Search tips
Search criteria 


Logo of emermedjEmergency Medical JournalVisit this articleSubmit a manuscriptReceive email alertsContact usBMJ
Emerg Med J. 2007 October; 24(10): A1–A10.
PMCID: PMC2658462

Roderick Little Prize Session

Wednesday 26 September 2007, Memorial Hall, Sheffield City Hall

001 Does pre‐hospital thrombolysis increase the number of patients with an aborted myocardial infarction?

L. Jackson1, J. Kendall2, N. Castle3. 1Bristol Royal Infirmary; 2Frenchay Hospital, Bristol; 3Frimley Park Hospital, UK

BackgroundAn “aborted myocardial infarction” is defined as an acute coronary syndrome where there is rapid resolution of existing ST elevation associated with a CK rise less than twice the upper limit of normal, or a small troponin release compatible with minimal myocyte necrosis. Previous research has shown that earlier thrombolysis is associated with a higher rate of aborted infarct. It is also known that pre‐hospital thrombolysis reduces the pain to needle time. Therefore, our hypothesis was that pre‐hospital thrombolysis would be associated with a higher incidence of aborted infarct, and we set out to test this.

MethodsRetrospective analysis of all patients given pre‐hospital thrombolysis in the Avon ambulance region between April 2004 and October 2006, and data from a similar period at Frimley Park Hospital. The control group were patients given in‐hospital thrombolysis at Frenchay or Frimley Park over the same period. 12‐h troponin, call to needle time, pain to needle time and incidence of aborted infarct were analysed.

Results69% of pre‐hospital patients had a pain to needle time of 2 h or less, compared to 40% of in‐hospital patients (p<0.001). The overall incidence of aborted infarction was 21.4%. Of those with aborted infarction with pain to needle times available (84) 55% had a pain to needle time <2 h, 26% 2–4 h and 18% >4 h. Despite this, 24.4% of in‐hospital patients had a troponin I (TnI) less than 0.5, compared to 12.7% of pre‐hospital patients (p = 0.01).

ConclusionAlthough pre‐hospital thrombolysis improved pain to needle time, and a shorter pain to needle time increased the incidence of aborted infarction, pre‐hospital thrombolysis was associated with a statistically significant decrease in the number of aborted myocardial infarctions, using troponin alone. Further work is required to understand this unexpected finding.

002 Back to the future: discharging chest pain patients without investigation. The MACS (Manchester Acute Coronary Syndrome) clinical decision rule

R. Body, S. Carley, G. McDowell, J. Ferguson, K. Mackway‐Jones. Manchester Royal Infirmary, Manchester, UK

BackgroundRisk stratification of patients with suspected acute coronary syndromes in the emergency department (ED) is a topical and contentious issue. Various algorithms exist but there is no consensus regarding their use and none facilitate immediate discharge without cardiac enzyme testing.

ObjectivesTo derive a clinical decision rule (CDR) incorporating clinical data available at presentation. The rule should enable confident early discharge of low risk patients while risk stratifying other patients to facilitate triage to an appropriate level of care.

MethodsConsecutive patients presenting to the ED with suspected cardiac chest pain within the past 24 h were eligible. The lead investigator (RB) was always contactable. Variables that have previously been shown to predict diagnosis of acute myocardial infarction (AMI) or to predict outcome as part of previously validated algorithms were prospectively recorded. Patients underwent 12‐h troponin T estimation and were followed up by telephone and chart review at 30 days for the composite primary outcome of AMI, death or urgent coronary revascularisation (UR). Variables that were univariate predictors (p<0.05) of outcome were entered into a multivariate analysis using recursive partitioning. It was estimated that recruitment of 800 patients would ensure adequate power for the derivation of a 14‐point CDR.

Results804 patients were recruited between January 2006 and February 2007 (90% by RB). Seven patients were excluded. None was lost to follow‐up. A 13‐point CDR effectively risk stratified patients (table). The no‐risk group (9.9% of the study population) may be eligible for immediate discharge (sensitivity for AMI 100% (95% CI 97.6–100%, sensitivity for 30‐day events 100% (98–100%).

ConclusionsThe MACS rule effectively risk stratifies patients with chest pain. A significant proportion may be discharged immediately, which could have tremendous implications for healthcare resources. Internal validation and six‐month follow‐up data are pending. External validation is also warranted before implementation.

Table thumbnail
Abstract 002

003 Is (3×age)+7 a better estimate of children's weights than the APLS formula in paediatric resuscitation?

B. Owens1, M. Luscombe2, D. Burke3. 1Kingsmill Hospital; 2Doncaster Hospital; 3Sheffield Children's Hospital, UK

AimsA recent study from Nottingham showed that the Advanced Paediatric Life Support (APLS) formula (weight in kg  = (age+4)×2) was a significant underestimate of children's weights and that (3×age)+7 was more accurate. The aim of this study was to see if this could be reproduced in a separate population.

MethodsThe age, sex and weight of 25 000 children (aged 1–10 years) attending Sheffield Children's Hospital Emergency Department from January to December 2006 was collected retrospectively. Their actual weights were compared to their estimated weights from both formulae and the difference was calculated and expressed as a percentage for each formula.

ResultsThe APLS formula was found to be a significant underestimate of children's weights at all ages with a mean of 22.2% (95% CI of +/−0.32%) with a maximum of 39% at age 10. (3×age)+7 was a mean underestimate of 2.36% (95% CI of +/−0.26%).

DiscussionIn paediatric resuscitation a quick, simple, accurate estimation of a child's weight is essential to calculate drug doses and fluid boluses, etc. This study confirms that the APLS formula is a significant underestimate of children's weights at all ages from 1–10 and that the formula weight(kg)  = (3×age)+7 is a more accurate estimation. The use of an inaccurate formula cannot be considered to be best practice. Therefore, in order to minimise the underdosing of resuscitation drugs the authors recommend the formula (3×age)+7 as a safe and accurate estimate of the weight of the modern day child.

004 Haemolysis and blood sampling technique: a prospective study

N. P. Laundy, C. Roche, R. Parry. Emergency Department, Countess of Chester Hospital, UK

Null hypothesisThere is no difference in haemolysis rate between samples collected with a vacutainer (BD Diagnostics, Franklin Lakes, NJ, USA) and those collected with a syringe.

DesignProspective study of 510 blood samples collected during October 2006.

SettingUK district general emergency department.

MethodsPractitioners completed a study proforma at the time of blood sampling—prior to sample analysis. Practitioners were specifically instructed to use their standard collection technique. Details of: syringe/vacutainer use; technique of transfer to tube (if syringe were used); needle/cannula size; time and date; hospital number; and rating of difficulty of collection were collected. Results were later checked to determine if a sample had haemolysed.

ResultsSee table.

Table thumbnail
Abstract 004

DiscussionHaemolysis causes frequent frustration, requiring further samples to be obtained. This may cause pain and inconvenience to the patient and has implications on 4‐h targets. The finger of blame is consistently pointed at the practitioner collecting the sample. This study has disproved our null hypothesis, and in fact shown a statistically significant higher haemolysis rate when the vacutainer is used, as opposed to the traditional syringe. It also confirms the association between perceived difficulty in taking the sample and haemolysis rate.

ConclusionUsing a syringe appears to have a lower haemolysis rate to that of a vacutainer, however this has to be balanced against risk of needle stick injury and trust risk management protocols.

005 In the Clinical Decision Unit does backache have to follow headache?

J. Cooper2, B. Smith1, T. Hassan1. 1Leeds General Infirmary; 2Leeds General Infirmary & St James's University Hospital, UK

Introduction and AimsLone acute sudden headache is a common presentation to the emergency department (ED) and the investigation thereof is a subject of much current debate. The exclusion of subarachnoid haemorrhage (SAH) as a cause of sudden onset headache is the main thrust of most investigative strategies and has historically been undertaken with non‐contrast CT head which, if negative, is followed by a lumber puncture with CSF analysis for xanthochromia. This study aimed to determine whether there was benefit in following a normal CT head with lumber puncture in every case.

MethodsRetrospective analysis of consecutive patients entered on to a standardised CDU pathway at two large tertiary urban hospitals for the exclusion of SAH as a cause of lone acute sudden headache were included over a three‐year period. Patients were excluded if they were <16 years, GCS was <15, there was clinical evidence of meningitis, seizures, raised ICP or pyrexia. Once entered on the protocol patients were analysed on an intention‐to‐treat basis. Results of CT head, lumber puncture, subsequent investigations and outcomes were analysed.

Results525 patients were included age 16–89 years (mean 39.9 years). Of these 521 patients proceeded to CT head scan. 13 patients (1.7%) had a positive CT scan for SAH, 12 patients had other CT diagnoses for their headache and 496 patients had normal CT head scans. Six patients, later reported as abnormal, had CT scans initially reported as normal and therefore 502 patients were eligible for lumber puncture. Lumbar puncture was not performed in 189 (36%) due principally to medical staff assessment of risk (98 patients) or refusal of patient consent (57 patients). Of the 313 lumber punctures performed, 10 were positive for xanthochromia or uniformly blood stained, 19 had other diagnoses from lumber puncture and 284 were normal. Of the patients who did not have lumber puncture there have been no adverse events to date. Of the 10 patients with positive lumber punctures, nine had normal cerebral angiography and only one proceeded to have surgical intervention.

DiscussionThe population of patients with lone acute headache reaching CDU certainly does not represent all comers to the ED and by definition only includes patients with a normal GCS. Therefore it is not surprising that non‐contrast CT head was positive for SAH in only 13 (1.7%) of this population. The utility of lumber puncture in these patients is matter of much debate and this is reflected by the finding that only 313 of 502 eligible patients proceeded to lumber puncture. Furthermore, of the 10 patients with positive lumber punctures only one required any surgical intervention.

ConclusionsAn lumber puncture is not without risk and the chance that it will change management in this population after a normal CT head is <0.2% and this should be involved in discussions about consent.

006 How crowded are English emergency departments?

T. Locker1, S. Mason2. 1Sheffield Teaching Hospitals Trust; 2University of Sheffield, Emergency Department, Northern General Hospital, UK

IntroductionEmergency department crowding has been extensively described in North American emergency departments, but little is known about its extent in UK practice. This study aimed to determine the extent of crowding in a sample of UK emergency departments and whether any departmental characteristics were predictive of crowding.

MethodsData from 23 emergency departments describing attendances over a one month period were analysed. Each time a patient arrived, was seen or left the department the number of patients in the department was determined and expressed as a percentage of the total cubicle spaces. Peak and mean daily occupancy levels were determined. Linear regression was used to explore factors predictive of mean daily occupancy.

Results112 299 episodes were included in the analysis. Maximum occupancy in each department ranged from 137.5% to 334.8% of the available cubicles. Mean daily occupancy in each department ranged from 41.7% to 147.1% (median 66.1%). In a multivariate linear regression model only one variable, the number of attendances per annum, was found to be predictive. This factor explained 49.5% of the variability in mean daily occupancy and showed that mean daily occupancy would be expected to increase by 7.2% for each additional 10 000 attendances per annum.

ConclusionThere is considerable variation in the degree of crowding both within and between emergency departments. Larger departments are more likely to be overcrowded. This may offer one explanation why larger departments have been found previously to perform poorly. The findings of this study should be taken into account when planning reconfiguration of emergency department services.

Free Paper Session 1: Diagnostic and Risk Tuesday 25 September 2007, 14:30–16:00, John Barbirolli Room, Sheffield City Hall

007 Does a symptom‐specific blood testing protocol at triage reduce the number of patients admitted for a short duration to the emergency department observation ward?

M. Baker, J. Coates, S. Hunter, S. Carss. Queen Alexandra Hospital, Portsmouth, UK

IntroductionQueen Alexandra Emergency Department is a large district general hospital in Portsmouth seeing in the region of 120 000 patients a year. A departmental audit of patients admitted to the observation ward during September 2005 highlighted a number of patients admitted for a short duration to await blood results. This admission was associated with paperwork for nursing staff, the transfer of patients by nurses, cleaning of bed spaces and may be unsettling for patients and their relatives.

MethodWe introduced a symptom‐specific blood testing protocol at triage so that bloods were sent early, these tests were agreed with in‐house teams to minimise the later adding on of tests. We hoped that results would then be available earlier in the patient journey, ideally at the time of clinical assessment. A re‐audit was conducted for September 2006.

ResultsSee figure.

figure emAbsJuly07.f1
Abstract 007 Waiting time for blood results.

DiscussionThe introduction of a symptom‐specific blood testing protocol at triage resulted in a reduction in the number of patients admitted to the observation ward for durations under 2 h. A longer duration of admission was associated with waiting for paracetamol 4‐h levels and 12‐h troponin levels. This simple intervention reduced the time patients were waiting in the department for blood results and improved patient flow within the 4‐h target. Although nursing staff were taking the blood tests it reduced their workload in terms of paperwork for the admission, the transfer of patients and the cleaning of bed spaces.

008 Value of serum procalcitonin, neopterin and C‐reactive protein in screening patients presenting with bacterial or viral respiratory tract infections: a prospective study

T. H. Rainer1, C. P. Y. Chan2, M. F. Leung2, W. M. Leung2, C. A. Graham1, M. Ip3, G. W. H. Cautherley4, D. Fuchs5, R. Renneberg2. 1Accident and Emergency Medicine Academic Unit; 2Department of Chemistry, The Hong Kong University of Science and Technology; 3Department of Microbiology, The Prince of Wales Hospital, Chinese University of Hong Kong; 4R&C Biogenius Limited, Hong Kong; 5Division of Biological Chemistry, Biocentre, Innsbruck Medical University & Ludwig Boltzmann Institute of AIDS‐Research, Austria

IntroductionRespiratory tract infections (RTIs) are a huge global challenge of which over 99% are of bacterial or viral aetiology. A pathological agent can be identified most often in less than 50% of RTI cases. The use of biomarkers to accurately rule in or rule out bacterial and viral infection may improve the appropriate management of RTI.

ObjectivesThis study is the first to investigate the roles of C‐reactive protein (CRP), neopterin and procalcitonin as screening tools for bacterial and viral infection in a mixed patient population presenting with lower respiratory tract infection (LRTI) of uncertain aetiology.

MethodologyWe investigated the incremental value of CRP, neopterin and procalcitonin to differentiate bacterial from viral aetiology in 429 patients with suspected RTI attending the Department of Accident and Emergency, and also to perform a sensitivity analysis on cases of uncertain aetiology.

ResultsIt shows that all three markers—CRP, neopterin and procalcitonin—may have a role in differentiating bacterial from viral causes, but that the CRP/neopterin ratio yields optimal differentiation. The areas under the ROC curves to distinguish bacterial from viral infections were 0.735 (95% CI 0.644 to 0.825; p<0.05) for CRP, 0.666 (95% CI 0.568 to 0.765; p<0.05) for PCT, and 0.772 (0.688 to 0.856; p<0.05) for neopterin. The area under the curve was higher if the markers were used in combination, with 0.834 (0.756–0.911; p<0.05) for CRP/neopterin ratio.

ConclusionsThe CRP/neopterin ratio may be used to rule in and rule out bacterial and viral infection with a high degree of accuracy.

009 Comparison of prediction rules for nursing home‐acquired pneumonia and community‐acquired pneumonia

C. A. Graham, S. Y. Man, P. S. K. Mak, A. H. Y. Yu, C. S. K. Cheung, P. S. Y. Cheung, T. H. Rainer. Accident and Emergency Medicine Academic Unit, Chinese University of Hong Kong, Hong Kong

BackgroundPrediction rules exist for the assessment of community‐acquired pneumonia (CAP) but their use in nursing home‐acquired pneumonia (NHAP) remains undefined.

AimThe objectives of this study were to evaluate the prognostic ability of five CAP decision rules (the pneumonia severity index (PSI), CURB‐65, the modified American Thoracic Society rule (M‐ATS), a modified rule for severe CAP (m‐ATSSCAP), and the España prediction rule) when applied to NHAP.

MethodsA prospective observational study of 1524 consecutive inpatients with pneumonia (1168 CAP and 356 NHAP) was performed in a university teaching hospital in the New Territories between 1 January 2004 and 30 June 2005. The primary outcome measure was the combined 30‐day all cause mortality and intensive care unit (ICU) admission.

ResultsThe overall mortality rate was 7.5% and the ICU admission rate was 3.2%. Compared with CAP, the NHAP group had a higher mean (SD) age (84 (9.1) vs 69 (17.6)), a higher proportion of cerebrovascular disease (31.5% vs 12.4%) and chronic obstructive pulmonary disease (19.9% vs 15.1%) and a higher combined poor outcome rate (14.0% vs 8.8%). The table shows the areas under the receiver operating characteristic (ROC) curves. Eight predictors for poor outcome for NHAP were identified which were similar to those identified in the CAP group.

ConclusionNHAP had similar disease prognostic variables as with CAP. There were no significant differences in the performance of PSI, CURB‐65 and m‐ATSSCAP between the two groups.

Table thumbnail
Abstract 009 Area under ROC curve (95% CI)

010 Novel role of TBET and GATA‐3 transcription factors mRNA in acute coronary syndrome

T. H. Rainer, R. Chan, P. Mak, C. A. Graham. Accident & Emergency Medicine Academic Unit, Chinese University of Hong Kong, Hong Kong

BackgroundAcute coronary syndrome (ACS) is the leading cause of death and the leading cause of disability‐adjusted life years worldwide. In this study we will determine the role of Th‐1/Th‐2 imbalance by measuring the mRNA ratio of TBET/GATA‐3 transcription factors in the diagnosis and prognosis of ACS.

MethodsIn a prospective study between 7 September and 8 December 2006, 77 patients with a diagnosis of ACS attended the emergency department of a university hospital in the New Territories of Hong Kong. We used the strictly quantitative real time RT‐QPCR to determine the mRNA expression of T‐BET, GATA‐3 in buffy coat. Expression of transcripts was normalised by GAPDH.

ResultsThe median TBET/GATA‐3 ratios increased in patients with unstable angina compared with control and were highest in patients with AMI (p<0.001). The areas under the ROC curve for TBET/GATA ratio and cTnT for predicting AMI versus no AMI (n = 16 vs n = 41) are 0.939 (95% CI 0.842 to 0.985) and 0.839 (95% CI 0.718 to 0.923) respectively (difference 0.1; p = NS). At a TBET/GATA‐3 ratio cut off of >1.5, the optimal sensitivity and specificity were 100.0% (95% CI 79.2 to 100.0) and 85.4% (95% CI 70.8 to 94.4). At a troponin cut off of [gt-or-equal, slanted]0.06, the sensitivity and specificity were 56.2% (95% CI 29.9 to 80.2) and 100.0% (95% CI 91.3 to 100.0). Allowing for interactions, both cTnT and the TBET/GATA‐3 ratio predict AMI and have an incremental effect in diagnosis.

ConclusionsThe TBET/GATA‐3 ratio is highly sensitive and specific for predicting AMI, and compares favourably with a first cTnT. Along with cTnT, it may add incrementally to the early diagnostic accuracy of ACS.

011 Does the timing of normal CT head affect the likelihood of a positive lumbar puncture in the investigation of headache on the CDU?

B. Smith2, J. Cooper1, T. Hassan2. 1Leeds General Infirmary & St James's University Hospital, Leeds; 2Leeds General Infirmary, Leeds, UK

Introduction and AimsThe investigation of lone acute sudden headache has traditionally involved a non‐contrast CT head scan followed by a lumbar puncture and CSF analysis. New generation CT head scanners have increased sensitivity in the diagnosis of subarachnoid haemorrhage (SAH) especially when performed close to headache onset. This study aimed to identify whether the timing of a normal CT head scan influenced the likelihood of a positive lumbar puncture in a population of patients with lone acute sudden headache and GCS 15.

MethodsRetrospective analysis of consecutive patients entered on to a standardised CDU pathway at two large tertiary urban hospitals for the exclusion of SAH as a cause of lone acute sudden headache were included over a three year period (2004–6). Patients were excluded if they were <16 years, GCS was <15, there was clinical evidence of meningitis, seizures, raised ICP or pyrexia. Once entered on the protocol patients were analysed on intention to treat basis. The timing of the CT head scan in relation to headache onset was documented and normal scans were spilt in to five groups (see table). These patients were then analysed further to determine the number in each group that proceeded to have a positive lumbar puncture result and their subsequent outcomes. A positive lumbar puncture was taken to be either xanthochromia positive or uniformly blood stained.

ResultsSee table.

Table thumbnail
Abstract 011 Total number of normal CT head scans (n = 508)

DiscussionThe role of lumbar puncture in patients with lone acute sudden headache is a subject of much conjecture. The yield of lumbar puncture in the population of patients GCS 15 presenting to the CDU is low with only eight patients (1.5%) having a positive result. However only one of these patients (0.2%) proceeded to have an abnormal cerebral angiogram and require intervention. The chances of a positive lumbar puncture in relation to the timing of normal CT head does not appear to be significant. Lumbar puncture also has a role in the diagnosis of other conditions causing headache in these patients, largely viral meningitis.

ConclusionsLumbar puncture is a procedure not without risk but the chance of missing a SAH remains a significant concern. The yield of lumbar puncture in this group of patients is low and this does not appear to be related to the timing of the CT head scan. These figures should aid discussion at the bedside in obtaining informed consent (or declination thereof) for lumbar puncture.

012 Modified Early Warning (MEW) score predicts in‐hospital mortality in unselected emergency department patients

C. Vorwerk, L. Gray, L. Florence, T. Fudge, C. Goss, T. Coats. Leicester Royal Infirmary, Leicester, UK

BackgroundEarly risk stratification of patients presenting to the emergency department is the key to preventing sudden catastrophic deterioration and adverse outcome. Risk assessment tools have to be easy to use and to be able to identify patients at risk. In recent years various physiological scoring systems have been introduced into ward based patient's care, some of which, such as the Modified Early Warning (MEW) score, are now in use in emergency departments. Very little is known as to the usefulness of an initial ED MEW score in risk stratifying unselected ED patients and to predict mortality.

ObjectivesTo assess the relation between MEWS on arrival in ED and 28‐day in‐hospital mortality.

MethodsProspective observational study of all ED patients (age >16) who presented to the Leicester Royal Infirmary during 1–13 March 2007, requiring subsequent hospital admission. Patients were excluded if they had no vital signs recorded on arrival, were transferred from another hospital, directly admitted to the ward or transferred to a different NHS trust. Each morning all patients admitted from ED during the previous 24 h were identified and their MEW score calculated. Patients were followed‐up for 28 days. Logistic regression, ANOVA and ROC analysis were performed.

ResultsA total of 1024 patients were enrolled, among these were 50 (4.9%) deaths. Non‐survivors had a significantly higher MEWS (4.5 vs 1.3, p<0.001) than survivors. Logistic regression showed MEWS to be a significant predictor of in‐hospital death in unselected ED patients (OR 1.7; 95% CI 1.5 to 1.8). ROC analysis for MEWS as predictor of mortality demonstrated a good performance (AUC 0.8).

ConclusionOur results indicate that the MEW score, on arrival in the emergency department, is a strong predictor of 28‐day in‐hospital mortality in unselected emergency department patients.

Free Paper Session 2: Interventions and Outcomes Wednesday 26 September 2007, 15:15–16:15, Memorial Hall, Sheffield City Hall

013 Evaluation of different tourniquet types in terms of pain, success rate and patient satisfaction

F. Wisniacki1, K. Jones2, E. Afolabi3, B. Pirnau3, A. Rendon3, R. Brown3. 1Ealing Hospital, London; 2South Thames Rotation; 3St Mary's Hospital, London, UK

BackgroundClinicians use a variety of tourniquet types during cannulation and/or venepuncture including disposable gloves. There is no evidence of the benefit of traditional purpose made tourniquets over such ad hoc tools. The objectives of this study were to explore the relative impact of the use of a glove, disposable tourniquet or reusable traditional tourniquet on pain, success rate and patient satisfaction.

MethodsPatients were recruited from the emergency department (ED) and a surgical ward and all patients gave consent. All of the venepuncture and cannulation procedures studied were performed by trained phlebotomists. On any one shift, the phlebotomist was randomly allocated to use glove, reusable tourniquet or the disposable tourniquet. Patients were asked to mark their pain on a visual analogue scale before and after the procedure. Patient satisfaction was assessed using the five‐point Likert scale and success of the procedure was recorded.

Results1378 patients were included in the study within a three‐month period (479 disposable/460 glove/439 traditional). Patients in whom the traditional tourniquet was used for the procedure had the lower median pain score (5 mm/interquartile range (IQ) 3–10) versus disposable (median 9.5/IQ 5.5–16) and glove (median 17/IQ 8–23.8) (p<0.001). This difference remains statistically significant after adjustment for covariates including site of procedure, age, gender and whether the patient was in the ED or on the surgical ward. The pain score following the procedure showed similar results. A greater number of procedures in the disposable group were unsuccessful (9.6%) versus glove (5.7%) and traditional (5.2%) (p = 0.01). There was no difference in overall patient satisfaction of the procedure.

ConclusionsThe traditional reusable tourniquet was shown to be associated with less pain and complications than the disposable or glove.

014 The ESCAPE Trial: Effectiveness and Safety of Chest pain Assessment to Prevent Emergency admissions

S. Goodacre1, L. Cross1, C. Lewis2 on behalf of the ESCAPE Research Team 1University of Sheffield; 2Research Associate, University of Liverpool, UK

BackgroundEmergency medical admissions are progressively rising, with approximately one in four being due to chest pain. Low risk patients with undifferentiated chest pain are less likely to be admitted if they are managed on a chest pain unit (CPU). We aimed to determine whether introducing CPU care would reduce the overall proportion of patients admitted with chest pain and reduce daily emergency medical admissions.

MethodsWe randomised 14 hospitals to either establish CPU care or continue providing routine care and measured outcomes for a year before and after intervention. A random effects multilevel model was used, with the individual hospital attended as a random effect and age, gender, hospital allocation (CPU or control) and time (before or after intervention) as covariates, to estimate the effect of introducing CPU care upon each outcome.

ResultsChest pain related attendances increased by 3.5% at control hospitals (22 858 vs 23 666) and 16.0% at CPU hospitals (20 784 vs 24 101). CPU care did not significantly change the proportion admitted with chest pain (OR 0.998, 95% CI 0.940 to 1.059, p = 0.945), but was associated with increased daily emergency medical admissions (1.7 per day, 95% CI 0.8–2.5, p<0.001). CPU care was associated with small increases in re‐attendances (OR 1.10, 95% CI 1.00 to 1.21, p = 0.036) and re‐admissions (OR 1.30, 95% CI 0.97 to 1.74, p = 0.083).

ConclusionThe introduction of CPU care may increase, rather than reduce, emergency medical admissions.

015 The relationship between distance to hospital and patient mortality in emergencies

J. Nicholl, J. West, S. Goodacre, J. Turner. University of Sheffield, Sheffield, UK

ObjectivesReconfiguration of emergency services could lead to patients with life threatening conditions travelling longer distances to hospital. Concerns have been raised that this could increase the risk of death. We aimed to determine whether distance to hospital was associated with mortality in patients with life‐threatening emergencies.

MethodsWe undertook an observational cohort study of 10 315 cases transported to hospital with a potentially life‐threatening condition (excluding cardiac arrests) in four English ambulance services and associated acute hospitals, to determine whether distance travelled to hospital was associated with mortality, adjusted for age, sex, clinical category and illness severity.

ResultsAmbulance journey distances ranged from 0 to 58 km with a median of 5 km, and 644 patients died (6.2%). Increased journey distance was associated with increased risk of death (OR 1.024 per km; 95% CI 1.014 to 1.034; p<0.001). This association was not changed by adjustment for confounding by age, sex, clinical category or illness severity. Patients with respiratory emergencies showed the greatest association between distance travelled and mortality (see fig).

figure emAbsJuly07.f2
Abstract 015.

ConclusionIncreased journey distance to hospital appears to be associated with increased risk of mortality. Our data suggest that a 10 km increase in journey distance is associated with an approximate 1% absolute increase in mortality.

016 The 3CPO trial: The clinical effectiveness of BiPAP and CPAP in patients presenting to emergency departments with acute severe cardiogenic pulmonary oedema

A. Gray1, S. Goodacre2, D. Newby1, J. Nicholl2, M. Masson1, F. Sampson2. 1Royal Infirmary of Edinburgh; 2University of Sheffield, UK

Objectives(1) To assess the clinical effectiveness of non‐invasive ventilation (CPAP or NIPPV) in addition to standard therapy against standard therapy alone in the early management of acute severe cardiogenic pulmonary oedema (ACPO). (2) To assess whether there is any difference in the effectiveness of CPAP and NIPPV in the early management of acute severe cardiogenic pulmonary oedema.

MethodsA multicentre randomised controlled trial in 26 UK emergency departments between July 2003 and April 2007 inclusive (46 months). Patients with ACPO with the following criteria were randomised after consent (written or verbal): clinical and radiological characteristics consistent with a diagnosis of ACPO, tachypnoea (respiratory rate >20) and a H+ on arterial blood gas analysis of >45 (pCO2 <7.35). Exclusion criteria include inability to consent, life threatening illness and requirement for time dependent intervention. Patients were randomised via a telephone service to one of three arms in a 1:1:1 ratio (NIPPV and medical therapy, CPAP and medical therapy or medical therapy alone). NIPPV was instigated at 8:4 and increased to 20:10 depending on tolerance and clinical status. CPAP was started at 5 and increased to 15. Primary outcome for the NIV versus standard therapy comparison is 7‐day mortality. The NIPPV versus CPAP comparison is a composite outcome of 7‐day mortality and intubation. Secondary outcomes include physiology, arterial blood gas analysis, patient perceived breathlessness and intubation rates. Myocardial infarction rates will be compared between groups. A sample size of 1200 patients will be collected to detect a 6% difference in mortality at a power of 80%. Analysis will be on an intention to treat basis.

ResultsData collection will be completed on the 30 April 2007 and we hope to be considered for presentation as a late breaking trial.

Free Paper Session 3: Systems and Safety Wednesday 26 September 2007, 16:15–17:45, Memorial Hall, Sheffield City Hall

017 Impact of time targets on quality and safety of care provided by emergency departments

S. Mason1, J. Freeman1, S. Croft2, C. Yapp2, S. Cross2. 1University of Sheffield; 2Sheffield Teaching Hospitals Trust, UK

BackgroundThe impact of targets (positive and negative), and particularly those that are based on a time‐frame, have not been extensively studied. In 2000, the NHS set a target of 4 h between arrival and discharge or admission of patients from the emergency department (ED). However, there are concerns about the unintended consequences for patients that might arise from this stringent time‐related target such as the possibility of ED clinicians making rushed decisions without complete information, resulting in missed diagnoses, unnecessary hospitalisations or revisits to the ED for the same problem.

AimsTo evaluate the effect of the 4‐h target on quality and safety of ED patient management.

MethodsRoutine data collected over a three‐year period examined change in patient outcomes in one large urban ED in Yorkshire since the introduction of the target. Anonymised patient‐level data compared: 4‐h performance, time spent in the ED, mortality, admission rate, one‐day hospitalisations, repeat visits to the ED and investigations performed for periods before, during and after the 4‐h target.

ResultsRoutine data from 580 000 new patient episodes between 2003–5 were used in the analysis. Median time to see a clinician fell from before the target was introduced to after it was established (57.0 min to 36.0 min). Median total ED time also fell (121.0 min to 100.0 min). Overall the mean number of investigations per patient rose (1.08 to 1.21). However, the proportion of patients being admitted fell (15.2% to 14.8%), as did the proportion of patients who did not wait (6.1% to 4.0%). There was no change in ED mortality.

ConclusionsThe introduction of a target has been associated with changes the quality of patient management within the ED. Overall patients are seen more quickly, spend less time in the ED, are less likely to be admitted to hospital, but more likely to have an investigation during their stay.

018 The development of openEHR Archetypes and Templates to support emergency medicine in the NHS

T. Shannon1, M. Simpson2, S. Schneidman3, S. Moore4, A. Neary5, L. Sato5, G. Stonham5, H. Leslie6, S. Heard6, T. Beale6. 1Leeds General Infirmary; 2Hull Royal Infirmary; 3Royal Lancaster Infirmary; 4Countess of Chester Hospital; 5NHS Connecting for Health; 6Ocean Informatics

ObjectivesAs part of the NHS Connecting for Health (CfH) programme, this work aimed to begin the development of clinical content required to support clinical documentation for the “Top 10” ED complaints, using a standardised methodology to ensure clinical content is shareable across multiple clinical applications.

MethodsThis work was based upon an international standardised methodology to defining clinical content for Electronic Health Records (openEHR, en13606). The “top 10” ED presentations chosen were:

  • chest pain
  • shortness of breath
  • abdominal pain
  • back pain
  • headache
  • overdose
  • head injury
  • joint injury
  • long bone injury
  • collapse of unknown cause

Using an online collaborative wiki tool clinical team members were able to collectively define maximal data sets which could be used for clinical data capture in EHRs. Multiple inputs were taken into the development including clinical brainstorms, NeLH library, DOAS library etc. to inform this work. This input material was then passed to the clinical informatics team who then produced openEHR based standardised archetypes and configurable templates. These have been designed to support key clinical processes and to be reused in any EHR system conformant to the openEHR standard.

ResultsThis initial work has resulted in international standards based clinical documentation for common emergency medicine presentations. Further development and input from NHS stakeholders will be required to inform the further development of this evolvable approach.

ConclusionsThe results of this emergency medicine oriented work has formed the first pioneering iteration of a new international standards based approach to developing clinical documentation within NHS Connecting for Health. Important lessons have been learned that have informed the development of clinical content across the rest of the NHS.

019 Can inter‐agency working reduce binge drinking and emergency department attendances?

J. Benger, R. Carter. Bristol Royal Infirmary, Bristol, UK

Background“Binge drinking” and associated harms in terms of health, crime and disorder have been highlighted by the government and media, causing considerable public concern. This research aimed to quantify the number of patient attendances to an urban adult and children's emergency department (ED) that are directly attributable to binge drinking, and investigate ways in which the inter‐agency sharing of anonymised information may be used to design, implement and monitor interventions to reduce these harms.

MethodsIntoxicated patients attending either the adult or children's ED were prospectively identified by nursing staff, and anonymised data collected by a dedicated researcher. Collaboration and data sharing between health, police, social services, university experts and local authorities was achieved through the establishment of steering and operational groups with agreed objectives and the formation of a shared, anonymised database relating to alcohol use and harms.

ResultsThe proportion of patients attending the ED as a result of binge drinking was 4% in adults and less than 1% in children. 70% of patients were male, with a mean age of 30 years, and 72% attended between the hours of 20:00 and 08:00. The commonest reason for ED attendance was accident (34%), followed closely by assault (30%). 27% of patients had done most of their drinking at home, 36% in a pub and 16% in a nightclub. Inter‐agency collaboration proved highly successful: pooling of anonymised data created a much clearer picture of the extent of the problem and immediately suggested strategies for intervention, which could be effectively directed and monitored by the steering group.

DiscussionOur figures appear accurate, but much lower than some previous reports. This may be attributable to a requirement that patients be both objectively intoxicated and attending the ED as a result. Specific issues were identified in relation to “pre‐drinking” and “drink spiking”, but the change in licensing law that occurred during November 2005 does not appear to have had a significant impact to date. The initiative to achieve inter‐agency collaboration and data‐sharing was highly successful, with considerable potential for the development and implementation of interventions that will reduce binge drinking and emergency department attendance.

020 Evaluating acute medical admissions through emergency departments in Hong Kong: can one adjust for case‐mix variation?

T. H. Rainer1,2, P. Sollich3, T. Piotrowski3, A. C. C. Coolen3, M. Y. Lam1,2, C. Y. Man1,2, N. K. Cheung1,2, C. H. Cheng1,2, B. Cheng4, C. A. Graham1,2. 1Accident and Emergency Medicine Academic Unit, Chinese University of Hong Kong; 2Trauma and Emergency Centre, Prince of Wales Hospital; 3Department of Mathematics, King's College London; 4Head Office, Hospital Authority

IntroductionThe purpose of this study was to develop a standardised system for assessing the performance of different hospitals in admitting patients with medical disorders.

ObjectivesThe objectives were firstly to derive a standardised admissions tool based on the average medical admissions for Hong Kong as a whole during 2001, secondly to validate the underlying statistical model using a second dataset from 2002, thirdly to use this model as a benchmark to assess each admitting hospital emergency department in both 2001 and 2002, and finally to test the tool in an extreme setting, namely 2003 in which there was the SARS outbreak and an introduction of charges for emergency department attendances.

MethodsThis was a retrospective study of patients attending emergency departments and who were coded “medical”. Data were retrieved from the HA database system, and a multiple logistic regression model was developed with four independent variables (age, gender, triage category and ambulance call) and one dependent variable (hospital admission or discharge).

Results912 992 patients attended emergency departments between 1 January 2001 and 31 December 2001, of whom 278 269 (30.5%) were admitted to medical wards and 12 965 (1%) died. A further 1 618 233 patients attended emergency departments between 1 January 2002 and 31 December 2003. A model was derived from 2001 data. It reproduced observed admissions probabilities very well, with a root mean square deviation between actual and predicted admissions probabilities of 0.72%. The model also ranked admissions probabilities of different patients well, giving an area under the ROC curve of 90.3% (95% CI +/− 0.11%). The root mean square deviations for 2002 and 2003 were 3.6% and 4.9% respectively, while the areas under the ROC curves for 2002 and 2003 were 89.9% (95% CI+/−0.11%) and 89.0% (95% CI+/−0.12%) respectively. Reliability diagrams showed that one hospital admits more patients across all groups than is the case in the average across all hospitals, that another hospital tends to admit fewer patients who are less severe, while yet another hospital tends to admit fewer patients who are more severe cases.

ConclusionIn conclusion, a tool for benchmarking hospital medical admissions and minimising case‐mix variation has been derived and validated. This may be used to evaluate the performance of departments against a common standard.

021 Effectiveness of the Doubt Kills campaign

A. A. Mohammed, M. Patel, M. Cooke. Heart of England NHS Foundation Trust, UK

BackgroundThe British Heart Foundation's Doubt Kills campaign aims to increase awareness of heart attack symptoms, to reduce pre‐hospital mortality from myocardial infarction and to reduce the time delay in seeking help in patients particularly targeting South Asian communities. This campaign started in November 2006 and will be continuing until October 2007.

ObjectivesTo evaluate the early effects (first 4 months) of this campaign on the emergency departments of our Trust hospitals with respect to numbers presenting with chest pain and detection of cardiac disease.

MethodologyPatients with a presenting complaint of chest pain or associated symptoms were included. The primary outcome is the number of patients who attended with chest pain before and during the campaign. Subgroup analysis was undertaken by eventual diagnosis, age sex, ethnic origin, investigations undertaken, speciality consultations and mode of arrival. Analysis is undertaken by comparison of means before and after including multivariate analysis and also using statistical process control (SPC).

ResultsThe average attendances changed from 712 per month to 884 per month, mostly in the non‐cardiac group (345 vs 494; p = 0.004) rather than the cardiac (366 vs 398; p = 0.163). The increase was more marked in the south Asian population, those discharged with a non‐cardiac diagnosis, those not referred to in patient teams. SPC analysis shows early changes beyond previous variation (see fig). SPC demonstrates that the total number of chest pains and those of non‐cardiac origin have shown an increase outside normal variation, whereas those with cardiac pain have only shown changes compatible with background variation with time.

figure emAbsJuly07.f3
Abstract 021 Patients attending Heart of England Trust with chest pain diagnosed as non‐cardiac.

DiscussionThe campaign seems to have had minimal effect on increasing the uptake of those with cardiac symptoms at the expense of many more patients presenting to the emergency department. Further analysis will be taken throughout the year‐long campaign.

022 REPS (a registry of emergency procedural sedation) early data

K. G. Jacques1, A. Dewar2, D. Kerslake1, A. Gray1, D. Beard3. 1Royal Infirmary of Edinburgh, Edinburgh, UK; 2University of Edinburgh, Edinburgh, UK; 3Scottish Trauma Action Group, UK

AimsTo study all emergency department (ED) patients requiring procedural sedation in Edinburgh. To examine their characteristics, management, and outcomes including complications.

MethodsConsecutive ED patients requiring procedural sedation were prospectively enrolled onto a registry of emergency procedural sedation (REPS). Data collected included: patient and procedural characteristics; process times; physiological parameters; agents; sedation depth; composition of staff and complications.

Results347 patients were identified during the first six months, 256 are currently available for analysis. Median patient age was 42 (range 13 to 97). Time of presentation peaked in the afternoon and evening. Shoulder fracture/dislocation 48.8% (n = 125) and ankle fracture/dislocation 22.2% (n = 31) were commonest indications for sedation. Other procedures requiring sedation included chest drain insertion, femoral line insertion, foreign body removal, and wound closure.The grade of the most senior doctor present included: consultant 21.5% (n = 55), SpR 52.3% (n = 134), clinical fellow (middle grade) 18.4% (n = 47), SHO 3.1% (n = 8), clinical assistant 3.1% (n = 8), FY2 1.6% (n = 4). 7 agents were used (nitrous oxide, fentanyl, ketamine, midazolam, morphine, propofol and sevofluorane). Nitrous oxide combined with morphine and midazolam 20.7% (n = 53) and fentanyl with propofol 14.5% (n = 37) were the most common combinations. Depth of sedation was graded by ASA Scoring. Most patients were graded as either 3; 42.6% (n = 109) or 4; 29.7% (n = 76). Complications occurred in 2.7% (n = 7) of cases: desaturation 1.6% (n = 4), agitation 0.8% (n = 2) and apnoea 0.4% (n = 1). 4 of these patients received reversal agents, flumazenil or naloxone. Naloxone or flumazenil was given in a further five patients.

ConclusionAudit of practice is essential to ensure continued patient safety and to drive quality improvement. Procedural sedation is a core element of emergency medicine. REPS is the first comprehensive prospective study in Europe of patients requiring procedural sedation in the ED. Early results suggest that procedural sedation by emergency physicians is safe; however, complications can arise.

Free Paper Session 4: Pre‐hospital and Triage Thursday 27 September 2007, 09:00–10:00, North Hall, Sheffield City Hall

023 Safety of paramedics with extended skills

S. Mason, E. Knowles, J. Freeman. Health Services Research Section, ScHARR, University of Sheffield, UK

BackgroundThe UK Department of Health's strategy has been to encourage the increased use of non‐medical staff to carry out assessments and treatments traditionally carried out by doctors. A UK service led by paramedic practitioners (PP) with extended skills aimed to manage minor acute illness and injury among older people in the home thereby avoiding transfer to the emergency department (ED).

ObjectivesTo evaluate the safety of clinical decisions made by the new service.

MethodsConducted as part of a cluster randomised controlled trial, patients aged >60 years contacting the ambulance service with a minor injury or illness were included in the study. The new PP intervention was compared with standard practice of ambulance transfer and treatment in the ED. Outcomes included mortality at 28 days, and unplanned ED attendance within 7 days of the initial episode. Clinical records were rated independently by two ED clinicians to identify related episodes, avoidable subsequent episodes and failures of care.

ResultsOf the 2010 patients included in this analysis, 219 (10.9%) went on to have a unplanned ED attendance within 7 days. Of these, 162 (74.0%) presented with a condition related to their initial episode. The independent raters agreed on failures of care 83.4% of the time. There were 16 agreed failures in care (0.80%). 60 (3.0%) patients died within 28 days of their initial episode. There were no significant differences found between the intervention and control groups in relation to presentation at hospital within 7 days, rates of assessed failure of care, or mortality.

ConclusionsThis study suggests that appropriately trained paramedics with extended skills treating older people with minor acute conditions in the community are as safe as standard ambulance transfer and treatment within the ED.

024 Assessment tool for patients with mental health needs in the emergency department improves risk identification and documentation

C. Vorwerk, F. Davies. Leicester Royal Infirmary, Leicester, UK

BackgroundMental health related attendances to emergency departments (EDs) are a common problem. In 2004 the National Institute for Health and Clinical Excellence (NICE) published a guideline on self‐harm, highlighting the importance of patients with mental health problems requiring a specific assessment on arrival in the ED to identify those at risk. To address these issues a tailored initial assessment form for use by nurses or doctors was implemented in our ED. Training was provided over a four‐month period, with reiteration and the introduction of a coloured assessment form following the first evaluation.

ObjectivesTo determine whether the use of the tailored initial assessment form in patients with mental health problems has improved assessment and documentation of patient's description, suicide risk, security risk and competence to take own decisions, as per NICE guidance.

MethodsRetrospective study of ED patients with the diagnosis of self‐harm or suicidal intent/thought. Data were collected from April 2005 (before implementation), October 2006 (after implementation) and March 2007 (after reiteration and coloured assessment form). The main outcome measures were documentation of patient's description, competence, security and suicide risk.

Results245 patients were identified, 214 case notes could be retrieved. Of those there were 57 seen in April 2005, 70 in October 2006 and 87 in March 2007, with documentation of patient's description (21% vs 57% vs 78%), security risk (4% vs 46% vs 75%), competence (4% vs 51% vs 74%) and suicide risk (32% vs 64% vs 71%).

ConclusionOur findings demonstrate that the use of a tailored initial assessment form for ED patients with mental health needs, with reinforced training and introduction of a coloured assessment form, greatly improves documentation of risk assessment and competence in these patients.

figure emAbsJuly07.f4
Abstract 024 Documentation of assessment of patients with mental health needs in the emergency department.

025 The triage trainer: serious games for major incidents

S. D. Carley1, K. Mackway‐Jones1, B. Tregunna2, S. Jarvis2, BlitzGames. 1Manchester Royal Infirmary; 2VEGA, Emergency Department, Manchester Royal Infirmary, Oxford Road, Manchester, UK

BackgroundSerious Games is an emerging field that uses video games technology and techniques to support education and training. A Department of Trade and Industry‐funded collaboration between Blitz Games, VEGA Group, several UK universities, and subject matter experts has developed a high fidelity game that teaches and tests learners in major incident triage. This first project is a proof of concept in the feasibility of meshing instructional and games design and developing an engaging games application to secure robust learning objectives in health care.

ObjectivesTo develop a triage training package, which includes a video game, that can teach, test and assess learners in primary major incident triage, as well as offering a virtual engaging challenge for CPD.

MethodsTriage Trainer has been developed using the XBOX 360 and will be delivered over a variety of platforms. Learners are able to enter a virtual realistic environment that simulates the scene of a major incident, with highly realistic, emotional human casualties. Learners are able to “play” in teaching or testing modes and can repeat or alter the game play sequence. Game play allows learners to move around the scene independently and to make decisions in any sequence. The program can determine the success and failure of the learner, compared to an optimal time and sequence. The results can be used by a trainer or for self‐assessment.

ResultsThe triage trainer represents the first high fidelity, believable serious games in the emergency medicine field, which ensures training objectives are embedded and tracked for assessment. Further work on evaluation is currently in progress, but this initial achievement represents a major step in embedding validated training protocols for major incident and emergency medicine education, while maintaining the engagement of an entertainment game. If selected for presentation we will be able to demonstrate the trainer using video projection.

ConclusionsMajor incident training (and emergency training in general) is challenging owing to the difficulty in creating high fidelity simulated environments. We believe that serious games are a major next step in the field.

026 Paramedic practitioners for older people: are they cost effective?

S. Dixon1, S. Mason1, E. Knowles1, B. Colwell1, J. Wardrope2, R. Gorringe3, H. Snooks4, J. Perrin2, J. Nicholl1. 1University of Sheffield; 2Sheffield Teaching Hospitals Trust; 3Yorkshire Ambulance Service Trust; 4Swansea University, UK

IntroductionThe use of paramedic practitioners (PP) for responding to selected 999 calls relating to elderly patients with minor acute conditions could have several important effects on costs and outcomes. As part of a randomised controlled trial (RCT) of the PP service an economic evaluation was undertaken.

MethodsCosts were estimated for paramedic practitioner responses, other responders, emergency department (ED) attendances, admissions, social services/integrated care assessments, primary and community care costs and nursing/residential care costs. Mean resource use and costs were compared between the trial study groups. Sensitivity analyses focussed on the unit costs of the paramedic practitioners, as there are uncertainties over salaries and the impact of implementing a comprehensive scheme.

ResultsCosts based on routine data show that the differences in pre‐hospital response between the two trial arms largely cancel each other out. However, the lower same day ED costs (−£32) and inpatients costs (−£492) in the PP group are not fully offset by the identified increases in costs in the following 28 days (+£0, +£6 and +£372, for assessments, subsequent ambulances, ED attendances and inpatient stays). When primary/community care and nursing/residential care are examined, the results change markedly, producing a much greater cost reduction associated with the paramedic practitioners. Increased salaries and/or reduced vehicle utilisation reduce the cost difference between the groups.

ConclusionsThe study shows statistically significant changes in the use of NHS resources when paramedic practitioners are used. More specialist PP resources are required, as too are more social services/integrated care assessments and secondary care in the days following the initial 999 call. These are offset, in cost terms, by reduced use of other ambulances for the initial call, reduced ED and inpatient resources. The net effect of these changes are lower costs of £150 per patient.

Free Paper Session 5: Information and Communication Thursday 27 September 2007, 10:00–11:00, North Hall, Sheffield City Hall

027 The patient, the doctor and the emergency department: a longitudinal study of patient‐centredness in the consultation

J. Dale1, H. Sandhu1, R. Lall1, E. Glucksman2. 1University of Warwick; 2King's College Hospital, London, UK

IntroductionAccident and Emergency (A&E) departments in the UK have undergone considerable changes over the past 10 years. The impact of structural, workforce and educational changes on the doctor‐patient consultation within A&E is unknown.

AimsTo explore how doctor‐patient communication in A&E primary care consultations has changed between 1991 and 2005, comparing communication skills of GPs, SHOs and Specialist registrars/staff grades.

Method430 videotaped A&E primary care consultations (180 (42%) recorded in 1991 and 250 (58%) in 2005) from 20 GPs (1991 n = 8, 2005 n = 12), 17 SHOs (1991 n = 7, 2005 n = 10) and 11 registrars (1991 n = 4, 2005 n = 7) all working in the same inner city A&E department were analysed using the Roter Interaction Analysis System (RIAS) allowing in‐depth coding of communication skills. Skewed data were log transformed and subjected to analysis of covariance adjusting for doctor and patient gender, case mix, length of consultation and age of patient.

ResultsOn average consultation lengths were 251 seconds (95% CI 185 to 316) longer in 2005 than in 1991. The difference was especially marked for senior house officers (mean duration 385 seconds in 1991 and 778 seconds in 2005; 95% CI 286 to 518). All doctor groups showed increased patient centred communication related to building a relationship and activating and partnering (involvement of patient during the consultation). While senior house officers demonstrated a greatly increased focus on data gathering, general practitioners substantially increased the amount of talk centred on patient education and counselling, with an emphasis on giving information related to the medical condition.

ConclusionBetween 1991 and 2005, patient‐centredness increased for all three doctor groups, together with greater consultation length. The limited emphasis given to information and advice providing by emergency department doctors suggests a need to further develop this aspect of communication if the quality of the consultation is to be further enhanced.

028 Job satisfaction and work‐related stress in emergency medicine specialist registrars: a focus group study

S‐A. Wilson. Bradford Royal Infirmary, Bradford, UK

ObjectivesTo explore the aspects of emergency medicine (EM) that give Specialist Registrars job satisfaction and those that cause work‐related stress.

DesignA qualitative study using focus groups as the primary source of data. A grounded theory approach using the methods described by Strauss and Corbin, was used for the analysis of the data.

Participants12 Specialist Registrars (SpR) on the Yorkshire Deanery's EM rotation.

ResultsThree main categories emerged from the data: time pressure, variety, and “making a difference”. Both job satisfaction and work related stress were found to be derived from the same phenomenon. The core category “variety with urgency” was derived, with the dimensional ranges being the experience of the SpR, the patients' presentation, and the interaction with the specialities.

ConclusionsDoctors contemplating a career in emergency medicine should be aware of the causes of job satisfaction and work‐related stress. Identifying causes of stress in this group of doctor will enable the strategies to be developed to minimise this stress.

029 Why do patients with minor conditions attend the Emergency department?

R. Penson, P. Coleman, S. Mason, J. Nicholl. Health Services Research Section, ScHARR, University of Sheffield, UK

BackgroundCurrent UK health service configuration provides numerous alternatives to the emergency department (ED) for patients with acute minor healthcare complaints. A questionnaire survey and notes review regarding patients' reasons for attending was carried out in large urban UK ED. This repeat of a previous study carried out nine years ago.

ObjectivesTo discover whether reasons chosen to attend the ED with a minor injury have altered over the past nine years.

MethodsData collection times were randomised over a period of 32 days. 261 patients completed a repeat of a 1997 survey enquiring about reasons for attending the ED, knowledge of and reasons for not accessing other services. If patients consented, a notes review was carried out to gain an impression of their appropriateness and strength of reason for attending. Findings were compared with those from 1997.

ResultsParticipants attending with a weak or medium reason increased (6.8% vs 18.1%, p<0.005). There were increases in the participants who sought advice from their general practitioner services (GP) (13.4% vs 53.3%, p<0.001) or NHS Direct/Walk‐In Centres to attend (0% vs 29.8%, p<0.001). There was an increase in the participants who attended because their GP was unavailable (4.7% vs 16.9%, p<0.001). More participants expressed a desire to see a doctor as soon as possible (23.5% vs 33.9%, p<0.05), or get a second opinion (4.7% vs 10.3%, p<0.02). In 2006, up to 48.4% of participants were unsure of other services that were available. However, there was an increase in those that could have been seen by their out‐of‐hours service (9.8% vs 29.0%, p<0.001) and decrease in those that could be seen exclusively in the ED (44.5% vs 26.2%, p<0.001).

ConclusionThis study shows that patients continue to use the ED for non‐urgent conditions. Though patients are aware of the existence of alternative providers of care, they appear to be unsure which conditions are appropriate for each provider. A major new factor causing attendance for non‐urgent care in the UK appears to be difficulty in accessing GP services.

030 Patient‐centred communication: a comparison between Senior House Officers and Emergency Nurse Practitioners consulting in A&E

J. Dale1, H. Sandhu1, R. Lall1, R. Crouch3, E. Glucksman2. 1University of Warwick; 2King's College Hospital, London; 3Southampton University, UK

IntroductionDespite the widespread introduction of Emergency Nurse Practitioners (ENPs) to UK A&E departments in recent years, there is little evidence about the impact that this has had on quality of care. In particular, no published research has explored how the consultation skills of ENPs compare with those of junior doctors, and in particular the degree of patient‐centredness.

AimsTo explore and compare communication skills of ENPs and SHOs consulting with patients identified at triage as attending A&E with a primary care problem.

MethodPatients were randomly allocated to either an SHO or ENP according to time/day. Consultations (all conducted within the same consulting room) were videotaped and analysed using the Roter Interaction Analysis System (RIAS) to allow in‐depth coding of their content.

ResultsIn all, 133 consultations were analysed (87 (65%) undertaken by SHOs and 46 (35%) by ENP) from 10 SHOs and 6 ENPs. Length of consultation differed with SHOs having a mean length of 778 seconds and ENPs a mean of 596 seconds (p<0.05). SHOs and ENPs showed differences in the extent of rapport building, data gathering and information giving and counselling. ENPs' consultations were more focused on the patient's treatment needs. They asked more open questions related to the therapeutic regimen (p<0.01), and gave more advice and information regarding treatment plans as well as lifestyle and psychosocial issues (p<0.05). On the other hand, SHOs placed more emphasis on establishing rapport and involving the patient in the consultation. For example, they gave more agreement statements (p<0.05), checked for the patient's understanding (p<0.01), and asked for the patient's opinion more frequently (p<0.01) than ENPs. As a consequence, patients expressed more concerns and worries in their consultations (p<0.05).

ConclusionThe findings from this study indicate significant differences between SHO and ENP consultations which are likely to have an impact on decision making and clinical outcome.

Articles from Emergency Medicine Journal : EMJ are provided here courtesy of BMJ Publishing Group