and management of SE in children in Australian EDs

Methods: Prospective, observational cohort study at 4 mixed and tertiary paediatric ED in Australia. Participants were included if presenting with SE (seizure duration >5 min). We present descriptive data on clinical characteristics and management as medians (IQR) using APLS dosing as gold standard.

Results:172 children were enrolled: age 4.0 years (1.8 to8.5), weight 18.0 kg (12.3 to 26.8) and 50.0% were male. Seizure duration was 15.0 min (7.0 to 40.0). The majority of patients arrived by ambulance (73.8%) and had a history of prior seizures/epilepsy (67.4%). Benzodiazepines were administered by parents/caregivers45/172 (26.2%), prehospital 58/172 (33.7%) and in ED,43/172 (25.0%) and were administered after seizure onset at 12.5 min (7–18.1), 13.1 min (9.3–19.3) and 125 min (70–250) respectively. Median doses of benzodiazepines were 0.2 mg (0.1–0.3)and 0.2 (0.1–0.2). Second line drugs (phenytoin or leviteracetam) was administered in ED 21/172 (12.2%) of patients at 125.0 min (70.0–250.0). On discharge, the most common diagnosis was convulsive SE and focal motor SE with prominent motor signs. Disposition included 6.9% ICU and 93.1% ward.

Conclusion: Children with SE are generally treated as per APLS guidelines. However, there is a large scope for earlier treatment with both benzodiazepines and second line drugs.