G60 MEDICAL PAEDIATRICS IN THE EMERGENCY DEPARTMENT: TRENDS IN ATTENDANCE RATES, SOURCES OF REFERRAL, AND ADMISSIONS, 1993–2005

N. P. France, J. L. Craze.Oxford Radcliffe Hospitals Trust, Oxford, UK

Introduction and Aims: Although there have been a number of studies exploring reasons why children with minor illnesses attend emergency departments, there have been no published studies looking at changes in the incidence and nature of these attendances with time. This is of particular interest following recent changes to primary care out-of-hours provision. We examined trends in the attendance rates, referral patterns, and outcomes for children with medical illnesses attending our emergency department (ED) over a 12 year period.

Methods: We used the computerised booking-in system in the emergency department to collected anonymised data for all children attending with a medical complaint for the month of January 2005. Results were compared with those of previous studies, using the same methodology, for January 1993, 2000, 2003, and 2004. We examined changes in total number of attendances, the source of referral (GP referral v self-presentation) and outcome (admission v discharge).

Results: We identified a year-on-year increase in the total number of children with medical problems attending the emergency department over the period studied. The total attendances in January 2005 were more than double those in January 2000. The increase in attendances was largely due to an increase in self-presenters, the number of which rose steadily from 1993 to 2004 (trebling within that time period) and steeply—by approximately 75%—between 2004 and 2005. The number of GP referrals remained unchanged with no identified trend. The ratio of GP referrals: self-presenters was 3:1 in 1993 but by January 2005 it had reversed to 1:3. There was a downward trend in the proportion of self-referrals being admitted to the wards over the time period 1993 to 2004. In 2005 there was a small increase in admissions for this group. There was no clear trend in the proportion of admissions for GP referrals.

Conclusion: These results might suggest an emerging crisis for paediatric emergency departments throughout the country. The steep rise in self-presenters between 2004 and 2005 followed the restructuring of out-of-hours primary care arrangements, but occurs against a background of a steady longer term increase. Possible explanations for this change in population behaviour include: perceived barriers to accessing primary care, increased anxiety regarding minor illness, and increased public expectations.

G61 THE IMPACT OF THE NATIONAL SERVICE FRAMEWORK FOR CHILDREN, YOUNG PEOPLE AND MATERNITY SERVICES (CHILDREN’s NSF) ON EMERGENCY DEPARTMENTS

T. Patel, I. Maconochie.St Mary’s Hospital, London, UK

Aims: The effects of managing 16–18 year olds as children, as per the Children’s National Framework, on our emergency departments.

Methods: Using our computerised system, Symphony, data on patients aged 0–16 attending the paediatric emergency service (Group 1), and aged 16–18 attending the adult emergency service (Group 2) were obtained over a one year period.

Results: Our results showed significant population differences between the two groups and therefore very different needs. The table below illustrates some of these differences:

Conclusion: Increasing the number of patients attending an already stretched service with a population of patients with very different needs will have significant implications for delivering unscheduled care.

G62 THE INCIDENCE OF SERIOUS BACTERIAL INFECTION IN FEBRILE CHILDREN, BELOW THE AGE OF 5, WHO ARE DISCHARGED THEN RETURN TO AN EMERGENCY DEPARTMENT DURING THE SAME FEBRILE ILLNESS

G. Haythornthwaite, M. McCaskill, G. Williams, M. Jones, J. Craig.Children’s Hospital at Westwead, Sydney, Australia

Aim: To analysis if returning to an emergency department (ED) during the same febrile illness is a predictor of serious bacterial infection (SBI) in children below the age of 5.

Method: Data were collected on all children below the age of 5 years, who presented to a tertiary paediatric hospital ED, with fever or a history of fever in the last 24 hours over a one year period July 2004 to June 2005. Children were excluded with known chronic disease, immunosuppression, oncology, those transferred from other hospitals, and those with prior reference tests. Data were collected by mandatory electronic assessment tools at triage and doctor’s consultation. Outcome information was then entered onto the database. SBI was defined by standardised reference tests within the six main bacterial diagnoses; pneumonia, urinary tract infection, bacteraemia, meningitis, osteomyelitis, and septic arthritis. A return visit within a febrile illness was defined as a visit to the ED within 24 hours of another febrile visit or when fever had persisted between the ED visits. Fever was defined as persistent if there was no gap in fever longer than 24 hours. The rate SBI was compared between the returning population and the main population over the same period. The returning population with serious bacterial infection was then subdivided according to nature of the SBI discovery based on review of the electronic patient notes.

Results: The rate of SBI to ED excluding recalls was 69/468 (14.7 (11.7 to 17.9)). The rate of SBI diagnosed after return excluding recalls and patients who were diagnosed on initial visit was 49/468 (10.4 (7.7 to 13.4))

Conclusion: Return to the ED during a febrile illness in itself is a predictor of SBI. Guidelines should highlight returning as a concerning feature in the history. The child who returns should be assessed carefully and have more senior staff involved.

G63 MANAGING HYPOGLYCAEMIA IN THE NON-DIABETIC CHILD: AN EVIDENCE BASED APPROACH

R. Bowker¹, T. Stephenson¹, M. Atkinson², D. Bond³, J. Bonham⁴, S. Hewitt⁵, M. Lakhanpaul⁶, S. Shipston⁷, S. Smith², H. Vyas², W. Whitehouse¹.¹Academic Division of Child Health, School of Human Development, University of Nottingham, Nottingham, UK; ²Department of Child Health, Queen’s Medical Centre, Nottingham, UK; ³Department of Child Health, Kings Mill Hospital, Sutton-in-Ashfield, UK; ⁴Department of Clinical Chemistry, Sheffield Children’s Hospital, Sheffield, UK; ⁵Department of Accident and Emergency Medicine, Derby Royal Infirmary, Derby, UK; ⁶Academic Division of Child Health, Department of Medical Education and Social Sciences, Robert Kilpatrick Clinical Sciences Building, Leicester Royal Infirmary, Leicester, UK; ⁷Department of Accident and Emergency Medicine, Queen’s Medical Centre, Nottingham, UK

Aims: When an unconscious child presents to hospital, hypoglycaemia needs to be recognised and managed effectively. There has previously been no national agreement on which investigations to request in this situation and how to appropriately treat these children. This study aimed to determine the appropriate tests to send when a child presents with hypoglycaemia and the initial management steps to take.

Method: As part of the development of an evidence based guideline, a systematic review of the literature was performed. When no evidence was available a formal Delphi consensus process was performed, using a multidisciplinary panel of 39 experts, to agree on best practice for junior doctors. Responses were made on a nine-point Likert scale. The statistical definition of consensus was 75% of responders marking a recommendation as 7, 8, or 9 on the scale. Three rounds could be performed to reach agreement.

Results: Published evidence draws conflicting conclusions about the level of hypoglycaemia which causes poor neurological outcome in children. The clinically important outcomes are more dependent on the cause of the low blood sugar than the level of it. There is a lack of published evidence to help determine which tests to request and which is the best treatment to correct hypoglycaemia in children. The Delphi panel agreed with the following recommendations: Hypoglycaemia needs actively investigating and treating if the capillary glucose is ⩽2.6 mmol/l. If the capillary glucose is between 2.6 and 3.5 mmol/l assess the laboratory glucose result. Before treatment ensure the following tests are sent: laboratory glucose, U+Es, liver function tests, plasma ammonia, full blood count, blood gas, blood culture, urinalysis, and save a sample of urine, plasma, and serum. From the saved plasma sample request a lactate, insulin, cortisol, growth hormone, free fatty acids, beta-hydroxybutyrate, acyl-carnitine profile. From the urine sample request an organic acid profile. Treatment is recommended as a bolus of 10% dextrose (2 ml/kg for neonates and 5 ml/kg for older infants and children) followed by an infusion to keep the capillary glucose between 4 and 7 mmol/l.

Discussion: Implementing these recommendations nationally will hopefully standardise and improve the diagnosis and treatment of non-diabetic children with hypoglycaemia.

G64 THE SPECTRUM OF ACUTE ALLERGIC AND ANAPHYLACTIC REACTIONS IN CHILDREN PRESENTING TO A DISTRICT GENERAL HOSPITAL PAEDIATRIC A&E UNIT

M. Absoud, S. Swain, N. Sellathurai, V. Chakravarti.Princess Alexandra Hospital, Harlow, Essex CM20 1QX, UK

Aim: There are limited data on emergency presentation of paediatric acute allergic reactions in the UK. We studied all acute allergic presentations to one DGH paediatric emergency department, pre and post publication of national resuscitation guidelines, to determine epidemiological, clinical, and outcome data.

Methods: Retrospective, case based study of patients aged 1 month to 16 years, attending an emergency department in a DGH. Medical records of patients satisfying the relevant ICD-10 criteria were scrutinised from 2004 (Study A) and compared with 1997 (Study B). We determined incidence, sex ratio, cause, referral source, clinical features, management, and patient disposal. Acute allergic reactions were graded as mild, moderate, and severe.

Results: 122 records were analysed and a total of 82 children were included, 52 from study A and 30 from study B, showing an increase in incidence of 75%. Data from study B is represented in parenthesis. Mean age was 6.2 years (6.7 years). 6/52 patients had anaphylaxis (2/30), giving a current presentation incidence of 1:200 casualty presentations for acute allergic reactions and 1:1500 for anaphylaxis. A cause was recognised in 48% of all cases (80%) most frequently foods, with nuts causing 23% of reactions (20%). Presenting cutaneous features were recorded in 88% (93%), and a past history of asthma in 20% (27%). Family history of atopy was poorly recorded. Of the children with anaphylaxis, one out of six presented with collapse due to egg (1/2, latex). The rest presented with respiratory distress; three from nuts, one egg, one dog (1/2, grass). In study A, three out of six with anaphylaxis received adrenaline prehospital (none in Study B). Half the patients received iv hydrocortisone and chlorpheniramine initially. In hospital one patient received neb adrenaline (iv, one patient). In all reactions 11 were admitted, and 6 referred to our allergy clinic from casualty. 10 children were “overtreated”. No investigations were recommended and few received advice regarding future management.

Conclusion: There is an increasing presentation of acute allergic reactions and anaphylaxis to our DGH. All allergic reactions may be managed in casualty; however, the importance of specialist follow up must be recognised. Despite publication of guidelines, an improved awareness of acute allergic reactions, and their management in the emergency department is required.

G65 ACCIDENTAL DISCHARGES OF EPIPEN DEVICES: A CASE SERIES

P. Davies, P. Shah, L. Goldsworthy.Emergency Department, Bristol Children’s Hospital, Bristol, UK

Aim: To investigate the incidence and effects of cases of accidental Epipen injections in patients where use was not indicated at the time.

Method: A retrospective review of admissions over six years with the search terms “Epipen”, “Epi Pen”, “Adrenaline”, or “Injection” in the presenting complaint as given to the receptionist. The setting was a dedicated Paediatric Emergency Department with a catchment population of 60 000 children. Cases were reviewed for history and treatment where possible.

Results: We found a total of six cases in six years. Three cases were seen in the last quarter of 2005. Five were boys. The age range spread from 5–13 years. We had access to the latter five case notes. Four of these were the owners of the device themselves. Three had blanching of the affected area, which all resolved spontaneously. All five were in digits. One had the Epipen still stuck intraosseously on arrival. None had systemic adrenergic effects. All were peanut allergic.

Discussion: Although the Epipen device has proven efficacy in the setting of severe anaphylaxis, presentation with accidental injection is occurring with increasing frequency. Four previous cases in children, all without serious sequelae, have been described in world literature (1 in 1999, 3 in 2004). However, inappropriate administration of intravenous adrenaline was the cause of death in a case of mild allergy in a child. Local infiltration with phentolamine is advised for severe cases of vasoconstriction, though none of our patients required this. In 2004, 36 700 prescriptions for Epipen Junior were written, the equivalent of three per 1000 of the <16 year population. In the study period, only one use of Epipen for anaphylaxis was identified in our department. The potential risks of carrying the device must be balanced against the potential life saving benefits in the case of anaphylaxis. Previous estimates of accidental discharge rate of 1:50 000 units is an underestimate in children. In the past year our calculated rate is 1:60 in children. Good patient and parent education (especially in boys) is necessary to prevent comorbidity from the device.

G66 TO USE IO ACCESS OR NOT TO USE IO ACCESS: DETERMINANT FACTORS ON SPECIALIST REGISTRAR’S DECISION DURING THE ACUTE CARE OF CRITICALLY ILL CHILDREN

T. Y. M. Lo, F. Reynolds.Birmingham Children’s Hospital, Birmingham, UK

Background and Aims: Emergency vascular access in the critically ill hypovolaemic paediatric patients can be difficult. Intraosseous (IO) access allows safe and rapid access to the vascular space, but little is known on the factors influencing the decision to choose IO access as the first choice of emergency vascular access in clinical practice. We aimed to determine the proportion of Specialist Registrars (SpR) among the West Midlands Deanery who would use IO access as their first choice of vascular access when managing critically ill hypovolaemic patients, and to identify factors influencing their decisions.

Methods: A prospective survey was conducted involving all Paediatric SpR from the West Midlands Deanery. The survey used a pre-design proforma to collect data on their previous paediatric experience including IO training and experience, and their likely first choice of vascular access using three hypothetical cases. These cases were modified from real life scenarios and included a critically ill hypovolaemic toddler with meningococal sepsis in a peri-arrest state with no hope of intravascular (IV) access (case i); an unwell infant with moderate dehydration but had some hope of IV access (case ii); and an arrested ex-premature infant with bronchiolitis and dehydration with no hope of IV access (case iii).

Results: Eighty four of the 105 SpR contacted participated in the survey with a median of 6.0 years previous paediatric experience (range 3.0 to 19.0 years). They were all APLS trained and 31 had instructor status. 71% of the SpR had previous real life experience in IO insertion. Significantly more of these trainees had more than six years’ previous paediatric experience (p<0.001, χ² test). Career intention and instructor status had no associations with real life IO insertion experience. 32% and 52% of the SpR chose to use IO as their first choice of vascular access for case i and case iii respectively and all trainees chose IV first for case ii. Those with previous real life IO insertion experience were significantly more likely to choose IO as their first choice of vascular access during resuscitation for cases i and iii (p<0.05, χ² test).

Conclusions: SpR are reluctant to use IO access as their first choice of vascular access when resuscitating critically ill shocked patients, but having previous real life IO insertion experience significantly reduces this reluctance.

G67 INJURY PATTERNS, DEMOGRAPHICS, AND RISK FACTORS IN PAEDIATRIC TRAMPOLINE INJURIES: AN AUSTRALIAN EXPERIENCE

C. Begg², J. Hort¹, F. Fahy¹, M. McCaskill¹.¹Children’s Hospital at Westmead, Sydney, Australia; ²Royal Hospital for Sick Children, Yorkhill, Glasgow, UK; ³Paediatric Department, Crosshouse Hospital, Kilmarnock, UK

Background: Trampolines are an increasingly popular family recreational purchase. They have previously been reported as a significant cause of injury to children. There are now an estimated 400 000 trampolines in Australia, most of which are in private backyards. Work in other countries suggests that most injuries occur unsupervised and at home.

Aims: To describe the epidemiology of trampoline injuries in an Australian paediatric population and to identify risk factors, injury patterns, and preventative strategies.

Design: A descriptive study of a consecutive series of patients.

Setting: The emergency department (ED) of an urban, tertiary children’s hospital.

Participants: All children presenting to ED with a trampoline related injury from 1 January 1996–31 December 2004

Methods: Patients were selected from the hospital Trauma Database, their medical records were reviewed, and data collated. Further case finding was carried out using the ED electronic patient records. Ethical approval was sought and obtained from the hospital’s research ethics committee.

Results: 503 patients presented to ED during the study period (average 63 per annum). The annual figure fell from 59 in 1996 to 33 in 1997; then peaked at 86 in 2000 before falling again to 23 in 2001. It has since climbed annually to reach 113 in 2004. Most injuries occurred in 3–8 year olds (47%); most were females (57%), and over 80% were at home. 144 patients (29%) required formal admission to hospital and 82% of those required general anaesthesia, mainly for fracture management. Upper limb injuries were commonest (>70%), followed by minor head injury. Upper limb injuries tended towards the severe end of the spectrum: supracondylar fracture of the humerus was commonplace. A smaller number had major head injury, cervical spine, or splenic injury. Amongst hospitalised patients, the mode Injury Severity Score (ISS) was 9 (1–16) and average length of stay was 1.7 days. Lack of adult supervision and multiple users on the trampoline were associated with injury.

Conclusion: Trampoline injuries are an increasing cause of morbidity in our paediatric population. While further work is needed to clarify risk factors, most injuries occurred in unsupervised domestic environments. There is thus a need for greater parent and user education.

G68 LIMB FRACTURES AS AN INDICATION OF NON-ACCIDENTAL INJURY

J. Bayreuther, A. MacGregor, T. Sajjanhar.University Hospital Lewisham, London, UK

Aims: The recognition of possible non-accidental injury (NAI) in children is an area of ongoing concern for all emergency departments. Limb fractures in non-mobile children should alert clinicians to consider NAI and refer the child for further management. We looked into the practice in our own paediatric emergency department (PED) to identify if this was occurring.

Method: The study took place in a large PED which sees 30 000 children per year. We carried out a retrospective review of the notes of all children under 1 year of age presenting with a diagnosis of a limb fracture, over a 4.5 year period. We identified patients by their discharge diagnosis on our computerised Accident and Emergency (A&E) system.

Results: In the time period studied, a total of 20 497 children under the age of 1 year presented to the PED. A total of 276 presented with limb injuries, of which 41 were found to have limb fractures. Of these only 37 sets of notes were available. Five infants under 6 weeks had clavicular fractures, presumed to be birth injuries and were therefore excluded. Of the 32 remaining, seven were thought to be probable NAI and admitted for further management. Of the 25 remaining, seven were referred to a senior paediatrician (specialist registrar or above) and of these, three were subsequently referred to social services. This left 18 children under 1 year who were seen and managed solely by a paediatric A&E senior house officer. All of these children were over 6 months. The types of fracture varied (2 clavicle, 2 tibia/fibula, 1 femur, 1 humerus, 6 radius/ulna, 4 fingers). Only two infants had their developmental history documented (1 by a nurse). None was discussed with social services. In the opinion of the reviewing clinician four of these cases exhibited worrying features (2 delayed presentations, 2 inconsistent mechanisms) and should have been referred for further management.

Conclusion: Limb fractures in non-mobile children account for a very small proportion of ED attendances. There needs to be increased awareness of the potential for NAI in this population. Documentation of developmental history must be emphasised to show that this has been considered in relation to NAI. To improve safeguarding children we feel that all children under 1 year of age with a limb fracture should be discussed with a senior paediatrician.