Background: Cohort identification from electronic medical records (EMRs) using structured data (ICD10 codes, lab results) is highly inaccurate. Document classifiers using Natural Language processing are ran primarily by computer science researchers, and while it is a promising strategy, to date the accuracy is still lagging.

 

Objective: To assess a clinician-supervised artificial intelligence methodology (combining manual review and machine learning techniques) to identify a cohort of 5000 Pediatric Emergency Department (PED) visits of children evaluated for acute appendicitis (EAA) from a set of EMRs with the ambitious goal of a 95% sensitivity/recall (Se) and a 95% positive predictive value / precision (PPV).

 

Design/Methods: We evaluated PED documents of all visits between 2007 and 2015 in a 3-step process.

(1) We used structured data (age, ICD10 codes) and a set of regular expressions to preselect an initial EMRs dataset.

(2) One year of data was annotated by a clinician-researcher into two categories (documents addressing EAA vs non-EAA). We then used 70% of the documents to train a ‘bag of words’ classifier and 30% as a validation set to assess the classifier performance. Based on the classifier metrics/performance, we set two cutoffs: a low threshold (LT) cutoff where the model had a Se of 95% and a high threshold (HT) cutoff where the model had a PPV of 95%.

(3) The classifier was then applied to all unclassified data: the documents scoring above the HT cutoff were automatically classified as EAA and those scoring below the LT cutoff were automatically classified as non-EAA. The documents scoring between these two thresholds were then manually reviewed by the clinician researcher (thus assigning an accuracy of 100% to this middle subset). For quality assurance, we manually reviewed 10% of the documents classified as EAA and non-EAA respectively.

 

Results: Overall our search involved 12,302 documents. Over half of our final cohort (2570/4803) was identified automatically as the model annotated 6,691 documents. Manual review of the other 4,267 EMRs resulted in additional 2,233 EAA cases.

The whole method required a manual review of 5,611 documents (15 sessions of 2 hours each)

Among a sample of 480 EMRs classified as EAA, 22 were false positives (4.6%). Among a sample of 616 EMRs classified as not-EAA, 15 were false negatives (2.4%)

 

Conclusion: Clinicians driven use of Natural language processing is highly successful and accurate by generating a robust accurate training set for classifier, and assessing middle ground cases that account for poor accuracy of a fully automated system. This method could be used to help gathering cohorts from emergency electronic medical records with high accuracy and acceptable time of manual review. 

Purpose: to compare myocardial dysfunction in patients with sepsis and septic shock and to evaluate its prognostic value in these two subgroups of patients.

Methods: We included patients diagnosed with sepsis/septic shock and admitted in our ED-High Dependency Unit between August 2012 and February 2018, in whom an echocardiogram was performed within 24 hours from admission. We evaluated LV systolic function using Global Longitudinal Strain (GLS, > -14% diagnostic for LV systolic dysfunction) and RV systolic function with Tricuspidal Annular Posterior Systolic Excursion (TAPSE, <16 mm diagnostic for RV systolic dysfunction). We divided our population in 3 subgroups: patients with normal LV and RV systolic function (G1), those with either LV or RV systolic dysfunction (G2) and those with a biventricular systolic dysfunction (G3). Day-7 and Day-28 mortality were our end-points.

Results: we included 238 patients (mean age 73±15 years, male sex 58%, T1 SOFA score 6.0±2.9), 41% with septic shock, day-28 mortality rate 27%. Age was comparable between sepsis and septic shock patients (72±16 vs 74±14 years), while MAP (72±12 vs 83±14 mmHg) was lower and SOFA score (7.8±2.8 vs 4.7±2.2, p<0.001) was higher in patients with shock. Patients with sepsis and septic shock showed comparable LV (end-diastolic volume index 44±18 vs 42±17 ml; end-systolic volume index 23±14 vs 22±15 ml, p=NS) and RV dimensions (basal diameter 3.6±0.6 vs 3.6±0.8 cm; mid-ventricular diameter 2.5±0.7 vs 2.7±1.0 cm; base-apex distance 7.7±1.1 vs 7.7±1.0, p=NS), as well as systolic LV (LV EF 51±15 vs 51±16%; LV GLS -12±4 in both groups, p=NS) and RV function (TAPSE 1.9±0.5 vs 1.8±0.5, p=NS). A similar proportion of patients in both groups showed an impaired LV GLS (respectively 67 and 69%) and a reduced TAPSE (55 vs 62%, all p=NS). Overall, G1 included respectively 25% of patients with sepsis and 22% of patients with shock, G2 54 and 55% and G3 21 and 23%. In the whole study population, day-7 (2% in G1, 14%in G2 and 24% in G3, p=0.001) and day-28 mortality rate (11% in G1, 25% in G2 and 49% in G3, p<0.001) were significantly increased in subgroups with more severe cardiac dysfunction. In the sepsis patients subgroup we confirmed these findings (day-7 mortality rate respectively 0%, 15% and 24%, p=0.002; day-28 mortality rate 9%, 26% and 45%, p=0.004); in patients with septic shock, mortality increased non significantly by day-7 end point (4%, 13% and 23%, p=NS) while the increase was significant by day-28 mortality (14%, 23% and 55%, p=0.008). A Kaplan-Meyer survival analysis confirmed a significantly reduced survival in patients with mono- or biventricular functional impairment by day-7 and day-28 mortality in the whole study population (p=0.005 and p<0.001) and in sepsis patients (p=0.016 and 0.002), while in patients with shock mortality increased significantly only by day-28 end-point (p=0.008).

Conclusion: patients with sepsis or septic shock showed a similar prevalence and severity of myocardial dysfunction; in both subgroups, LV and RV systolic dysfunction were associated with an increased mortality.

Aim: to evaluate the prognostic value of early variation of lactate dosage in septic patients.

Methods: between November 2011 and December 2016, 263 patients were enrolled in a prospective analysis aiming to find reliable biomarkers for an early sepsis diagnosis. Patients admitted to our High-Dependency Unit from the Emergency Department (ED) with a diagnosis of severe sepsis/septic shock were eligible. We evaluated lactate dosage (LAC) at ED-admission (T0), after 2 hours (T2), 6 hours (T6) and 24 hours (T24) from the initial diagnosis. Lactate clearance was calculated at T2, T6 and T24 (T2: (LAC T2 – LAC T0/LAC T2)*100; T6: (LAC T6 – LAC T0/LAC T6)*100; T24: (LAC T24 – LAC T0/LAC T24)*100). Primary end-points were day-7 and day-28 mortality rate.

 

Results: Mean age of the study population was 74±14 years, 58% male gender; mean Sequential Organ Failure Assessment (SOFA) score at admission was 5.3±2.7. The most frequent infection source was respiratory (45%), followed by abdominal (17%) and urinary tract (14%).  Day-7 mortality was 16% and day-28 mortality was 25%. Lactate dosage was higher in non-survivors compared with survivors at day-28 at all evaluations (T0: 3.9±4.2 vs 2.7±2.5; T2: 3.6±4.0 vs 2.4±2.5; T6: 3.5±3.8 vs 1.8±1.4; T24: 4.1±5.6 vs 1.5±1.4, all p<0.05). Considering day-7 mortality, lactate dosage was significantly higher in non-survivors compared with survivors only at 6 and 24 hours (T6: 3.9±4.3 vs 1.9±1.7; T24: 4.8±6.3 vs 1.7±2.1, p<0.05). We dichotomized lactate dosage values (≤ o >2 meq/L). A dosage >2 meq/L was more frequent in non-survivors compared with day-7 survivors at T2 (60 vs 41%), T6 (58 vs 34%) and T24 (52 vs 22%). Considering day-28 mortality, a lactate dosage >2 meq/L was more frequent at T6 (49 vs 33%) and T24 (44 vs 21%, all p<0.05) in non-survivors. A survival analysis using the Kaplan-Meier method showed a decreased day-7 survival in patients with lactate >2 at T2 (77 vs 88%), T6 (76 vs 89%) and T24 (72 vs 91%) and a decreased day-28 survival in patients with lactate >2 at T6 (67 vs 80%) and T24 (59 vs 81%, all p<0.05). A lactate clearance <10% at T6 (54 vs 29%) and T24 (51 vs 32%, all p>0.05) was more frequent in day-28 non-survivors compared with survivors. Considering day-7 mortality, a lactate clearance <10% at T6 was significantly more frequent among non-survivors (55 vs 32%, p=0.021), and had a tendentially significant association with survival at T24 (56 vs 34%, p=0.05). A Kaplan-Meier survival analysis confirmed an association between lactate clearance <10% at T6 and T24 and decreased day-7 survival (respectively 77 vs 90% and 81 vs 91%) and 28-day survival (respectively 62 vs 82% and 67 vs 82%, all p<0.05).

Conclusions: a lactate dosage >2 meq/L taken as early as 2 hours from ED admission was associated with a short-term poor prognosis; a lactate clearance <10% was associated with an increased mortality only when evaluated at least 6 hours from ED admission.

Purpose: PCSK9 (proprotein convertase subtilisin/kexin type 9) plays a critical role in regulating circulating cholesterol levels, through the reduction of the membrane-associated low-density lipoprotein (LDL) receptor. In experimental models plasma PCSK9 levels were increased in sepsis. At normal levels, PCSK9 has no influence upon hepatocyte bacterial endotoxin clearance, but as levels rise, there is a progressive inhibition of clearance. Recently it has been shown that decreased function of PCSK9 increases survival of septic patients. Aim of this study was to assess prognostic value of an early assessment of PCSK9 levels in septic patients.

Methods: Between November 2011 and December 2016, 263 patients were enrolled in a prospective analysis aiming to find reliable biomarkers for an early sepsis diagnosis. Patients admitted to our High-Dependency Unit from the Emergency Department (ED) with a diagnosis of severe sepsis/septic shock were eligible. We evaluated vital signs and laboratory data at ED admission (T0), at 6 hours (T6) and after 24 hours (T24); Sequential Organ Failure Assessment (SOFA score) was calculated at every evaluation point. Primary end-points were 7-day and 28-day mortality.

Results: Mean age of the study population was 74±14 years, 58% male gender; mean SOFA score at admission was 5.3±2.7. The most frequent infection source was respiratory (45%). Day-28 mortality was 25%. PCSK9 normal values are lower than 250 ng/ml; in septic patients, at every evaluation point, PCSK9 level was significantly higher than normal range reported in previous studies (T0 661±405 ng/mL, T6 687±417 ng/mL, T24 718±430 ng/mL). There was no significant difference between patients with Gram+ or Gram- pathogen infection (T0: 641±493 ng/mL vs 701±406 ng/mL; T6: 652±433 ng/mL vs 769±389 ng/mL; T24: 690±397 ng/mL vs 811±501 ng/mL, all p=NS). Only at T0 non-survivors by day-28 showed a significantly lower level than survivors (549±437 ng/mL vs 696±390 ng/mL, p=0.016); all the other evaluations were comparable regardless of the outcome, both considering day-7 and day-28 mortality rate. An Analysis for Repeated Measures between T0 and T24 levels did not show any significantly different trend between day-7 (T0: 668±389 vs 623±492; T24: 702±388 vs 843±669 ng/mL) and day-28 (T0: 696±390 vs 549±437; T24: 718±392 vs 719±553 ng/mL) survivors and non-survivors (all p=NS). There was no correlation between SOFA score values and PCSK9 levels at all evaluation (non-parametric correlations: all p=NS). We divided the study populations in two subgroups according to the level of T0 SOFA score (≤ and >5): PCSK9 levels were comparable regardless the severity of sepsis-induced organ damage (T0: 696±381 vs 614±444; T6: 716±417 vs 631±421; T24: 709±413 vs 716±470 ng/mL, all p=NS). Finally we compared PCSK9 levels in patients with T0 lactate level ≤ and >2: even in this analysis we did not find any significant difference (T0: 672±485 vs 642±365; T6: 689±449 vs 664±373; T24: 711±392 vs 698±452 ng/mL, all p=NS).

Conclusions: PCSK9 levels were significantly increased in septic patients. However, the levels did not show any significant association with indexes of hypoperfusion and with the severity of organ damage, as well as with the short- and medium-term mortality rate.

Introduction:The 2015sepsis definitions suggest using the quick SOFA (qSOFA) score for risk stratification of sepsis patients. Our aim was to compare it to our local score of triage in order to predict mortality.

Methods:This was a retrospective cohort study based on data from our local sepsis registry.We retrospectively calculated the qSOFA ,our triage score for all patients admitted in our center with the diagnosis of sepsis. Statistical analysis was performed using SPSS 22.

Results:268 patients were included in the study. The mean age was 63,6 ± 16,8 years with a sex ratio of 1.3 (56,7% of patients were men). 42,5% were diabetic and 45,9% were hypertensive.Based on ourresults we have shown thatqSOFAhas a poor performance in predicting mortality;our triage score was more sensitive and specific in predicting mortality of such patients when compared to q SOFA score.

Conclusion: In conclusion, we find that in our settings, our triage score was better than q SOFA in predicting mortality. Further studies are needed to reevaluate theqSOFA score.

 

The incidence of VTE ranges from 75 - 265 per 100,000 population; Pulmonary Embolism (PE) accounts for 40% of these events with high morbidity and mortality. Different scoring systems are available to predict the presence of PE; Well's and Revised Geneva scores are most widely used. Scoring system catagorise the patients into Low; Moderate or High risk groups. D-Dimer is non-specific for PE, its role is important in low-risk patients, where negative results can rule out PE. CT and V/Q scans are available imaging modalities to diagnose PE. CT pulmonary angiogram is more preferred, as it is more quicker, accurate and gives alternate diagnosis (if any). Both these imaging techniques are associated with exposure of the patient to high radiation dose.

An audit was performed in our department to evluate whether the protocols for diagnosing PE are being adhered to in terme of scoring systems, D-Dimers and imaging and to determine if we were over doing the CTPA. Different co-relates, e.g. age, sex, risk factors, Wells score, presenting symptoms and investigations were studied and their association with PE was determined. The study included 119 consecutive patients, who underwent CTPA in 3 months time. 

The mean age was 62.7 years, with 43% male and 57% female patients. Mean age of PE positive patients was 66.7 years. The most common presentation was shortness of breath, followed by chest pain. A rare prrsentation was abdominal pain. 50% of the patients had no risk factor and 35% had single risk factor. Immobility was the commonest risk factor. Intrestingly, 26.5% of the patients were smokers. 61% of the intermediate probability patients were positive for PE, whereas 31% were positive from the low probability group. D-Dimer was positive 41% in and negative in 54% of patients. 2 patients with negative D-dimers were positive CTPA. 

CTPA was diagnostic for PE in 31% and negative for 65% of the patients. In the rest it was inconclusive. Chest infection was the most common alternative diagnosis.

We concluded that if the protocols were adhered to strictly, we could have avoided CTPA in 26 of the patients. 

Background

Chronic obstructive pulmonary disease (COPD) accounts for approximately 10% of medical hospital admissions in the UK and the number of these admissions is rising.  As such - the accurate management of acute exacerbations of COPD (AECOPD) in the emergency department (ED), in order to improve outcomes such as safe discharge, length of inpatient stay, patient morbidity/mortality and cost effective treatment, is increasingly important.  The use of care 'bundles' as standardised, evidence-based treatment guides has been shown to improve the management of these patients in EDs. 

 

Aim

To improve the management of AECOPD in the Queen Elizabeth University Hospital (QEUH, Glasgow) ED with structured implementation of a COPD treatment pathway and supporting departmental education. 

 

Methods

Data from a random sample of AECOPD cases presenting the QEUH ED (a large, inner city, university hospital department) between 01/07/17 and 31/03/18 were analysed.  A standard of care for AECOPD was developed with reference to national COPD treatment guidelines and consensus of emergency medicine (EM) consultants within our ED.  This standard was comprised of nine key performance indicators (KPIs) and allowed ‘percentage treatment success’ (PTS) - in terms of adherence to those indicators - to be calculated on a case by case and also mean weekly basis.  Five cases per week were randomly selected and assessed retrospectively for PTS. 

Intervention with a COPD treatment pro-forma and a structured program of related departmental education was implemented during the data collection period and subsequent response of the PTS was monitored. 

 

Results

A total of 120 cases were analysed.  The total mean PTS was 63.88%.  Mean PTS pre-intervention was 59.71% whilst total mean post-intervention PTS was 65.96%.  The PTS following each intervention was calculated.  After COPD pro-forma introduction and informal discussion PTS was 60.73%.  Following presentation at an ED ‘show and tell’ session (a junior staff educational meeting) PTS rose to 67.99%.  After departmental posters were displayed and two teaching sessions regarding COPD management and pro-forma use were delivered the PTS was 66.66%.

In total the number of pro-formas used in the analysed cases was seven (5.83% of the total number of cases) however the average PTS of cases where the pro-forma was used was 87.29% in comparison to 63.75% were it wasn’t.  4 of the 7 pro-formas analysed were used following departmental teaching sessions – none were found to have been used prior to formal discussion of the project in ED. 

 

Discussion/Conclusion

Given the volume and impact of AECOPD cases on national emergency departments, measures to improve efficiency and accuracy in treating them are of increasing importance. While uptake of the pro-forma was limited, particularly amongst senior staff, compliance with the pro-forma and PTS improved following interventions.   

This project has shown the use of a structured pro-forma can improve the accuracy of COPD treatment.  It also suggested that formal presentation of such measures within a department as well as targeted teaching sessions to relevant staff is vital in gaining clinician engagement and ultimately improving treatment.

 

 

Background: Treatment of opioid use disorder with buprenorphine has expanded significantly, with 58% opioid treatment programs now offering buprenorphine. Additionally, 2.1 million ambulatory visits reported uptake of buprenorphine in 2013. Buprenorphine films, released for use in October 2010, with a single dose foil packaging are considered child-resistant and abuse deterrent. The objective of this study is to evaluate the trends, and characteristics of exposures to buprenorphine film formulations.

Methods: We retrospectively queried the National Poison Data System (NPDS) for all confirmed exposures to buprenorphine films from 1/1/2012 to 12/31/2016 as specified by the American Association of Poison Control Center Code (AAPCC) generic code and product name. We also assessed the distributions of several key characteristics of the exposures, including demographic characteristics, reason of exposure, clinical effects, medical outcomes, and therapies. We generated descriptive statistics after having segmented the relevant characteristics of exposures into appropriate categories. Frequencies and rates of buprenorphine film exposures (per 100,000 human exposures) were evaluated using Poisson regression methods, with the percent changes and corresponding 95% Confidence Intervals (95% CI) reported.

Results: Overall, there were 6,205 reports of exposures to buprenorphine sublingual films to the PCs during the study period. The reports of buprenorphine film exposures increased from 852 to 1,425 during the study period. Children under 6 years of age represented 26.0% of the sample, while adults between 20 and 39 years of age accounted for 42.1% of the cases. The most common reason for exposure was unintentional (35.8%), with intentional abuse (21.4%) and suspected suicides being common (15.1%). Single substance exposures accounted for 61.6% of the cases and ingestion was the most common route of exposure. In 57.6% of cases, the patient was enroute to a healthcare facility. The case fatality rate for such exposures was 0.3%, with 4.2% cases demonstrating major effects. Among children under 6 years of age, majority were single substance (96%), and unintentional exposures (98%), with resulting minor clinical effects (36.4%). Overall, multiple substance exposures resulted in a higher number of deaths (16 vs 2 cases) and major clinical effects (193 vs 65 cases). Similarly, the proportion of major effects was highest among suspected suicides (11.4%) and abuse (5.3%) in comparison to cases of other exposure reasons such as unintentional. New Mexico (32.8) demonstrated the highest rate of buprenorphine film exposures (per 100,000 population). Naloxone therapy was reported for 18.4% cases. The frequency of buprenorphine film exposures increased by 67.2% (95% CI: 42.3%, 81.7%; p<0.001) over the study period, and the rate of such exposures increased by 75.9% (95% CI: 59.8%, 93.4%; p<0.0001). 

Conclusions: Analysis of national data from the NPDS exhibited a significantly increasing trend in the exposures to buprenorphine film products, with such exposures being frequent among children under 5 years of age. Considering the complexity of film packaging, it is imperative to explore in greater detail, the reasons for the observed rise in these exposures.  There were fewer major outcomes in unintentional overdoses, compared to cases of suicidal ingestion and abuse, especially when used with other substances. 

n/a

Background:

Urinary tract infection (UTI) is a common infection in the Emergency Department. The diagnosis requires the presence of symptoms alongside bacteriuria. The presence of bacteria in the urine without symptoms is known as asymptomatic bacteriuria (ABU), which is common with increasing age, and has no significance outside pregnancy.

The diagnosis of UTI can be difficult, particularly in the elderly. Over diagnosis of UTI’s can lead to unnecessary tests and inappropriate use of antibiotics and antibiotic resistance, at significant cost.

The Royal College of Pathologists Australasia (RCPA) guidance suggests urine microscopy, culture and sensitivity (MCS) is only indicated in those with: Dysuria/frequency, cystitis, UTI, screening in pregnancy, post-partum fever, epididymo-orchitis, unexplained fever and renal involvement of systemic disease.

It is not clear to what extent this guidance is followed in clinical practice, and this study aimed to assess this, alongside estimating costs of unnecessary processing of samples.

Methods:

All urine MCS sent over a 1 month period (Sept 2016) in the ED of our hospital (QEII, Brisbane, Australia) were identified. A chart review of all patients was performed, and relevant data was collected, including: indication, dipstick test results, culture results, antibiotic use, final diagnosis, and demographic data. A comparison was made to the RCPA guidance, and statistical analysis was performed to look at the performance of dipstick in predicting culture results, and a simple cost analysis was performed to estimate costs of over-diagnosis. 

Results:

A total of 536 urine MCS were sent over a 1 month period (out of 4680 total ED presentations) and of these, 416 charts were used for the results. The figure difference is explained by repeat urine MCS samples sent for the same patient, and an inability to access charts. 

31% of urine samples sent had a positive urine culture, and 37% of samples sent were contaminated. 

30% of the urine samples sent had negative nitrites and leukocytes on urine dipstick. 

Nitrites on urine dipstick were found to have a specificity of 92.31%, and leukocytes in all samples sent had a negative predictive value of 87.5%. 

With a urine sample costing 22.60 AUD, the monthly cost of processing urine samples came to 12,113.60 AUD alone, before costs associated with laboratory staff, and of over-diagnosis and over-treatment. 

Discussion and Conclusions:

Urine MCS is an undoubtably necessary tool in day-to-day practice in ED, with results often effectively guiding treatment for practitioners. 

The following key findings can be taken from this audit:

We need to educate patients prior to MSU collection considering a 37% contamination result. 

Almost 2/3 of samples sent were not in line with RCPA recommendations suggesting further education of staff is required. 

It has also been shown that we do not trust our urine dipstick- with 30% being sent despite completely negative dipstick results. 

Nitrites are a good predictor of infection when positive, and leukocytes do well to predict absence of infection when negative. 

After changes have been implemented it will be necessary to re-audit this process to see how education has changed our practice. 

Background

In 2016, a new clinical concept termed ‘Quick Sepsis-Related Organ Failure Assessment’ (qSOFA) was introduced to identify high-risk patients with suspected infection outside of intensive care settings. This is a part of Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Systemic Inflammatory Response Syndrome (SIRS), the previous criteria (Sepsis 2) were de-emphasised from the current sepsis definition. In the present study, we aim to validate qSOFA in an emergency department in Hong Kong. The prognostic value of qSOFA was compared with SIRS as well as another commonly used early warning score, the National Early Warning Score (NEWS).

 

Methods

This was a single-centre, prospective study conducted in the Emergency Department (ED) of Prince of Wales Hospital, Hong Kong between July 2016 and June 2017. The ED in PWH receives over 144,000 new patients per annum and admits 30% of those attending. We have recruited 1,253 patients presenting to the ED triaged as category 2 (Emergency) and 3 (Urgent). All variables for calculating qSOFA, SIRS and NEWS were collected. The primary outcome measure was 30-day mortality. The prognostic value of qSOFA, SIRS and NEWS to predict 30-day mortality was compared. Venous lactate was also measured to investigate whether lactate level provide additional value for the prediction of 30-day mortality. Receiver Operating Characteristic analysis was performed to determine the Area Under the Curve (AUC), sensitivity, specificity, positive and negative predictive value, positive and negative likelihood ratio for qSOFA≥2, SIRS≥2 and NEWS>5.

 

Results

Of 1,253 patients recruited, median age was 72 years (IQR: 59-84); 638 (50.9%) were male. Overall 30-day mortality was 5.7%. The prognostic value for prediction of 30-day mortality, with AUC of for qSOFA≥2, SIRS≥2 and NEWS≥5 were 0.56 (95%CI 0.53-0.58), 0.61 (95%CI 0.58-0.64) and 0.61 (95%CI 0.58-0.64) respectively. Using pairwise comparison of ROC curves, NEWS≥5 was better at predicting 30-day mortality in ED patients (p=0.036). The AUC of lactate level≥2 mmol/l of predicting 30-day mortality was 0.64 (95% CI 0.61-0.66). The combination of lactate level≥2 mmol/l with qSOFA≥2, SIRS≥2 and NEWS≥5, AUC were 0.54 (95%CI 0.51-0.56), 0.61 (95%CI 0.58-0.63) and 0.59 (95%CI 0.56-0.61) respectively. In addition, positive likelihood ratio of qSOFA≥2, SIRS≥2 and NEWS≥5 to predict 30-day mortality were 13.66 (95%CI 5.57-33.48), 2.00 (95%CI 1.49-2.69) and 2.71 (95%CI 1.90-3.86).

 

Discussion & Conclusion

Among emergency and urgent patients presenting to the ED, the prognostic value for using NEWS was greater than qSOFA, while there was no difference between qSOFA and SIRS. Combinations of lactate level with qSOFA, SIRS or NEWS did not improve the prognostic value in predicting 30-day mortality for ED patients.

Trial Registration: The study is registered with ClinicalTrials.gov (NCT02817581). Funding: This study did not receive any specific funding Ethical approval and informed consent: Ethics approval was obtained from Institutional Review Board of the Chinese University of Hong Kong to conduct this prospective study (CREC 2015.624). Written consent was obtained either from the patient or from a relative in all cases

Dear SIRS, we are sorry to say you are ousted but NEWS is that we have made an IMPACT!

 AIM:

To increase awareness and recognition about sepsis.

 

INTRODUCTION

Sepsis is the leading cause of death and is responsible for nearly 44,000 deaths in the UK alone. Therefore, improving compliance of sepsis screening and delivery of sepsis six within one hour was one of our aims in this quality improvement project. Currently, in most hospitals in Wales, NEWS of 3 is being used as an early warning system to identify acutely unwell patients who could potentially be unwell with sepsis. There is evidence to show that NEWS of 6 is associated with more likely rapid deterioration if they are septic. Therefore, we wanted to show improvement in this group of patients.

METHODS

This study was conducted in a single centre University Teaching Hospital Emergency Department that sees approximately 100,000 patients annually. SeGMED (Sepsis Group Morriston Emergency Department) was formed in 2016. Our mission statement was, “To inspire, educate and promote early recognition, treatment and management of sepsis”. The group incorporates receptionists, a data analyst, Health Care Assistants, Staff nurses, Advanced Nurse Practitioners, junior doctors, a registrar and a consultant. Sepsis has now become a part of the departmental mandatory teaching. We have developed a new sepsis screening tool, created a staff notice board with weekly updates, compliance charts, star of the week and pictures of the SeGMED group for staff to communicate. Since the groups’ creation in 2016, we have done four sepsis awareness days. We had prominent guest speakers and patients who have had their lives affected by sepsis come and narrate their side of the story. The whole process has helped us improve our screening tool and also improve compliance. 

RESULTS

The results are from patients registered between June 2017 to February 2018, a total of 9 months. Our data showed steady progress. There was a dramatic increase in screening tools being commenced from 305 per month in June 2017 to nearly 700 a month in December 2017. For patients with NEWS ≥ 6 there was a 27% increase in patients receiving all elements of sepsis six. Focussing on the number of septic patients who were given antibiotics within the hour, there was an increase from 35% in June 2017 to 40% in January 2018. In the group with NEWS >6, our 1-hour antibiotic compliance rose from 42% in June 2017 to 47% in February 2018.

 

 

 CONCLUSION

For nearly 80% of the patient’s sepsis journey begins outside the hospital. Whenever a patient or relative comes to us and asks if one of their relatives or friends could have sepsis because of the visual prompts and awareness around sepsis, we know we have made an impact. Our results are proof that our busy staffs are more aware of sepsis and conscientious of commencing and completing screening forms for one main cause – TO SAVE SOMEONE’S LIFE.

Background

Alcohol related attendances are a growing and potentially preventable burden for the NHS. Between 2008/9 and 2013/14 there was a 104.6% increase in alcohol related attendances at Emergency Departments (ED), with a 53.9% increase in associated emergency admissions. Currently, there is limited understanding about the proportion of alcohol related attendances that may not require emergency attention and could be diverted into specialist services away from the ED.

We conducted a retrospective multi-site analysis of routinely available data to gain an understanding about the pathways of care through the ED of acute alcohol related attendances and their outcomes.

Methods

Routine NHS patient level data for adult ED attendances across 18 EDs in Yorkshire and Humber for a one year period from January 2014 to December 2014 were analysed.

Alcohol related attendances were identified using a staged process: (1) the first diagnosis variable was searched for a comprehensive list of all alcohol related terms. Those patients with an alcohol related term then had their presenting complaint variable searched; (2) identified terms were used to search all nine diagnosis variables and presenting complaint variable; (3) “acute alcohol” related attendances were separated from “chronic alcohol” related attendances and the latter cases were returned to the “all other attendances” group.  

Age, mode and time of arrival, number of investigations and treatments, length of ED stay and ED outcome were analysed. We applied an adapted version of a previously published process based definition of “low acuity attendance” to identify the proportion of acute alcohol related attendances that could have been seen and treated in an alternative primary or urgent care setting.

Results

Of the 1,312,539 ED attendances, 1.5% (n=20,052) were related to acute alcohol intoxication. The odds of arriving by ambulance in the “acute alcohol” group were significantly higher than the “all other attendances” group (OR 7.44, 95% CI 7.2-7.7). Peaks in attendance for the “acute alcohol” group were during the early hours of Saturday and Sunday morning, whereas peaks in attendance for the “all other attendances” group were around midday across all days of the week. Once in attendance the “acute alcohol” group had a significantly longer length of stay (Median: acute alcohol = 184 minutes vs all other attendances = 139 minutes, p<0.001) and were more likely to leave before treatment / refuse treatment (OR 3.27, 95% CI 3.14-3.42). Of those in the “alcohol group” who did complete their care pathway the majority were discharged (n=11,079; 55.3%) but around a third were admitted (n=6,113; 30.5%). 30% of patients in the “acute alcohol” group could have been seen and treated in alternative healthcare setting.

Discussion/Conclusions

Our findings show that acute alcohol related attendances are placing a small but potentially preventable burden on emergency care services. The management of patients presenting with acute alcohol intoxication is resource intensive, both within the hospital and for ambulance and police services. Therefore, the development of effective interventions prior to ambulance calls would allow for better management elsewhere, thereby reducing pressures on emergency services.

Funding: The research was funded by the National Institute for Health Research (NIHR) Collaboration and Leadership in Applied Health Research and Care, Yorkshire and Humber (CLAHRC YH): Avoiding Attendances and Admissions in Long Term Conditions Theme (AAA). The views expressed are those of the authors, and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. Ethics: A UK National Research Ethics committee granted ethical approval for the data analysis (REC ref: 14/YH/1139) and Confidentiality Advisory Group (CAG) approval was also obtained (CAG ref: 4/CAG/1015).

A retrospective cohort study of venous thromboembolism risk in patients treated with a removable splint for acute knee injury.

Dr Ashleigh Philips MBChB DRCOG; Dr Jamie Cooper MBChB FRCEM

Introduction

Temporary lower limb immobilisation in a cast following injury is a risk factor for developing venous thromboembolism (VTE) and United Kingdom guidance recommends prophylaxis with low molecular weight heparin (LMWH) to high risk patients. However, the current evidence only pertains to patients managed in rigid circumferential casts.  Patients with knee injuries are of often managed with a removable knee splints in the initial period but evidence is lacking regarding the VTE risk they confer and, consequently, no guidance on how to approach patients from a VTE perspective is available.

Aim

To evaluate the incidence of symptomatic VTE in ambulatory adult patients with acute knee injuries who are managed in a removable knee splint.

Methods

This was a retrospective cohort study performed in the Emergency Department (ED) of Aberdeen Royal Infirmary (ARI). Adult patients (≥14 years) with an acute knee injury, discharged from the ED to a specialist Orthopaedic clinic for follow up between 1^st January 2010 and 31^st December 2012 (before standardised VTE tools were used in this population at ARI) were identified from computer records. Both ED and Orthopaedic records were scrutinised. Patient exclusions included: unavailable notes; treatment commenced elsewhere; non-traumatic pathology; postoperative patients; patients not immobilised in a removable knee splint; patients whom 3-month post immobilisation follow up was not reliable.

Population demographics and evidence for confirmed symptomatic VTE up to 3 months post-immobilisation was extracted using electronic patient records and descriptive statistics were calculated.

Results

In total, 1110 patients were referred to the Orthopaedic knee clinic. Following exclusions, a study population of 566 was achieved. Male patients accounted for 62.5%, ages ranged from 14-90 years (mean 32 years), 445 patients (79%) suffered soft tissue injuries and 84% were immobilised in a removable splint for

Prophylactic LMWH was not prescribed for any patient during the study and only one patient (0.17%) was diagnosed with a symptomatic VTE; a 36 year old male with no VTE risk factors who developed bilateral pulmonary emboli within the 3 months post immobilisation. There were no deaths during the study period. 

Discussion

Even without structured risk assessment and no prophylaxis administered, the incidence of VTE in this population is small (0.17%) compared with an estimated 2.1% for casted patients.  This is likely due to the young age of the population and the removable nature of the splints. Major risk factors were identified in 34 patients within the study population who would qualify for LMWH prophylaxis if current guidance for casted patients had been followed. The patient sustaining VTE did not have any risk factors, therefore wouldn’t have had prophylaxis in any case.

Conclusion

These results should help inform decisions about LMWH prophylaxis in patients immobilised in knee splints. We would recommend risk stratification in these patients but raising the threshold for prophylaxis compared with casted patients.

Background:

Venous cannulation is one of the most commonly performed invasive procedures in the Emergency Department. Many of the patients coming to ED, despite severity, will require a vein cannulation to receive intravenous fluids or medication. Difficult peripheral intravenous cannulation is the situation arisen when at least two punctures are performed by professional experienced nurses without success, even after using facilitating techniques (i.e. application of tourniquet 5-10 cm above puncture point, vasodilation with alcohol, hanging the forearm downward, tapping of visible veins). 

Venous dilation can increase success of vein cannulation with fewer attempts and decrease procedural time, specially focusing on patients in which difficult cannulation is expected, in addition of minimizing discomfort for the patient and the practitioner. 

Although few decades ago nitroglycerin ointment was successfully described for this purpose, there is no current available data for supporting the nitroglycerin spray solution would have the same effect. 

We aim to determine whether the topical spraying of a solution containing 0.8 mg of nitroglycerin (Trinispray®) could increase the vein cross-sectional area (CSA) in the upper extremity.

Methods:

We conducted a clinical trial in a tertiary university hospital. We recruited a convenient sample of 24 healthy volunteers. Those who meet the following criteria were eligible for the study, at least 18 years of age and obtained a written or verbal informed consent. After enrollment, subjects underwent ultrasound of two veins: CSA of median vein of right forearm (RV) , and CSA of dorsal arch vein of left hand (LV). Both were measured at baseline and 10 minutes after spraying. All the measurements were obtained using a super-high frequency ultrasound machine (Vevo MD). To avoid cofounding factors, other variables were measured and recorded as well, such as age, blood pressure, SpO2 and heart rate before and after intervention. The compression of the tourniquet was applied with a sphygmomanometer applying a pressure 10mmHg over the systolic blood pressure. All statistical data were analyzed using Stata, and due to type and data distribution, Wilcoxon Test was applied. 

Results: 

From our sample, 10 (42%) were female, with a mean age of 35.7 years (SD 9.3). Basal RV CSA was 21.6 mm² (p25= 13.9, p75= 29.4), and basal LV diameter was 5.68 mm² (p25= 3.45, p75=8.0). After intervention, RV CSA was 27.7 mm² (z= -4.129, p=0.00), and LV CSA was 8.3 mm²  (z = -4.229; p=0.00). No statistical differences were found in mean systolic blood pressure (SBP) , diastolic (DBP), heart rate (HR) nor SpO2.  No side effects were observed amongst volunteers. 

Discussion and Conclusions: 

This study suggests that topical application of nitroglycerin in spray (trinispray®) is fast, safe and produced a significant venous dilation both in the dorsal arch of the hand and in the median vein of forearm. This procedure could ensure a higher probability to have a successful venous cannnulation. Probably the weaknesss of our study is that is based on healthy volunteers. More research is needed to assess the effect in real patients. 

All authors have contributed equally to this work. All authors read and approved the final manuscript. All authors have no disclosures. This study did not receive any specific funding. This work has not been presented at any conferences or published before. We certify that this study was conducted in conformity with ethical principles of our institutions.

Background and Aims: Acute hepatic porphyrias (AHPs) are rare, often misdiagnosed genetic diseases caused by an enzyme mutation responsible for heme synthesis. This results in accumulation of neurotoxic heme intermediates, aminolevulinic acid and porphobilinogen, that can cause life-threatening attacks, often requiring urgent medical care and/or hospitalization. EXPLORE (NCT02240784) is the first international, prospective study (currently ongoing) characterizing the natural history and clinical management of AHP in patients with recurrent attacks. High healthcare utilization has been previously reported in the overall EXPLORE population. The aim of this analysis is to report European country specific differences in patient reported emergency department (ED) attendance and hospitalisation for porphyria related issues.

Methods: EXPLORE enrolled patients with AHP with recurrent attacks (≥3/year) or receiving hemin prophylactically. Patients’ medical history and questionnaires on porphyria symptoms, quality of life, and healthcare utilization were collected at prespecified intervals and during attacks. ED visits, and hospitalizations from 13 European countries are presented.

Results: Sixty-three patients from 13 European countries were enrolled mean age, 41 years; 87% female; and 97% with AIP, 3% VP, and 0% HCP). Thirty-two (51%) patients did not attend the ED for porphyria-related issues in the 12 months prior to enrolment; 31 (49%) patients made 165 ED attendances; mean 2.6 (SD=4.72); median 0.0 (min=0, max=20). Nineteen (31%) patients had zero overnight stays for porphyria-related issues and 42 (69%) patients had a total of 241 admissions overnight; mean 4.0 (SD=7.69); median 2.0 (min=0, max=48) in the 12 months prior to enrolment.

Discussion and Conclusions: Three possible patterns of ED attendance and overnight admissions were observed in different European countries. In England, Wales and Finland very similar numbers of ED attendances and admissions for porphyria may suggest patients attend the ED to access the porphyria service. In Germany, Switzerland, Italy and Spain, more ED attendances than admissions may suggest porphyria is managed in the ED without necessarily requiring admission. In France, Netherlands and Poland, fewer than half of the admissions for porphyria were via the ED, which may indicate more direct access to porphyria services without ED attendance. Patients with AHPs with recurrent attacks have a high degree of healthcare utilization in Europe and may attend the ED for porphyria-related issues. ED staff should be aware of the symptoms and signs of AHPs and local guidelines for management of acute attacks and the regional availability of specialist porphyria services.

Trial Registration: ClinicalTrials.gov Identifier: NCT02240784 Funding: Study sponsored by Alnylam Pharmaceuticals Ethical approval and informed consent: IRB/IEC approval was obtained prior to study onset and written informed consent was obtained from each patient as required by national regulations and ICH GCP.

Background

Up to 20% of febrile young children between three to thirty-six months old present to the emergency department (ED) with fever without source (FWS). A minority of these children are clinically well in spite of having serious bacterial infection (SBI). Identifying these children may be difficult on clinical grounds alone. Biomarkers, such as C-reactive protein (CRP) and procalcitonin (PCT), may be useful in identifying those patients with SBI; and their role has been extensively studied. It is important to view these studies in the context of the evolving epidemiology of SBI, especially following the introduction of Haemophilus influenzae type b and pneumococcal immunisations. The primary aim was to explore the utility of CRP and PCT as markers for SBI in young children with FWS in the ED; and to review their applicability in the setting of variable pneumococcal vaccine uptake.

Methods

This systematic review was performed using electronic databases and reference tracking. Studies of SBIs in children aged three to thirty-six months with FWS in the ED were included, as long as they contained enough data for a 2x2 table. The yield for both tests was assessed by performing meta-analysis at the various thresholds, with the ability of both tests to predict SBI being evaluated by using Receiving Operator Characteristic (ROC) curve analysis. Metaregression analysis was used to assess for any association between performance for the studied biomarkers and the background pneumococcal vaccination rate.

Results

A total of eleven studies were identified. All eleven studies (3,112 patients) were included for CRP analysis; nine studies (2,779 patients) were used for PCT analysis. CRP and PCT test characteristics were compared for identifying SBIs in young children. Both markers had a similar performance: Area Under ROC curve (AUROC) for CRP was 0.84 (95% CI 0.80 – 0.87) and for PCT was 0.87 (95% CI 0.84 – 0.90), with overlapping sensitivity, specificity, positive and negative likelihood ratios and diagnostic odds ratios. Optimum cut-off for CRP was identified at 40 mg/l and for PCT at 1 ng/ml; PCT had better specificity (0.91 versus 0.80). PCT performed better at identifying SBIs within the first eight hours of presentation, with an AUROC of 0.95 (versus 0.77 for CRP) and specificity above 0.9 (comparable values for CRP not available). There was no advantage when combining both tests – sensitivity increased at the expense of specificity. The performance of both tests was sensitive to the pneumococcal vaccination rate, but this difference could not be quantified with the available summary data.

Discussion

CRP and PCT had similar diagnostic accuracy in identifying SBI in young children with FWS in the ED, but PCT performed better within the first hours of fever onset. Both biomarkers lacked adequate sensitivity and specificity to be used as sole criterion in deciding on antibiotics or admission to hospital. There was insufficient evidence to assess how their performance was affected by the rate of pneumococcal immunisation.

Conclusion

This analysis does not favour CRP or PCT in assessing young children with FWS presenting to the ED. 

TITLE:  

Risk stratification of atraumatic limp. Before and after: a retrospective cohort study.

AIMS: 

To review the investigation and outcomes of children presenting with atraumatic limp to a tertiary Paediatric Emergency Department (PED) over a 12-month period (May 2016 - April 2017) following introduction of a revised guideline advocating targeted investigations where red flags exist.  

To compare the data to that from the 12 months preceding introduction of the revised guideline to ensure safety and efficacy. 

To assess adherence to the guideline by local doctors and emergency nurse practitioners (ENP). 

METHODS: 

All patients who attended the PED clinic from May 2016-April 2017 with limp/lower limb pain in the absence of clear trauma were manually identified and the following information gathered: age, sex, presenting complaint, duration of symptoms, examination findings (including ability to weight bear), pyrexia preceding/during PED attendance, blood results (white cell count, neutrophil count, C-reactive protein, erythrocyte sedimentation rate), X-ray and ultrasound (USS) results, clinical diagnoses and final outcomes (including those diagnosed at a later date by other specialists). This was compared with results from the 12 months preceding the new guideline when all patients regardless of clinical findings had blood tests and USS. 

RESULTS:

386 patients attended the PED review clinic with atraumatic limp after introduction of the revised protocol. Of these 226 patients (59%) had investigations on their first PED attendance, and of those investigated, 93.6% had a documented appropriate “red flag” justification for doing so. All febrile patients had investigations. Five patients (1.2%) had duration of history that merited investigation but did not have them performed at first PED attendance. 

30 patients (8%) had a significant pathology including seven patients with juvenile idiopathic arthritis (JIA), four with osteitis/osteomyelitis, two with Perthes’ disease, two with acute lymphoblastic leukaemia (ALL), one with Langerhans Cell Histiocytosis (LCH) and one with metachromatic leukodystrophy. 23/30 had investigations appropriately at first presentation. Of the 7 that did not, one ought to have had for duration of symptoms and subsequently had a diagnosis of JIA.  

The remaining 92% had insignificant pathology, predominantly transient synovitis (73%) with a small number of soft tissue injuries and unexpected minor fractures. 

Compared to the 498 patients attending in 2014-15 who all had investigations (with significant pathology diagnosed in 10%) we are doing 62% fewer blood tests and 72% fewer USS.  

CONCLUSIONS:

Children presenting with atraumatic limp in the absence of red flags can be safely managed at first presentation without blood and radiological investigations. Prevalence of significant pathology in those who do have investigations remains low but it is pertinent to follow these patients up until symptom resolution or the underlying pathology declares itself given the serious nature of the conditions we may find.

Background

Paediatric attendances to Emergency Departments in the UK have been rising over the past decade.   Neonatal attendances pose a particular challenge to healthcare professionals due to non-specific presentations and vulnerability to infection.  Previous studies have shown that many neonates present with low-acuity problems that may historically have been dealt with in the community.  This study aimed to evaluate the characteristics and disposition of neonatal attendances to a tertiary Children’s Emergency Department (CED).

 

Methods

Retrospective observational study of all neonatal (≤28 days of age) attendances to the CED at Bristol Royal Hospital for Children (BRHC) over 12 months (01/01/2016-31/12/2016).  Clinical notes for each child were reviewed and if admitted, further information gained from investigation results and discharge summaries.  Information gathered included sex, age, referral method, presenting complaint, diagnosis, investigations, treatment and admission location.  Data was analysed using STATA V. 12.0.

 

Results

There were a total of 1,205 neonatal attendances with a mean age of presentation of 13.6 days.  The most common presenting complaints were breathing difficulty (18.1%), vomiting (8.3%), poor feeding (8.2%), referred for a contrast study (7.4%) and fever (5.9%).  The most common diagnoses after assessment by a medical practitioner were no significant medical problem (41.9%), bronchiolitis (10.5%), suspected sepsis (10%), reflux (6.1%) and surgical problem (5.6%).  Half of all neonatal attendances had no investigations performed.  The most common investigations done in the remainder were bloods tests (25%), blood gases (22%), urine samples (18%), lumbar punctures (8%) and contrast studies (8%).  Most infants did not require any definitive treatment with 77.7% of attendances requiring advice only.  Just over a third of neonates were admitted (12% to the Short Stay Ward (SSW) and 23% to an inpatient bed).  Mean length of stay was 7.8 hours (range 1-48 hours) on the SSW and 3.5 days (range 0.1-43 days) for an inpatient ward.  The most common diagnosis for admitted patients was suspected sepsis.  Out of 106 septic screens, there were 23 positive bacterial cultures (19 urine, 2 blood, 2 CSF).   Less than 1% required a PICU admission and there was no reported mortality.

 

Discussion and Conclusions

This is the largest recent UK study looking at neonatal presentations to the CED.  With half of neonatal attendances requiring no investigations and over 75% requiring advice rather than a specific medical intervention, this study suggests that some elements of routine newborn care may be becoming the domain of the CED.  This has significant implications for training and service provision.  The most common reason for admission to an inpatient bed was suspected sepsis, however rates of proven bacterial infection were low.  This study had higher rates of admission and investigations than some previous studies, however 12% of attendances had been transferred for specialist care.  Moving forward, strategies to support community services and new parents should be investigated in addition to reviewing neonatal knowledge for those working in the CED.  Regional review of the management of suspected infection may decrease admission rates as may work into developing out of hospital models of care.

N/A

Introduction: the study of mortality in pre-hospital traumatology reveals a quality of care quality and the analysis of causes of death "avoidable" could inform us about the development of trauma network and optimize our quality of care in pre-hospitalized serious traumatized.

Objective: to evaluate the mortality of the traumatized in pre-hospital.

Methods: We conducted an analysis of the database of our electronic registry including all severe traumatic deaths who died in prehospital for a period of 2 years. The time, response time and cause of death, ISS, were analyzed for two age groups (

Results:

We included 42 post-traumatic deaths over a 2-year period. The average age was 46 ± 8. The circumstances of occurrence were by far the highway accident (93%). The vast majority of these deaths (71%) occur in the "golden hour". The most common causes of traumatic death were hemorrhagic and hypoxia was 68% and 33% respectively. An average time for medical intervention was 128 +/- 11 minutes. Patients aged 65 years or older had an increased risk of death (OR = 2.1, 95% CI 1.8-3.4, p = 0.04). 43% of elderly subjects with an ISS> 16 and died within 72 hours of admission, regardless of age and causes directly related to trauma. The rate of preventable deaths was 18% This death was attributable to a long prehospital care delay (> 1h), which was by far the most common hemorrhagic etiology in this group of patients.

Conclusion: Elderly patients regardless of their Injury severity score have an increased risk of death compared to younger patients. Avoidable deaths are correlated with an extended period of care. This encourages us to optimize this delay and the development of a trauma network in the center-east.

Introduction

Physiological deterioration often precedes clinical deterioration as patients develop critical illness. Use of a specific Paediatric Early Warning Score (PEWS), based on basic physiological measurements, may help identify children prior to their clinical deterioration. NHS Scotland has adopted a single national PEWS – PEWS (Scotland). Objective

We aim to look at the utility of PEWS (Scotland) in unselected paediatric ambulance patients.

Methods

A retrospective cohort of all ambulance patients aged under 16 years conveyed to hospital in Scotland between 2011 and 2015. Patients were matched to their 30 day mortality and ICU admission using data linkage.

Results

Full results were available for 21,202 children and young people (CYP). On multivariate logistic regression, PEWS (Scotland) was an independent predictor of the primary outcome (ICU admission within 48 hours or death within 30 days) with an odds ratio of 1.403 (95%CI 1.349 to 1.460, p<0.001) – see Table 1. Area Under Receiving Operator Curve (AUROC) for aggregated PEWS was 0.797 (95% CI 0.759 to 0.836 ,p<0.001). The optimal PEWS using Youlden’s Index was 5.

Discussion

These data show PEWS (Scotland) to be a useful tool in a pre-hospital setting. A single set of physiological observations undertaken prior to arrival at hospital can identify a group of children at higher risk of an adverse in-hospital outcome. Paediatric care is becoming more specialised and focussed on a smaller number of centres. In this context, use of PEWS in the pre-hospital phase may allow changes to paediatric pre-hospital pathways to improve both admission to ICU and child mortality rates.

This project was funded by a grant from the Laerdal Foundation for Acute Medicine, grant no. 1868

Background

There are many challenges to obtaining informed consent in pre-hospital research. Consent waivers have been granted where a patient is seriously compromised, for example in the RePHILL study. The situation is less clear when a patient is fully conscious but experiencing severe pain. Patient perspective on the provision of informed consent whilst in pain has not been previously investigated. To support a feasibility study investigating pre-hospital analgesia, we sought these views through patient consultation.

 

Methods

Over two separate one-month periods (January and April 2018), face-to-face interviews were conducted on the trauma wards at a UK major trauma centre university hospital by three interviewers. Participants were identified through convenience sampling using the inclusion criteria of the planned feasibility study re; had sustained a traumatic injury, had been conveyed to hospital by ambulance and could recall their pre-hospital treatment. Patients provided verbal consent to be interviewed. Bias between interviewers was minimized through set questions.

The primary outcome was to determine if patients felt they could have provided consent during the pre-hospital phase of their treatment. Secondary measures included patient demographics, injury pattern, information about pre-hospital treatment and qualitative data from open questions relating to patient perspective.

 

Results

Results were obtained from 31 patients (age range 18-94 years). 94% of respondents received analgesia in the pre-hospital setting and 100% expressed that research to improve analgesia access in a pre-hospital setting was a good idea.

57.1% reported pre-hospital analgesia was effective or very effective but 38.7% of respondents reported inability to access pain relief at some point during their pre-hospital and Emergency Department care.

Specific questions focusing on consent revealed that 61.3% of respondents felt they would have been unable to provide informed consent for the proposed trial, comments were themed around acute pain, urgency of care and perceived inability to retain information. Of note, 9 of the 10 respondents that felt informed consent could be given were all patients that had fallen at home with an isolated lower limb fracture. Across all respondents, 83.9% felt that they would have been happy to have informed consent delayed until arrival in hospital.

 

Conclusion

Patient perspective on pre-hospital informed consent in this group has not been defined. Pre-hospital analgesia research is important to patients and it appears that mechanism and injury type could be an important factor in a patient’s ability to provide informed consent. This survey identified that further work is needed to provide guidance for those conducting research in the pre-hospital setting. We aim to achieve this through further patient survey’s linking clinical data to perceived ability to give consent and wider focus groups to understand influencing factors.

Ultimately patient involvement in setting these standards is fundamental.

Introduction

Use of cognitive-type checklist in the management of crisis situations has proved its effectiveness in operating rooms. The objective of this study was to evaluate the added value of using a checklist in case of cardiac arrest on ventricular fibrillation (CA/VF) in Emergency Medical Service (EMS).

Methods

A professional practice assessment was conducted in March 2017, focusing on the management of CA/VF by an EMS team in a simulation scenario.

Each EMS team was composed of 3 members: emergency physician, nurse and  paramedic. Teams were randomized into 2 groups: a checklist (CL) and a control (CG) group. The first one benefited from a cognitive type of checklist during the medical simulation session, the second one did not. The checklist was produced using AHA guidelines for management of CA. Performance of teams (adherence to recommendations of good practice) was studied.

The same scenario was introduced to each team: 1. a simulated patient, Resusci Anne Simulator manikin from Laerdal (SimMan®), equipped with SimPad SkillReporter, had ST-segment elevation myocardial infarction. 2. He suddenly had a CA/VF that could be detected on monitoring (time 0 of simulation). 3. VF lasted for 10 minutes, followed by 5 minutes of asystole. The simulation was performed under the same conditions as in an advanced life support (ALS) ambulance.

The manikin software displayed different parameters as the scenario progressed. Results from CL and CG were compared for chest compressions (CC) quality parameters and evaluation of decision algorithm.

Results

The use of CL has improved time from VF set up and start of chest compressions (CC), from 27 sec in CG to 12 sec in CL (p=0.03), and the Amiodarone injection (300 mg) after the 3rd EES, from 25% in CG to 80% in CL (p=0.02).

Evaluation of basic cardiac life support (BLS) shows a trend of improvement, as percentage of CC during cardiopulmonary resuscitation (73% CG/77% CL, p=0.23) or time from 1st external electric shock (EES) and 2^nd EES (93 sec CG/128 sec CL, p=0.13). Some parameters were similar as mean time of no flow (8 sec CG and CL), CC mean frequency/min (119 CG/120 CL), CC mean depth (46 mm CG and CL).

Evaluation of ALS shows a trend of improvement, as the time from VF set up and emergency call (10 min 8 sec CG/7 min 32 sec CL, p=0.15), Amiodarone injection (150 mg) after the 5th EES (25% CG/70%CL, p=0.06), time from VF set up and tracheal intubation (6 min 57 sec CG/5 min 50 sec CL, p=0.16). Some parameters were similar as time from VF set up and venous access (108 sec CG/110 sec CL, p=0.67).

Conclusion

Use of a checklist allows teams to begin more quickly the external cardiac massage, and to give Amiodarone at the right time and right dose, in line with AHA recommendations. Use of checklists in an EMS to manage situations such as a CA/VF seems to improve team’s performance. Further work will be required to study the impact of their use in real life conditions.

Background:

It is widely accepted that handover of patients carries significant potential risk.¹ The handover of critically ill patients in the resuscitation room from the pre-hospital to Emergency Department (ED) team is fraught with potential for communication error.² Anecdotally this process within the ED at the Queen Elizabeth University Hospital (QEUH ) was felt to be suboptimal in relation to its lack of standardisation and structure, raising significant patient safety concerns.

 

The use of a structured handover tool, IMIST (Identification, Mechanism / Medical complaint, Injuries / Information regarding medical complaint, Signs / Symptoms, Treatment / Trends) has been shown to reduce variation, improve information volume, and reduce interruption, duration and repetition at the handover³.

 

Aim:

This quality improvement project’s overall aim was to improve patient safety at the handover in the resuscitation room. This was carried out through the introduction of a standardized and structured approach was introduced to the face-to-face handover in the resuscitation room. The project involved the SAS crews as well as the clinical team in the ED, receiving the patient.

 

Methods:

6 key performance indicators (KPI’s) were drawn up through reviewing current literature and discussion with clinical team. Compliance with these standards was assessed for 50 handovers between Scottish Ambulance Service (SAS) crews and the receiving clinical team in resus. Previous work with the SAS had led to the introduction of IMIST as a standardised handover tool between SAS crews and the ED.  The compliance of SAS crews with this method of handover was also assessed.    

 

Following this a number of interventions were implemented using the model for improvement and Plan- Do- Study- Act (PDSA) cycle, involving both paramedics and hospital clinicians. Interventions included: discussing findings at clinical governance meeting, information posters displayed within resus, face to face ‘hot’ debrief with paramedic teams following the handover, increasing distribution of pocket- aide memoirs for SAS crews, skills & drills sessions and simulation training.

 

 

Results:

On analysis of pre-intervention data, it was noted that compliance with KPI’s is currently 58% and the most common KPI’s missed are: no interruptions from clinician during paramedic handover and clarification sought on next of kin. Compliance with IMIST handovers by SAS staff was only 10% and the mean time of handover was 170 seconds (30-752.)

As the staggered changes are implemented, a run chart of results will demonstrate increased compliance with preset standards, an increase in clinician and paramedic satisfaction as well as a decrease in total time of handover.

 

 

 

Conclusion:

It is well recognised that communication errors account for a significant number of adverse clinical events. The handover of critically ill patients from the pre-hospital to the receiving hospital team carries additional potential hazard due to the time critical nature of the handover and multiple human factors involved in dealing with stressful clinical events.

 

Through the implementation of sustainable interventions involving both ED and ambulance staff, communication at handover is improving. This will ultimately benefit the safety of time critical and vulnerable patients

Background: Ultrasound guidance of peripheral venous catheter (PVC) insertion may increase the success rate of the procedure. However, this approach has not been verified in the setting of pre-hospital emergency care so far. Therefore, we decided to perform randomized clinical study to compare two different techniques of pre-hospital ultrasound guided PVC insertion and conventional cannulation technique with regard to the success of the first attempt of PVC insertion, the overall success of cannulation, the number of attempts required for successful PVC placement and the time required to introduce PVC.

Methods: We performed pre-hospital prospective randomized clinical trial. Both physicians and paramedics were involved. Patients treated by emergency medical service and meeting the inclusion and exclusion criteria were randomized to either undergo PVC insertion fully controlled by ultrasound (target vein identification and ultrasound guidance of the tip of the PVC until it penetrates the lumen; Group A); or to undergo PVC insertion partially controlled by ultrasound (target vein identification only, without ultrasound control of placement; Group B); or to receive PVC by conventional approach without any ultrasound guidance (Group C). The goals of the study were compared between the groups.

Results: A total of 300 adult patients were enrolled (100 in each group). The success of the first attempt (group A: 88%, group B: 94%, group C: 76%, p<0.05) and overall success rate (group A: 99%, group B: 99%, group C: 90%, p<0.05) were significantly higher in the groups A and B than in the group C. The differences between the group A and B were not statistically significant. The number of attempts required for successful PVC placement in each patient was significantly lower in the group B than in the groups A and C (group A: 1.18±0.54, group B: 1.05±0.22, group C: 1.22±0.57, p<0.05). Time required for the procedure (regardless of the number of attempts and overall success) was significantly shorter in the group B than in the groups A and C (group A: 75.3±60.6 s, group B: 43.5±26.0 s, group C: 82.3±100.9 s, p<0.05). An analysis of a subgroup of the patients in whom the first attempt was successful showed the time required for PVC insertion comparable in the groups B and C and longer in the group A (group A: 59.0±15.6 s, group B: 40.6±22.2 s, group C: 40.0±13.1 s, p<0.05).

Discussion & Conclusions: Ultrasound guidance of PVC placement was associated with higher success rate of the first cannulation attempt and with higher overall success rate than conventional method, irrespective of the technique of ultrasound guidance. However, PVC insertion partially controlled by ultrasound was superior to full ultrasound control and conventional method in the number of required cannulation attempts for successful insertion and the same technique was not associated with the delay in comparison with conventional method.

Approved by Ethics Committee, University Hospital Hradec Kralove, Sokolska 581, 500 05 Hradec Kralove, Czech republic, Number 6201603S12. Informed consents were obtained.

Introduction: Acute chest pain (ACP) is one of the main causes of acute care in the prehospital setting. Depending on the case, it may need early on-site treatment or a simple medical evaluation determining the eventual need for transportation to the hospital.
In crowded towns, the use of a medicalized motorcycle as first response vehicle could be a safe, fast and economic effector.

Methods: This prospective, monocentric, observational study was conducted from March 2016 to March 2017, on daytime, 3 days a week, in a two-tiered system with medicalized ambulances. It included all adult patients calling the Dispatch Center for ACP. For each case, a moto-team with a pilot trained for emergency drive and an Emergency Physician (EP) was sent on an especially equipped motorcycle (with an electrocardioscope, an oxygen tank, equipment for fluid and drug administration, ventilation, intubation) together with the rest of the Mobile Intensive Care Unit team (ambulance-team), usual effector composed of an emergency medical technician and a nurse. The principal endpoint was the measurement of time-intervals from call for departure to site arrival with each team. Secondary endpoints were stress and comfort scales on the motorcycle, and rate of transportation to the hospital.
Continuous variables were described as medians [Interquartile Range] or mean (min;max) according to the underlying distribution and categorical variables as proportions. Delays were compared using paired Wilcoxon tests. There was an ethical committee approval.

Results: Eighty patients were enrolled in the study (median age: 56.8 years old [46.4;68.0]). Time to reach the patient was systematically faster with the moto-team (9 min [8;10] versus 11 min [8;13], p<0.008). It was comfortable (9.8 [9;10] on a 0-10 numerical rating scale (NRS)), unstressful (0.6 [0;7] on a 0-10 NRS) and safe (0 accident).
Seventy two patients could be managed by the moto-team alone. Among them, all had monitoring and electrocardiogram, 12 had a troponin test, 23 received a treatment (painkiller, sublingual nitroglycerine, adenosine, nebulizations, dextrose), 11 received oxygen. For the last eight patients, the EP preferred to involve both teams from the start.
Eight patients were left at home (4 after treatment), 37 patients were sent to the hospital via regular ambulance, 35 needed medical transport via the medicalized ambulance (the ambulance-team was involved when decision of transportation was taken).

Discussion: The moto-team seems to be a relevant effector for prehospital ACP. This small team reaches the patient faster, makes clinical evaluation, management, and allows no transportation or a transportation with a non-medicalized ambulance for the majority of the patients. Special attention to equipment and training of the moto-team provided safety and comfort.

Conclusion: Medicalized motorcycle seems to be a safe, fast and relevant additional effector for patients with ACP in the prehospital setting. New and larger studies should be undertaken in order to more precisely define a wider range of indications in emergency medical services systems.

Background: Major trauma is a leading cause of morbidity and mortality worldwide. Trauma care systems in the UK are designed to optimise patient outcomes by balancing stabilisation at the scene and safe transfer within a regional network. Current evidence on prehospital time parameters is conflicting; suggesting its relationship with mortality is complex. There are few UK studies addressing transport time and mortality in trauma patients specifically. 

Study objective: To evaluate the association between transport time (TT) and mortality in trauma patients in West Yorkshire. Functional status at discharge was recorded as a secondary outcome using the Glasgow Outcome Scale.

Methods: This was a secondary analysis of a retrospective cohort of major trauma tool positive patients recorded in the Trauma Audit and Research Network registry presenting to one Major Trauma Centre (MTC) during two 12-month periods. These intervals reflect a year prior and a year following a change in local ambulance bypass policy within the Yorkshire region, which extended maximum recommended TT to an MTC instead of a local Trauma Unit (TU) from 45 to 60 minutes. Patients who died at the scene, children under 16 years old and patients transferred from a TU to MTC were excluded. Covariates including injury severity score, age and other prehospital time intervals were controlled for in logistic regression analysis.

Results: There were 355 patients from the period May 2014 - May 2015 and 417 from October 2015 - October 2016. Survival rates to hospital discharge were 86.7% (n=308) and 90.6% (n=378), respectively. There was no significant difference in mortality over the two intervals (p=0.34). TT did not influence survival in 2014-2015 (OR 0.999, 95%CI 0.976-1.023, p=0.947). However, in 2015-2016, there was a significant association with survival (OR 1.042, 95%CI 1.004-1.082, p=0.030). In both years, TT was not a predictor of poor functional status at discharge. Additionally, journeys over 45 minutes were not associated with increased mortality (2014-2015: n=41; p=0.960; 2015-2016: n=51; p=0.484).

Conclusion: This study provides new information for a UK regional trauma network. Longer transport times do not show an increase in mortality or worse functional status in major trauma tool positive patients in West Yorkshire. These findings require validation by an external data set.

Registration: University of Leeds and Yorkshire Ambulance Service

Funding: This study did not receive any specific funding.

Ethical approval and informed consent: « Not needed. »

 

Registry data from Trauma Audit and Research Network. Learning Development Agreement in place with Health Education Yorkshire and Yorkshire Ambulance Service through the University of Leeds.

Diagnostic performance of the quick Sepsis Organ Failure Assessment criteria (qSOFA) score for the early identification of sepsis in patients presenting to the emergency department (ED)

Myrto Bolanaki, Martin Möckel, Stella Kuhlmann, Antje Fischer-Rosinsky, Anna Slagman

Department of Emergency and Acute Medicine (CVK, CCM),

Charité -Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin

 

Introduction

The qSOFA score was suggested from the Third International Consensus Definitions for Sepsis to be used in the ED as a screening tool to identify patients likely to have sepsis.  Data on the diagnostic performance of qSOFA in the ED for the early identification of patients with sepsis are limited.

 

Objective

To evaluate the diagnostic performance of qSOFA for the early identification of patients with sepsis in the ED.

 

Methods

Routine data for all non- trauma patients over 18 years of age who presented in the ED of two large tertiary hospitals in Berlin center within two weeks (11/11/2016-17/11/2016) were extracted from the hospital information system. Data were extensively checked for validity and plausibility and vital parameters were manually completed. The qSOFA score was calculated based on available routine information and one point was assigned for either one of the following criteria SBP≤100 mmHg, respiratory rate ≥22 breaths per min, or altered mentation (Glasgow coma scale<15). Sepsis was defined based on the routine diagnoses and following ICD-codes: A41.X, A40.X, R65.1, A39.2, A39.4, A22.7, A26.7, A42.7, B37.7, A32.7, A20.7.

Results

Overall, 1,334 patients were included. The prevalence of sepsis was 1.1% (n=15) and 4 patients developed a septic shock (0.3%). 91.4% had no qSOFA points (n=1,221), 8.1% had 1 qSOFA point and 0.5% had 2 qSOFA points. The prevalence of Sepsis was 0.6% in patients with no qSOFA point (n=7), 6.5% in patients with 1 qSOFA point (n=7) and 14.3% in patients with 2 qSOFA points (n=1). The area under the receiver operating characteristics curve (AUROC) for the detection of sepsis by qSOFA was 0.728 (95%-CI: 0.570-0.885, p=0.002; figure 1). Logistic regression analysis revealed an odds ratio of 7.64 (95%-CI: 3.50-16.66; p<0.001) for the prediction of sepsis by qSOFA.

 

 

Conclusions

In our study, we were able to demonstrate a moderate accuracy of the qSOFA score for predicting sepsis in non-selected patients who presented to the ED.  Further evaluation could be conducted in order to determine if other parameters increase the accuracy and sensibility of early sepsis diagnosis in ED patients.

INTRODUCTION. 

The Surviving Sepsis Campaign (SCC) guideline recommends transfusion of blood in sepsis when the haemoglobin concentration drops below 7.0 g/dL. Thus, blood transfusion in septic shock is generally reserved as a later-stage treatment option after administration of crystalloids, colloids, vasopressors and inotropic agents. In the ProCESS trial, more patients in the early goal-directed therapy (EGDT) group received packed red blood cell (PRBCs) transfusion than in those in the protocol-based standard-therapy or the usual-care groups, without any apparent differences in the overall outcomes evaluated. 

OBJECTIVES. 
We therefore conducted a randomised pre-clinical trial evaluating whether early administration of blood transfusion has a role in haemodynamic and volume resuscitation for endotoxaemic shock by comparing effects of fresh versus stored blood transfusion in a validated ovine model of hyperdynamic endotoxaemic shock.

METHODS. 
Endotoxaemic shock was induced in sixteen anaesthetised and mechanically ventilated merino ewes by infusing an escalating dose of lipopolysaccharide, LPS (E.coli 055:B5) over 4 hours (total LPS dose 11.25mcg/kg).  During the last hour of LPS, 10mls/kg of either fresh (≤5 days); n=8 or stored (≥30 days); n=8 ovine PRBCs was administered followed by a 12-hour monitoring period. Blood samples were taken for measurement of inflammatory cytokines and coagulation profile.

RESULTS. 
Compared to fresh PRBC, transfusion of stored PRBC in ovine endotoxaemic shock was associated with haemodynamic changes including increased central venous (p<0.0001) and pulmonary artery (p<0.0001) pressures; decreased mixed venous oxygen saturation (p=0.0036) and cardiac index (p=0.0007). Additionally, there was an increase in pro-inflammatory cytokines (IL-6, p=0.0053; IL-8, p=0.0332), a decrease in the anti-inflammatory IL-10 (p=0.0215), and a decrease in the fibrinogen concentration (p=0.0036).

CONCLUSIONS. 
There were significant changes in the haemodynamics, inflammatory response and coagulation profile with stored ovine PRBC transfusion. These results indicate a longer storage duration of PRBCs prior to transfusion could be associated with systemic effects that predispose septic shock patients to poor clinical outcomes.

Australian National Health and Medical Research Council (NHMRC) APP-ID 1061382

Bacterial meningitis in adults is fatal in 20% of patients. Accurate diagnosis is necessary for prompt treatment. In our hospital, there is no bacteriologist on site from 7 PM to 8 AM. During this time, only chemistry parameters of cerebrospinal fluid (CSF) sample were on site analysis, cytological and microbiogical analyses were previously outsourced to another hospital. One year ago, outside bacteriological laboratory hours, we introduced a CSF cytological analysis completed by a multiplexed molecular diagnosis if more than 10 elements /mm3. This system detect 13 organisms (E Coli K1, Haemophilus influenza, listeria monocytogenes, neisseiria meningitis, streptococcus agalactiae, streptococcus pneumonia, cytomegalovirus, enterovirus, herpes simplex virus 1 and 2, herpes virus 6, human parechovirus and varicella zoster virus), the result is available in about an hour.

Patients and Methods: It is a 10 months, monocentric, retrospective, observational study. We evaluated the impact of multiplex molecular diagnosis system in meningitis diagnosis. All patients presenting to emergency department and who had a lumbar puncture with CSF multiplex PCR analysis were included.

Results: 25 patients had a CSF multiplex PCR analysis during the study period. The mean age was 45 year old (15-88). 19 patients had headache, 10 altered consciousness. Fever was present in 76%, nauseas or vomiting in 60% and neck stiffness in 32% of patients. 2 patients had streptococcus pneumonia meningitis, the twice had altered consciousness and had symptoms of otitis. Cell count and CSF protein level were increased, glucose level was low. 4 patients had VZV infections: mean age 31, 2 had zoster eruption, 1 had eruption secondary. They had no inflammatory syndrome, predominant cells were lymphocytes, CSF protein concentration was increased and GSF glucose concentration was normal. 4 patients had enterovirus meningitis, their mean age was 24; 25% had a CSF protein concentration increased and CSF glucose concentration was normal, predominant cells were polymorph. One had herpes encephalitis. 14 had negative multiplex PCR test. All negative tests had negative culture, PCR VZV were positive in a case (< 200 copies/ml). The other final diagnoses were malaria in a case, HIV encephalitis for a patient and a case of toxoplasma. 10 patients were discharged with the diagnosis of “viral meningitis”.

Discussion and Conclusion: In this study, no bacterial meningitis had been misdiagnosed. CSF multiplex PCR analysis is a good alternative when gram staining is unavailable and can give rapid microbiogical diagnosis. Detecting causative organisms is helpful to organize the treatment and can reduce the unnecessary use of antibiotics and length of stay. However, this method is expensive.

Not registred (no appropriate register)

Objectives: We aimed to assess whether a treatment with ceftriaxone/cefotaxime or amoxicillin-clavulanate was associated with in-hospital mortality in patients hospitalized in medical wards for community-acquired pneumonia.

Methods: We conducted a retrospective and multicentric French series of patients hospitalized from the Emergency Department (ED) in medical wards for community-acquired pneumonia between 2002 and 2015. Treatments with ceftriaxone/cefotaxime or amoxicillin-clavulanate were defined by a start in the ED for a duration of 5 days or more and no other beta-lactam. A logistic regression was performed on the overall population, and a propensity score analysis was restricted to patients treated with either ceftriaxone/cefotaxime or amoxicillin-clavulanate.

Results: 1698 patients (median age, 80 y) were included, of which 716 and 198 were treated with amoxicillin-clavulanate and ceftriaxone/cefotaxime, respectively. In-hospital mortality was 10% (9%-12%). In multivariate analysis, factors associated with in-hospital mortality were treatment with ceftriaxone/cefotaxime (aOR, 2.9 [1.4-5.7]), Pneumonia Severity Index class 4 or 5 (aOR, 7.8 [4.3-15.7]), do-not-resuscitate order in the ED (aOR, 8.7 [5.2-14.6]) and fluid therapy and in the ED (aOR, 6.3 [2.5-15.1]). A propensity score analysis was performed on 178 patients treated with ceftriaxone/cefotaxime who were matched with 178 patients treated with amoxicillin-clavulanate, and showed that a treatment with ceftriaxone/cefotaxime was not associated with in-hospital mortality (OR, 1.5 [0.7-3.0]).

Conclusion: In the largest study aiming to compare amoxicillin-clavulanate and ceftriaxone/cefotaxime in community-acquired pneumonia, ceftriaxone/cefotaxime was not associated with a lower in-hospital mortality than amoxicillin-clavulanate. Our results suggest that ceftriaxone/cefotaxime should not be preferred over amoxicillin-clavulanate for patients hospitalized in medical wards with community-acquired pneumonia.

not applicable

OBJECTIVE: To review the utility of Point-Of-Care Ultrasound (POCUS) in Monash Health Emergency Departments (ED). The primary aim of our study was to evaluate the prevalence of ultrasound use in ED patients. The secondary aims were to assess the purpose, indication, and the therapeutic impact of US in the ED.

BACKGROUND: Ultrasound scan (USS) has been recognised as a great imaging tool owing to its non-invasive, non-radioactive nature, with rapid provision of clinical information at the bedside without the use of ionising radiation. Although the utility of USS has been extensively advocated in the emergency setting, the actual rate and indication of utility remain unknown.

Methods: A 31-day, prospective observational study in three Emergency Departments of the Monash Health Network, Victoria, was performed in October, 2016. Data in regard to the frequency and indications of utility, the mode of POCUS use, its diagnostic values, and the corresponding operator were achieved and analysed across these three sites. Factors associated with diagnostic impacts were also identified. 

RESULTS: In the 31-day study period in ED, the prevalence of POCUS was 2%. POCUS was utilised for assessing diagnosis in 87.6% of cases and procedural guidance in 12.4% cases. Majority of the diagnostic POCUS cases were indicated for abdominal pain(73%), chest pain (14%) and multi-trauma/injury (6%); while most procedural POCUS cases were indicated for IV access (85%). The most frequently used modality were eFAST/AAA (46%) and BELS (17%).The sensitivity and specificity of diagnostic POCUS were found to be 91.3% and 94.7%, respectively. Majority of cases were performed by the senior medical staff (51.9%). 

CONCLUSIONS: The prevalence of POCUS in ED was found to be low, despite that fact of it being highly recommended. The utility was mainly for diagnostic assessment of intra-abdominal and cardiac issues. Given its high sensitivity and specificity rate, as demonstrated in this study, utility of POCUS in ED should be further encouraged, particularly among junior medical staff.

Comparison between EXTEM hyperfibrinolysis and FIBTEM hyperfibrinolysis in severe trauma patients : Retrospective study using thromboelastometry

 

Introduction

Rotational thromboelastometry (ROTEM) is a useful method for detecting hyperfibrinolysis (HF) in trauma patients. However, ROTEM detects HF only after high degrees of fibrinolysis. The aim of this study was to identify whether FIBTEM HF could be used to increase the sensitivity of HF diagnosed by ROTEM. We hypothesized that FIBTEM HF have a higher mortality than non-HF.

Method

This is a single center observational cohort study performed in the level 1 trauma center in Korea (Pusan National University Hospital). Trauma patients with an Injury Severity Score (ISS) >15 and ROTEM performed in the emergency department from January 2017 to December 2017 were included. Patients <15 year were excluded. EXTEM HF was defined as maximal lysis (ML) > 15% in EXTEM, and FIBTEM HF was defined as ML > 15% in FIBTEM. We divided patients into three groups, EXTEM HF group, FIBTEM HF group, non-HF group. And we compared the mortality rate of three groups.

 

Result

One hundred and ninety four trauma patients were enrolled into the study over an 1 year period. The mean (SD) ISS was 29.7(9.7). EXTEM HF group, FIBTEM HF group, and non-HF group were 22(11.3%),73(37.6%), and 99(51.0%), respectively. Of the total 194 patients, 48 (27.4%) died. In 17 of 22 patients (77.3%) with EXTEM HF, in 19 of 73 patients (26.0%) with FIBTEM HF, and in 12 of 99 patients (12.1%) with non-HF died (P <0.001 for EXTEM HF vs. FIBTEM HF, P=0.27 for FIBTEM HF vs. non-HF).

 

Discussion & Conclusion

The mortality rate was highest in the order of EXTEM HF group, FIBTEM HF group, and non-HF group. FIBTEM HF as compared with non HF, resulted in significantly higher mortality. FIBTEM HF may be used as a diagnostic modalities to improve the sensitivity of HF diagnosis in trauma patients.

This study was not registered. (I am sorry, it is not our usual practice for an individual retrospective scientific research) This study did not receive any specific funding.

Background :

Trauma is a leading cause of death in young people and hemorrhagic shock is the second responsible of the mortality rates. Hypoperfusion is hard to diagnose clinically and might stay patent especially in young patients in whom adaptative response mechanisms are efficient. In this context, arterial Base Excess (BE) has been proposed to be used as an early indicator of hypoperfusion. The aim of this study was to evaluate the prognostic value in terms of mortality of admission BE in severe trauma patients admitted to the Emergency resuscitation room.

Methods :

It was a prospective observational and prognostic study. We included severe trauma patients with at least one high velocity criteria and admitted to the resuscitation room over a 33 months period. Arterial blood gas sample was immediately withdrawn and BE calculated. Multivariate analysis with logistic regression was performed to identify the predictive factors of mortality at Day one and Day-7 after trauma. Moreover, ROC characteristics and survival curves were underwent. P < 0,05 was considered significant.

Results :

We included 479 patients. Median age was 37 (18-90) with sex-ratio=4.2. Road traffic accident was the most frequent cause n(%) : 358(75).Clinical characteristics were n(%) : Glasgow coma scale <13 : 170(35) ; Systolic blood pressure<90 mmHg : 64(13) and Pulse oximetry <90% : 82(17). Mean injury severity score was 22 ± 13. Rates of mortality were respectively at day 1 and day 7: 2,2% and 27,3%. Median BE was -3,2 mmol/l (-25 ; 28). Forty-five per cent had a BE ≤ -3,5 mmol/l. The Roc curve determined a cut-off value of BE of -6,5 mmol/l. In multivariate analysis, initial BE ≤ -6,5 mmol/L was an independant predictive factor of first day mortality with an adjusted Odds Ratio; [Interval Confidence 95%] = 3,17 ; [1,4-7,1] ; p=0,005. Similar results were found while studying 7 days mortality with an adjusted Odds Ratio; [IC95%] = 1,5 ; [1,14-1,96] ; p=0,003. BE showed high prognostic value for both mortality rates as showed the characteristics of the ROC curves. Survival curve was significant for BE> -6,5mmol/l with better results and p < 0,001.

Conclusion :

In this study, BE showed a significant prognostic value towards immediate and early mortality and could be proposed as a marker of detecting severity in trauma patients admitted to the emergency ward at the very early stage of management.

INTRODUCTION Every year, trauma accounts for about 30% of all consultations at Emergency Reception Services. The patient's self-diagnosis abilities regarding the presence or absence of a fracture following acute trauma would optimize their management. GOAL We wanted to evaluate the ability of patients to suspect and diagnose the presence of a fracture as a result of acute trauma. METHOD It is an observational, prospective and monocentric study. When taken care of by the Reception and Orientation Nurse, the patients who consulted for acute and isolated trauma gave their opinion and their degree of confidence on the possibility of a fracture and the need for an x-ray. . The primary endpoint was the concordance of the patient's self-diagnosis at the final diagnosis at the emergency exit. The secondary endpoint was the degree of confidence with respect to their self-diagnostic capacity assessed by a visual analogue scale (out of ten). RESULT Over the study period, 35% (25/66) of patients suspected the presence of a fracture with a confidence level of 6.23 (+/- 2.36) versus 7.17 (+/- 2.18) in those who did not suspect no fracture. The diagnostic performance of the patient's self-diagnosis was: 57% sensitivity, 69% specificity, 40% VPP and 81% VPN. The agreement between the self-diagnosis of the patient to rule out the presence of a fracture and the absence of fracture objectified in radiology was 67%. CONCLUSION The patient's ability to rule out the presence of a fracture is good. Integrating the self-diagnosis of the patient would optimize the management of isolated trauma in emergencies

Infants ≤90 days old with fever without source (FWS) are in a higher risk of having an invasive bacterial infection (IBI) and a more aggressive management is usually recommended for this population. Few is known about the value of the fever degree for predicting the risk of IBI. Although most of the guidelines do not recommend modifying the management of these patients according to the fever degree, some authors recommend a different approach in well-appearing infant >28 days old depending on the maximum temperature detected.

 

Our objective was to analyze the value of temperature for predicting an IBI or and herpes infection in well-appearing infants 29-90 days old with FWS.

 

METHODS

 

Subanalysis of a prospective multicenter study including febrile infants ≤90 days old with FWS, carried out in 19 hospitals included in the Spanish Pediatric Emergency Research Group (RISeuP-SPERG) between October-2011 and September-2013. Axillary or rectal temperature ≥38°C at home or in the emergency department was considered fever. An IBI was diagnosed when a single pathogen was isolated in blood or cerebrospinal fluid (CSF). Herpes infection was diagnosed when there was a positive chain polymerase reaction (CPR) for Herpes virus in blood or CSF.

 

RESULTS

 

We recruited 3,401 infants. Of them, 2,253 were well-appearing infants >28 days old. In this subgroup, 43 (1.9%) were diagnosed with an IBI (39 positive blood culture, 3 positive CSF cultures and 1 patient with both positive cultures). One patient presented a positive CPR for herpes in CSF.

Area under the ROC curve for temperature for predicting an IBI was 0.623 (0.603-0.643) and for predicting specifically a bacterial positive culture or a herpes CRP in CSF was 0.651 (0.631-0671).

There were 17 IBI among well-appearing patients with less than 38.6ºC. There were no differences in rate of meningitis between patients with <38.6ºC and ≥38.6ºC (0.34% vs 0.31%, p=0.9).

 

Results were similar when analyzing specifically infants 29-60 days old and 61-90 days old.

 

CONCLUSION

Temperature itself has a low accuracy for ruling out an IBI in well-appearing infants 29-90 days old with FWS. Considering the prevalence in patients with lowest temperatures, we do not recommend modifying the management of these patients according to the maximum registered temperature.

Social and judicial reports of children suspected of abuse in one pediatric emergency department : description and follow up.

 

Introduction

Pediatric emergency departments are key actors in child welfare. However, children at risk of abuse may be difficult to detect and may remain unrecognized. One of the barriers to report children at risk of abuse to the child protective services (social report SR) and/or to the prosecutor’s offices (judicial report JR) is the fear to be mistaken.

This study described the reported cases of children suspected of abuse in one pediatric emergency department and their social and judicial feedbacks.

 

Materials and method

Retrospective analysis of visits of children 0 to 18 years old in one academic pediatric emergency department between January 2015 and December 2016 that were followed by a SR or a JR.

Demographics data, previous visits in an emergency department and the reasons of the SR or the JR were gathered from the electronic medical records. The follow-up and outcomes of the SR or the JR were searched for by contacting the child protective services and the prosecutor’s office.

 

Results

During the study period in our hospital, 180 children that visited the emergency department had a SR or a JR, that is a prevalence of ,0.17% of visits. The median age was 4.4 years (19 days - 17.3 years), 68 children (38%) were less than 2 years old. A previous visit in an emergency department was found in 84 children (47%) mainly for "skin lesion, hematoma" (N = 35, 19%) and "burn" (N = 17, 9%). The child abuse was mainly suspected because of the parents' attitude, the features of the lesion and/or the fact that the child or his parents had reported the abuse. Physical abuse was the main diagnosis (N = 73, 40%). There were 122 SR and 58 JR. A follow up was available for 155 children (98 SR and 54 JR). Fifty eight SR (38%) resulted in a socio-educational intervention. All of the JR (N = 54) and 40 SR resulted in an initial judicial process. From these 94 judicial processes, 32 Provisional Placement Orders (34%) were made. Finally, 3 of these 155 children had a differential diagnosis of child abuse and were dismissed by the child protective services and/or the prosecutor’s offices.

 

Conclusion

Most of the SR and JR performed during the study period resulted in social or judicial measures. Moreover, the differential diagnoses of child abuse in children with a SR or a JR were extremely low. However, the prevalence of SR and JR in our pediatric emergency department was low. These results should encourage physicians that suspect child abuse to report it and endorse researchers and projects that aim to improve child abuse detection, management and follow-up.

Background:

Pain is the most common symptom encountered when patients are in the care of the emergency services and acute trauma is a common cause.  The 2012 Royal College of Emergency Medicine (RCEM) audit of 166 UK Emergency Departments (EDs) revealed a significant gap between the stated standards and resultant outcomes in the management of paediatric pain.  The process of care describes the interactions between the health care providers and patients (e.g. timeliness, safety, efficiency, best practice and equity).  The outcomes of the care include clinical measures (e.g. pain scores).

The aim of this study was to establish the processes and outcomes of current practice for the management of pain in children and young people presenting to EDs with minor injuries.

Methods:

A prospective chart review service evaluation was performed in December 2016 by the Paediatric Emergency Medicine in the United Kingdom & Ireland (PERUKI) network.  Participating sites identified and abstracted data over a one week period.  All children under 16 years, presenting with injuries, were eligible for inclusion.  Major trauma alert calls were excluded.

Anonymised data were collected, including demographics, injury characteristics, disposition, and analgesic processes and practices both before and during the ED visit.

Results:

34 sites accepted the invitation to participate and data sets were submitted by 31 sites (comprising 21 tertiary/university hospitals and 10 rural/district general hospitals) (response rate 91.2%).  Data was provided for 3888 patients.  The number of patients per site ranged from 11 to 292 (median 104, interquartile range 80 – 159.5).

Injury types included sprains (20.0%), lacerations (18.8%), contusions/abrasions (18.4%), fractures (17.1%), minor head injuries (5.9%), burns/scalds (3.3%), dislocations (1.9%) and other injuries (14.6%).

818 patients (21.0%) received analgesia prior to ED attendance.  The proportion of pre-hospital analgesia administration was 51.4% of 170 patients who arrived by ambulance or 53.2% of 47 patients who were transferred from another ED.

Initial pain assessment documentation was present for 2235 (57.5%) patients, with only 3.5% of charts containing a repeat pain score.

Analgesia was offered to 1812 patients (46.6%).  Administration of analgesia occurred for 1533 patients, representing 84.6% of those who were offered it.  Analgesia within 30 minutes of registration occurred in 56.7% of patients.  Paracetamol was the most commonly administered analgesic (72.8% of recipients).  Opiate/opioid was administered in 86 cases (5.6% of patients receiving ED analgesia).

Moderate or severe pain was encountered in 456 patients with 197 (32.0%) receiving analgesia within 20 minutes of arrival (RCEM standard).  Of the 160 patients with severe pain, only 15 (9.4%) received intranasal or intravenous opioids (RCEM standard).

Discussion & Conclusions:

We present the benchmark assessment of the processes and outcomes of current practice for the management of pain across the PERUKI network.  This study illustrates areas of both excellent and suboptimal performance.

Initial pain assessment, pain reassessment and the timely administration of appropriate analgesia to patients with moderate or severe pain are key targets for improvement.

The protocol was reviewed by the Research and Development department at the lead site and was adjudged to be service evaluation, with no requirement for ethics review. The study was unfunded and unregistered (no appropriate register).

Background:The endotracheal intubation is an essential skill for pediatric residents who are faced with critically ill children.In that failed endotracheal intubation may lead to systemic complications secondary to hypoxia and mechanical complications of trauma.Direct laryngoscopy is the most preferred method as it is easily achievable and utilizable.The direct visualization of vocal cords is required for direct laryngoscopy, but it can be impossible for critically ill children occasionally.Although videolaryngoscopes can be utilized in these cases, there is limited knowledge about efficacy of different videolaryngoscopes.We aimed to compare efficacy of McGrath MAC® and Storz-PM® videolaryngoscopes with direct laryngoscope in infant and child airway manikins when used by pediatric residents.Methods:This study was performed in Dokuz Eylul University Department of Pediatrics between March 2016 and June 2016.Residents of pediatrics subscribed written informed consent, were enrolled.Each participant performed intubation with three different laryngoscopes and seven different blades which were used according to the computerized random number generator.Two different manikins were used sequentially.Primary endpoints were intubation time and success rate of intubation.Secondary endpoints were the Cormack & Lehane grade of best glottic view and assessment of each device based on visual analogue scale (VAS).ANOVA was used to compare intubation time and VAS among groups. Student’s t-test was used to compare these parameters between the  training status groups.The rates for intubation success and Cormack & Lehane grade were analyzed with Pearson’s Chi-Squared test.Results:Thirty four participants had airway management training.In child manikin, the intubation times were significantly shorter in DL group compared with all other device groups (p<0.001).The difference of success rates were statistically significant (p=0.014), but intubation success rates were similar for DL, McG and ST groups (p=0,591). In infant manikin, the intubation times were significantly different for attempts (p=0.005) and the intubation time of DL group was the shortest with 15.79 ± 9.41 seconds.The intubation success rate of McG group was lower than other device groups (p=0.041).The intubation time and success rate of trained group were significantly different when compared with untrained group (p<0.001 and p=0.020 for attempts in child manikin; p<0.001 and pDiscussion & Conclusions:We showed that compared with direct laryngoscopy the utilization of McGrath MAC® and Storz-PM® videolaryngoscopes led to longer intubation time without improvement in the intubation success rate.Also we found that the intubation attempts of trained pediatric residents were faster and more successful than untrained residents.A large number of studies have been compared direct laryngoscopy with indirect laryngoscopy in children.In a metaanalysis performed by Sun et al, it was suggested that videolaryngoscopes provided better glottis view with longer intubation time and similar success rate which supports our results. Also recent metaanalysis have been suggested that indirect laryngoscopy leads to longer intubation time with increased failure rate when compared with direct laryngoscopy.In conclusion McGrath MAC® and Storz-PM® videolaryngoscopes can’t provide increased success rate and decreased time of pediatric intubation and trained pediatric residents performed faster and more successful intubations than untrained residents.The airway management training with both direct laryngoscope and videolaryngoscopes can lead to improvement in pediatric intubation.

There is no appropriate register. This study did not receive any specific funding.