The objective of this study is to evaluate the prognostic impact of response to immediate short-term intensive care therapy within the Emergency Department (ED) on mortality of septic patients.

Methods

This is a prospective single center study on a cohort of septic patients in an urban academic Emergency Department (ED) with about 36.000 patients per year and an inpatient rate of 55%. In this center all emergency patients are seen and treated in the ED. Direct tranferrals on the ICU do not occur. The ED offers six beds for immediate and short-term intensive care therapy. In the study, patients were assessed by nurses at arrival and triaged according to the EmergencySeverityIndex. In case of suspicion of sepsis according to the SIRS criteria they were immediately admitted to a monitor bed in the ED based intensive care area, received a q-SOFA and MEWS scoring and were treated according to current sepsis guidelines. Patients above 18 years and a final diagnosis of sepsis (according to ICD-10-coding) have been included in the study. Vital signs have been measured with GE Healthcare Solar 8000M and treatment effects on vital signs including q-SOFA parameters were assessed after 4 hours of ED based intensive care and compared with the measurements at the beginning of the treatment period. Data were collected from April to October 2015.

Results

117 patients with sepsis were identified in the study period of 6^th month. The overall in hospital mortality of this group was 14,5%(n=17). Patients with a q-SOFA ≤ 1 at the beginning of the treatment period had a mortality of 8,3%.

When only those patients with a q-SOFA > 2 at the beginning of the treatment period were analyzed (38,5%;n=45) the mortality in this group was 24,4%(p<0,05) which was significantly higher than the the overall mortality of all septic patients. Successful initial intensive treatment in ED had a profound effect on mortality in this group. An improvement of the initial q-SOFA-score ≥2 within four hours to a qSOFA-score ≤1 (27 out of 45) correlated with a mortality of 7,4%(n=2) which was not significantly different from those patients with a q-SOFA ≤ 1at ED arrival. However, in those cases without successful improvement of q-SOFA and a score of ≥ 2 after 4 hours intensive ED treatment (40%;n=18) mortality was 50% (n=9) with 50% of the patients dying within the first 3 days of hospital stay.

Conclusions

This study shows that ED based early intensive care intervention in severely septic patients can downgrade q-SOFA scores ≥ 2 to sofa score ≤ 1 within 4 hours in more than 50% of patients. Furthermore, the treatment response in this period is a strong prognostic factor. Good treatment reactivity profoundly reduces the risk of in hospital mortality to levels of septic patients with an initial q-SOFA ≤ 1 at ED arrival. Thus, successful early ED based intense sepsis treatment helps to reduce mortality and allows discriminating those patients who will need further ICU treatment from those who can be treated without using further ICU resources.

Objective To determine if intravenous paracetamol was superior to oral paracetamol in the management of moderate pain in out of hospital setting.

Methods:

This was a prospective, randomized, preliminary study over 3 months.

We Included adult patients (age >14ans), with mild to moderate pain (visual analogue scale (VAS) = <6). They received analgesia with 1 gram of paracetamol and this in the absence of contraindication to the prescription of the product or its oral form (a known allergy to paracetamol, hepatocellular insufficiency, renal insufficiency, deterioration of the Consciousness). The patients were randomly assigned to receive either the intravenous paracetamol (IV group) or oral paracetamol (PO group).

Demographic data, hemodynamic parameters, visual analogue scale (VAS), paracetamolemia before taking medication and thirty minutes afterwards were collected.

The primary judgement criterion was Visual Analogue Scale (VAS) pain reduction at 30 min. A clinically significant change in pain was defined as 15 mm.

Our secondary endpoint is the difference of paracetamolemia 30 minutes after administration of the drug in both groups. A value of p <0.05 was considered significant.

Results:

Twenty patients were identified, 7 in PO group and 13 in IV group. Male predominance was noted (sex ratio = 1.5), mean age was 32.8 old  years +/- 16. No significant difference was demonstrated in this trial with intravenous paracetamol compared with oral paracetamol in terms of the average of pain scale reduction at 30 minute (42.5 +/- 12.81 mm in PO group vs 38.33 +/- 15.85 mm in IV group; p=0.54). However, a significant difference in paracetamolemia at 30 minutes was noted between the two groups (4.33 +/- 3.39 mg/l in PO group vs 18.58 +/- 7.4 mg/l in IV group; p< 0.01).

Conclusions:

Our preliminary study, showed a better bioavailability of the IV form of Paracetamol. However there was no significant difference in analgesia. Multicenter studies including a larger number of patients are needed in order to confirm this result.

Background: During winter season, influenza viruses are responsible for a large proportion of acute respiratory illness in emergency department (ED) patients. To confirm the diagnosis of an influenza virus infection, standard polymerase chain reaction (PCR) test is accurate but takes at least 24 hrs to have the results available. In ED patients, it is important to decide timely whether a patient has to be isolated due to influenza virus infection in case of in-house admission. Rapid PCR tests have equivalent diagnostic accuracy to standard PCRs but produce a result in less than one hour. The aim of our study was to determine the prevalence of influenza virus infection in ED patients with influenza-like-illness and to identify risk factors for in-house admission.   

Methods: In a retrospective analysis, we enrolled consecutively ED patients with influenza-like-illness from two winter seasons (December to April 2015/16 and 2016/2017). During the ED stay, these patients were isolated according to the guidelines for acute respiratory illness. In case of in-house admission, rapid PCR tests were performed in these patients. The primary endpoint was to assess the prevalence of influenza virus infections and secondary, to identify signs and symptoms as well as further predictors that were associated with in-house admission due to the influenza virus infection. Descriptive, univariate and multivariable logistic regression analysis was performed.  

Results: The rapid PCR test was performed in 842 ED patients with influenza-like-illness who were supposed to be admitted in-house. Hundred-eighty-two patients (21.6%) were influenza positive, mostly influenza A (n=149). There was no difference in age between influenza positive and negative patients (median 60 vs. 63 years), in contrast more female patients were influenza positive (48.4% vs. 38.8%). Mostly a symptomatic treatment was sufficient, whereas influenza positive tested ED patients received more often antiviral therapy comparing to negative tested patients (36.8% vs. 1.8%). Three main symptoms leading to ED presentations were: 68.5% coughing, 52.9% fatigue and 44.9% dyspnea. Female ED patients (p=0.039) presenting with cough (p<0.001), fever (p=0.002) and myalgia (p=0.010) were at increased risk for an influenza virus infection. If dyspnea was additionally present to these symptoms, ED patients were more likely to have an influenza virus caused pneumonia when presenting in the ED (p=0.006). Senior ED patients (p<0.001) with influenza virus infection, known coronary heart diseases (p<0.001) and symptoms such as fatigue (p<0.001) and dyspnea (p<0.001) were identified to be at high risk for in-house admission.

Discussion & Conclusion: The rapid PCR test was positive in every fifth ED patient with influenza-like-illness and the likelihood increased if symptoms like cough, fever and myalgia led to ED presentation. Rapid PCR tests are useful to plan rapid isolation measures in case of in-house admission to avoid contamination of other patients, to identify ED patients at risk for pneumonia and to reduce unnecessary antibiotic use but increase focused use of antiviral therapies if needed.

Background

Sepsis was redefined as “life-threatening organ dysfunction caused by a dysregulated host response to infection” in 2016. As the concept of systemic inflammatory response was superseded, a new clinical criterion for identifying patients with high risk of organ dysfunction – the quick Sepsis-related/ Sequential Organ Failure Assessment (qSOFA) score - was recommended. qSOFA consists of altered mentation, tachypnoea (respiratory rate ≥22 bpm) and hypotension (systolic blood pressure ≤100mmHg). The aim of this systemic review was to review the existing evidence to determine the validity of qSOFA in the prediction of prognosis to justify its use in the emergency setting.

Methods

All studies that recruited adult patients where qSOFA was calculated and reported were included, except case series, case reports and conference abstracts. Studies that only included patients with neutropenic fever were excluded. Literature search strategies were developed using Medical Subject Heading (MeSH) and text words related to qSOFA. The Cochrane Central of Controlled trials, EMBASE, BIOSIS, OVID MEDLINE, OVID Nursing Database and the Joanna Briggs Institute EBP Database were searched using the OVID interface. The WHO International Clinical Trial Registry Platform, Web of Science, Scopus, ClinicalTrials.gov were searched independently. Risks of biases were assessed using an adapted version of the QUality In Prognosis Studies instrument. The validity of qSOFA was determined by comparing the Area Under the Curve (AUC) of Receiver Operating Characteristic (ROC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) for predicting in-hospital mortality. Prognosis was defined as in-hospital mortality, which was also our primary outcome. Other outcomes include 28/30 day mortality, ICU admissions, ICU length-of-stay, hospital length-of-stay, and diagnosis of sepsis.

Results

Our search identified 529 records. After duplicates were removed, 251 records were screened and 43 papers were included in the final analysis.

For in-hospital mortality AUC, we reviewed 26 studies that had 381,029 subjects and found the median to be 0.68 and ranged from 0.55-0.82. In-hospital mortality median sensitivity and specificity was 0.53 and 0.83; it was compiled with data from 22 studies with 116,490 patients, ranging from 0.16-0.98 and 0.19-0.97 respectively. Positive and negative predicted values ranged from 0.07-0.4 and 0.89-0.99 respectively. Selection bias and bias in definition were the most common biases. The included studies differed mostly on altered mentation and how this was determined.

Meta-analysis of 376,723 subjects confirmed high heterogeneity among studies (I²=99.0%, 95%CI 98.8-99.1). Caution must be exercised when interpreting the results of the pooled in-hospital mortality AUC of 0.69 (95%CI 0.66–0.71).

Discussion

For screening scores or systems to be useful in the emergency setting, sensitivity of the test should be high to ensure that potentially critically ill patients are not missed. qSOFA may have been suggested to be a useful predictor of organ dysfunction, however our systematic review shows that it is not a clinically reliable score clinically for predicting patients for in-hospital mortality. 

Background and Objectives

Lumbar puncture (LP) is an emergency department (ED) procedure for evaluating children with suspected central nervous system (CNS) infection. We aimed to determine whether extensive testing of cerebrospinal fluid (CSF) obtained from pediatric ED patients was useful and to demonstrate the distribution of abnormal bacterial or viral pathogens.

 Methods

Charts of all patients who underwent LP in a children’s hospital ED between January 2014 and December 2017 were reviewed. Patients demographics, clinical characteristics, complete blood count, C-reactive protein (CRP) and CSF results (bacterial culture, viral serology and biochemistry) were recorded. Based on the CSF findings patients were categorized into three groups; normal CSF, abnormal CSF indicating viral meningitis (VM) / encephalitis, and abnormal CSF indicating bacterial meningitis (BM). 

Results

A total of 260,000 patients were presented to the PED during the study period, LP were ordered to perform for 130 of them (1/2000). However, 23 patients for consent issues, and 6 patents with missing data were excluded. Final analysis performed for 101 patients; the mean age was 48 ± 12 months and 60 (59%) were males. The most common causes for ordering LP were altered mental status (36%) and fever without source (25%). Normal CSF was detected in 45% of the patients, CSF indicating of VM or encephalitis in 41% and BM in 14%. Patients with abnormal CSF indicating BM were more likely to have headache and less likely to have seizure (respectively p = 0.012, p = 0.023). No patient with seizure had CSF findings indicating VM. The mean age, absolute neutrophil count and CRP values were higher in patients with abnormal CSF indicating BM (p = 0.001, p = 0.003 and p = 0.001, respectively). Bacterial cultures grew 5 pathogens in 8 patients ( 2 Pneumococcus, 2 Hemophilus influenza non-type b, 2 Neisseria meningitidis, 1 Escherichia coli and 1 Streptococcus Hominis) whereas only three viral agents were detected in  8 patients (6 Enteroviruses, 1 HSV type 1 and 1 HHV 6).

Conclusion

Since the results of LPs contributed directly to patient management in majority of cases, either by identifying an organism, allowing unnecessary antibiotics/antivirals to be stopped after 24 hours, or by permitting an earlier discharge from the ED, it should not be delayed when clinical and laboratory tests indicating CNS infections.

Background

An international task force has suggested the use of the quick Sequential Organ Failure Assessment (qSOFA) score instead of systemic inflammatory response syndrome (SIRS) to identify patients with high risk of mortality. Only few studies have evaluated the new sepsis criteria prospectively in emergency department (ED) settings.

Purpose

To determine the prognostic value of qSOFA compared to SIRS in predicting 28-day mortality in a prospective study of infected patients admitted to an ED.

Methods

A prospective observational cohort study of all infected patients aged 18 years or older admitted to the ED of Slagelse Hospital during October 1 to December 31 2017. The ED is a tertiary care center with 26,500 visits per year. All patients with suspected or documented infection on arrival to the ED, and who treated with antibiotics, were included. Admission variables included in the SIRS- and qSOFA criteria were prospectively obtained from triage forms. Survival status after 28 days from admission was obtained from the Danish Civil Registration System. The diagnostic performance of qSOFA and SIRS score for predicting 28-day mortality was assessed by analyses of sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating curve (AUC) with 95% confidence intervals (CI).

Results

A total of 1,075 patients (50.4% male) were included. The median age was 73 years (interquartile range 59-83 years). A total of 91 (8.5%) met at least two qSOFA criteria, and 523 (48.7%) met at least two SIRS criteria on admission. The overall 28-day mortality was 9.7% (95% CI 8.1-11.7%). Unadjusted odds ratio of qSOFA and SIRS for 28-day mortality was 4.0 (95% CI 2.4-6.8) vs 1.6 (95% CI 1.0-2.4).

A qSOFA score of at least two for predicting 28-day mortality had a sensitivity of 23.1% (95% CI 15.4 -32.4%), a specificity of 93.1% (95% CI 91.3-94.6%), a PPV and NPV of 26.4% (95% CI 17.7 -36.7%) and 91.9% (95% CI 90.0-93.5%), respectively. A SIRS score of at least two for predicting 28-day mortality had a sensitivity of 58.7 (95% CI 48.6-68.2%), a specificity of 52.4% (95% CI 49.2-55.6%), a PPV and NPV of 11.7% (95% CI 9.0-14.7%) and 92.2% (95% CI 89.7-94.3%), respectively. The AUC for qSOFA and SIRS was 0.58 (95% CI 0.54-0.62) vs 0.55 (95% CI 0.50-0.61).

Conclusion

Use of qSOFA had improved specificity, but with poor sensitivity, in predicting in 28-day mortality.  qSOFA and SIRS showed similar discrimination potential for mortality.

Background:

Patients experiencing sepsis can rapidly develop organ dysfunctions whose intensity and duration have been linked with deleterious outcomes. In addition, a substantial proportion of sepsis survivors suffer from long-term sequelae, such as recurrent sepsis with hospital readmissions, functional and cognitive impairments, increased cardiovascular and renal complications and overall a decreased life expectancy. As early management of sepsis has been proven successful, accurate triage of these patient when admitted to hospital is key. While, no biomarker is available, endothelial damage plays a major role in the pathophysiology of organ dysfunction and biomarkers associated with this early sepsis events could be an early sign of worsening in ED.

Objectives:

We investigated whether biomarkers could predict the deterioration of in patients admitted at emergency department with a suspected infection.

 Methods:

TRIAGE was a prospective, multicentre (14 sites in France and Belgium) observational study. Adult patients admitted in the ED with a suspected infection and at least 2 SIRS criteria were included. Blood samples were collected at 0, 6 and 24 hours after patient admission. Main outcome was subsequent deterioration (defined as any of the following: death, ICU admission, increase of SOFA score) within 72 h. This primary endpoint was assessed by an independent adjudication committee of sepsis experts including emergency physicians and intensivists. Biomarkers association with primary endpoint and prognostic performances were assessed with and without adjustment on clinical variables such as age, sex, Charlson score, SOFA score, qSOFA, lactates, using logistic regression models. AUC under the ROC curve and their 95% confidence interval were computed using DeLong’s method. Predictive performances (sensitivity, specificity, Negative Predictive Value, Predictive Positive Value) were assessed using classification threshold optimized for high sensitivity.

Results:

sVEGFR2 and sUPAR protein levels were measured in 462 ED patients with a suspected infection at 3 time points: H0 (patient’s admission), H6 and H24. Of these 462 patients, 124 (27%) worsened within the 72-hour study period. Compared with other biomarkers, the sVEGFR2/sUPAR protein combination was the most differentially expressed between worsening and non-worsening patients at H0 and H6 (p-value= 2.19e-10, and 2.36e-6, respectively, Mann-Whitney test).Interestingly, based on the new definition of sepsis (sepsis 3), 233 patients of our cohort were no longer considered as sepsis patients at enrolment (SOFA<2 at inclusion). Of these, 36 (15%) however developed organ dysfunction within the 72-hour study period. Again, the combination of sVEGFR2 and sUPAR protein levels was, at inclusion (H0), the best predictor of worsening (AUC=0.73, sensitivity= 0.92, NPV=0.95), compared with other biomarkers, CRP (AUC=0.57), Lactates (AUC=0.48), qSOFA (AUC=0.54) and PCT (AUC=0.61). Moreover, this performance was increased at H6 (AUC=0.79, sensitivity=0.94, NPV=0.98).

Conclusion:

While no vital signs nor clinical score or parameters can predict worsening of non-severe patients (SOFA<2 at inclusion), the expression of the sVEGFR2 alone or combined with sUPAR significantly predicts their progression to more severe status, within 72 hours after their admission to the emergency department. Such biomarker(s) could enhance early identification of severe patients in the ED for an appropriate and rapid management in order to decrease risks of deleterious outcome.

Background

An accurate assessment of septic patients at risk for deterioration  is challenging for clinicians in the emergency department (ED). Even so, early recognition of life-threatening conditions could result in improved outcomes.

Objectives

In this study, we aimed to evaluate the prognostic accuracy of procalcitonin (PCT) in patients with a suspected infection in the ED for predicting septic shock within 72 hours.

Patients and Methods

TRIAGE was a prospective, multicentre (14 sites in France and Belgium) observational study. Adult patients admitted in the ED with a suspected infection and at least 2 SIRS criteria were included. Blood samples were collected at 0, 6 and 24 hours after patient arrival. Accuracy and prognostic performances of biomarkers were assessed along with clinical variables such as age, sex, Charlson score, SOFA score, qSOFA, lactates, using logistic regression models. AUC under the ROC curve and their 95% confidence interval were computed using DeLong’s method. Predictive performances (sensitivity, specificity, Negative Predictive Value, Predictive Positive Value) were assessed using classification threshold optimized for high sensitivity.

Results

Serum PCT was measured in 462 ED patients with a suspected infection at 3 time points: H0 (patient’s admission), H6 and H24. Of these 462 patients, 12 (2.6%) developed a septic shock within the 72-hour study period. PCT and Lactates were the most differentially expressed markers at H0 between patients that developed septic shock or not (p-value<0.0001, Mann-Whitney test). Multivariate analysis showed that PCT and Lactates were independent prognostic factors of septic shock occurrence (IQR OR: 1.12, 95%CI: 1.05-1.19, P= 0.0005 and IQR OR: 1.47, 95%CI: 0.94-2.11, P = 0.049, respectively). Moreover, PCT protein level was, at inclusion (H0), the best predictor of septic shock (AUC=0.91 (0.85 - 0.98), sensitivity= 1, specificity=0.63, for a PCT threshold of 2.52ng/mL), compared with other biomarkers, Lactates (AUC=0.86 (0.78 - 0.93)), CRP (AUC=0.63 (0.48 - 0.77), or score SOFA (AUC=0.78 (0.65 – 0.91)).

Conclusion

Early prognostic assessment in sepsis in ED is essential to adjust therapeutic protocols, prevent deterioration and reduce mortality. In this context, PCT showed good performances in identifying patients at-risk of septic shock among patients suspected of infection at ED admission

Background: Early and appropriate recognition of Sepsis is essential both to start the treatment and improve the prognosis in the Emergency department (ED). Traditional biomarkers such as Procalcitonin (PCT) and C Reactive Protein (CRP) are of limited value in Sepsis with a significant overcost in the ED when done systematically. Lymphopenia, known to be Sepsis-associated in the ICU, has not yet been evaluated neither used by clinicians as a Sepsis biomarker in the ED while white blood cell (WBC) count is routinely performed in almost every ED patient as part of admission. In addition, lymphocyte count is a cheap and simple biological criterion, easy to implement into daily practice.

Objective: We investigated whether isolated or clinically associated severe lymphopenia in ED could be a marker of Sepsis.

Methods: We conducted a descriptive single-center study over a 1-year period in a teaching hospital. Adult patients admitted in the ED with severe lymphopenia (defined by a lymphocyte count < 0.5 G/L) were retrospectively analyzed. Patients with hematological or oncological diseases, HIV infection, hepato-cellular deficiency, immune depression and over 85 years old were excluded. Demographic data, reason for admission, qSOFA score, SIRS criteria, WBC, CRP and lactates were collected. Prevalence of Sepsis was evaluated by an independent adjudication committee based on clinical, biological and microbiological data available. Correlation between lymphopenia and Sepsis was assessed using a univariate analysis and multi-variable logistic regression.

Results: From January to December 2017, 953 patients were admitted in ED with severe lymphopenia and 245 were eligible (148 men; mean age 63±19 years). Sepsis was confirmed in 159 patients (65%) by the adjudication committee (bacterial: 60.4%, viral: 30.2%, other: 9.4%). Initially, only 61 patients (25%) were referred to ED for suspected infection. Of the infected patients, 18% had no diagnosis of infection after discharge from the ED. In “Sepsis group”, 116 patients (73%) met SIRS criteria and 46 patients (29%) had a qSOFA score ≥ 2 points at admission. There was a difference between both groups regarding CRP (Sepsis group: 109±133 vs No sepsis group: 46±62, p<0.0001) but not regarding the WBC count (Sepsis group: 10.8±5.5 vs No sepsis group: 10.9±5.9, p=0.82) and lactate measurements (Sepsis group: 2.3±1.6 vs No sepsis group: 2.5±1.7, p=0.38). In univariate analysis, WBC count (OR = 0.23; 95% CI [0.95 ; 1.04]; p=0.82) and CRP (OR = 1.01; 95% CI [1 ; 1.01]; p = 0.0003) did not appear as factors associated with sepsis unlike clinical criteria such as fever (OR = 10.95; 95% CI [5.39 ; 22.26]; p < 0.0001). In the multi-variable logistic model, SIRS criteria (OR = 3.45; 95% CI [1.41 ; 3.52]; p=0.0006) and fever (OR = 2.54; 95% CI [1.32 ; 8.66]; p=0.01) were identified as independent variables associated with Sepsis.

Conclusion: Prevalence of Sepsis is high in patients with severe lymphopenia regardless the reason of admission in ED and traditional biomarkers results. This easy measurable marker could be used routinely by emergency physicians to assist them in the early diagnosis of Sepsis.

Background.

Pain is a frequent reason for Emergency Department (ED) attendance. The aim of the POEM (Prescription Of analgesia in Emergency Medicine) study was to provide insight into the management of acute pain in the ED. We present a post-hoc comparison of the pain management provided to children (under 18yr) and adults.

Methods

The POEM study was an observational cross-sectional multicentre study in eleven UK EDs during 2015-2016. Patients of any age with a confirmed diagnosis of an isolated long bone fracture or dislocation were included. The patients were identified from each Trust’s clinical information system. Pain scoring and analgesia provision were compared to the Royal College of Emergency Medicine (RCEM) Best Practice Guidelines (2014).  The recruiting EDs were five major trauma centres (one adult only, two combined adult/paediatric, two paediatric only) and six trauma units (combined adult/paediatric). We present a post-hoc comparison of children vs adults after noting an apparent difference during planned analyses.

Statistical methods

All analyses were performed using the R Statistics program (R Foundation for Statistical Computing, Vienna, Austria) and standard statistical methodology.

Results

3196 (38%) of the 8346 patients in the POEM study were under 18 years old. The statistically significant differences between adults and children included: children received analgesia more often than adults (66% vs 53%), were more likely to have a pain score documented on initial assessment (55% vs 47%), and were more likely to have that initial assessment within 20 minutes of arrival (63% vs 50%). The analgesia provided was more likely to be deemed appropriate to the pain score than it was in the adult population (67% vs 53%). Therefore compliance with the RCEM guidance was significantly more likely for children than adults.

Interestingly, children were less likely than adults to have a reassessment of their pain score documented (7% vs 11%) and furthermore, if they arrived by ambulance children were less likely than adults (14% vs 40%) to have received pre-hospital analgesia.

Discussion

Overall, children had better pain management before arrival and in the ED than adults. However there is still need for improvement in the management of pain for patients of all ages in the ED, in line with recommendations from previous national RCEM audits of painful conditions. Our data does not yet explain why there is an observed difference in the pain management of children vs adults. We plan to review our data further to understand these differences in more depth.

Conclusions

We have found that children with an isolated long bone fracture/dislocation receive better pain management in the ED than adults. Until factors influencing pain management are better understood, we recommend that all clinicians seek to improve pain management for patients of all ages.

Ethical approval and informed consent

The study was approved by the Berkshire Research Ethics committee (REC 14/SC/0167) and waiver of consent was approved. Approval was gained from the Confidential Advisory Group (CAG 3-02(c)/2014) for collection of postcodes (required to calculate the index of multiple deprivation).

Background

Pain as the 5th vital sign was a concept introduced by the American Pain Society in 1996 and subsequently adopted in the UK.  It encourages the routine use of pain scoring in the acute setting in addition to standard vital signs, and remains an essential component of any pain management plan, where underassessment of pain risks the undertreatment of pain.  We investigated pain scoring in emergency departments (EDs) across the UK, in light of the 2014 Best Practice Guidelines by the Royal College of Emergency Medicine (RCEM).  These state that all patients should have their pain assessed within 20 minutes of arrival in ED, with a re-evaluation within 30 minutes of their first dose of analgesia.

Methods

The POEM (Prescription Of analgesia in Emergency Medicine) study was an observational cross-sectional  multicentre study in eleven UK EDs during 2015-2016. Patients with a confirmed diagnosis of an isolated long bone fracture or dislocation were included. Pain scoring and analgesia provision were compared to the RCEM Best Practice Guidelines.  Here we discuss the adequacy of pain scoring for the patients studied.

Statistical methods

All analyses were performed using the R Statistics program (R Foundation for Statistical Computing, Vienna, Austria) and standard statistical methodology.

Results

Data were collected from 8346 patients, with 50% (4160 patients) having a pain score recorded on arrival, and only 19% (768) of these then having a follow up pain score.  Thus only 9% of the total study population had both an initial and an early follow up pain score, despite RCEM guidance recommending both.  Interestingly, a higher proportion of patients received analgesia in ED (58%, n= 4845) than had pain scores documented.

Departments ranged from documenting admission pain scores in 7% of their patients to 99% (median 62%).  Repeat pain scores were even more rarely recorded with an inter-hospital range of 1-29% (median 11%).

We found that there is more chance of a pain score being recorded in younger patients; if the patient arrives by ambulance (rather than self-presentation); if the patient is seen by a consultant rather than an Emergency Nurse Practitioner; and if the patient is non-white or is from an area with a higher deprivation index.

Conclusions

Pain scoring is documented poorly in EDs, and the RCEM Best Practice Guidelines have not been fully adopted.   Patients should be asked about pain to guide appropriate analgesic management, and pain scores reassessed to determine its effectiveness.  The fact that more patients received analgesia than had a pain score documented may indicate that pain is being discussed but not recorded.  Routine pain scoring remains an important part of demonstrating suitable and effective pain relief for patients attending ED.

Ethical approval and informed consent

Approved by the Berkshire Research Ethics committee (REC 14/SC/0167), including waiver of consent; and by the Confidential Advisory Group (CAG 3-02(c)/2014) for collection of postcodes to calculate the index of multiple deprivation.

Background

It is estimated that 7 out of 10 attendances to the Emergency Department (ED) are related to pain and Royal College of Emergency Medicine (RCEM) national audits of painful conditions conclude that  wide variation in performance exists between EDs across the United Kingdom (UK).  The RCEM has published standards for adequate acute pain management and acknowledged that the current management of acute pain in UK EDs is inadequate and the evidence base is poor.  Good pain management has been shown to correlate with patient satisfaction and departmental factors, including ED crowding have also been associated with standards of pain management.  The POEM (Prescription Of analgesia in Emergency Medicine) study aimed to identify factors associated with management of acute pain in the ED.

Methods

A retrospective cross-sectional observational study was carried out in eleven UK EDs during 2015-2016. All patients with a confirmed diagnosis of an isolated long bone fracture or dislocation were included from a convenience sample.  Pain scoring and analgesia provision was compared to the RCEM Best Practice Guidelines and those who received adequate pain management were examined against those patients who did not receive adequate pain management looking at factors including: 4-hour target; time to assessment; re-attendance rate; total number of admissions from ED; staff:patient ratio; numbers of patients who left before being seen.

Logistic regression of the outcome variable was used to assess whether any of the explanatory variables were associated with individual and combined aspects of adequate pain management as per RCEM guidance.  All analyses were performed using the R Statistics program (R Foundation for Statistical Computing, Vienna, Austria) and standard statistical methodology.

Results

Data was collected from 8346 patients with a fracture or dislocation.  1346 patients received adequate pain management as per RCEM Best Practice Guidance and 2740 patients did not.  Considering the combination of timely assessment and provision of analgesia, improved 4-hour performance is associated with improved delivery of adequate pain management (OR = 1.027 (95% CI p < 0.001)).

Improved staff to patient ratios was significantly associated with the documentation of a pain score (OR = 1.095 (95% CI, p < 0.001)).  It would appear that departmental crowding (4-hour performance) does not alter the recording of pain scores.

Discussion

EDs are crowded when there is poor performance against the 4hr target.  At these times patients are less likely to be assessed within 20 minutes, and less likely to be given analgesia.  Our results indicate that as an ED becomes more crowded with reduced 4-hour performance and reduced staff to patient ratios, the probability of achieving the recommended pain management standards for patients reduces.  This is in keeping with our clinical experience of crowded departments. 

Conclusions

These findings contribute to the growing evidence base of the detrimental effects of ED crowding on patient care and further work is required in this important area.

Ethical approval and informed consent

Approved by the Berkshire Research Ethics committee (REC 14/SC/0167), including waiver of consent; and by the Confidential Advisory Group (CAG 3-02(c)/2014) for collection of postcodes.

Background

There is evidence to show that effective, early analgesia in patients with a fractured neck of femur (fNOF) improves final outcome. There has been debate as to the best form of analgesia, especially in the elderly, but fewer published studies as to why some fNOF patients are not given any analgesia. This analysis looks at staffing characteristics of the Emergency Department (ED) and demographic factors of patients with an isolated fractured NOF and their impact on measurement of pain scores and provision of analgesia.

Methods

The POEM (Prescription Of analgesia in Emergency Medicine) study was an observational, cross-sectional, multicentre study in eleven UK EDs during 2015-2016 of patients with a confirmed diagnosis of an isolated long bone fracture or dislocation. This is a post-hoc analysis of a subgroup of these patients who had a diagnosis of fNOF.  Logistic regression was used to examine patient and staffing factors which may impact on pain management. Patient factors included age, gender, ethnicity (white or non-white) and socio-economic status based on the index of multiple deprivation (IMD) whereby a higher score suggests greater deprivation. Departmental factors included time of the day and day of the week that the patient arrived in ED, arrival by ambulance, and crowding measures (ED staff:patient ratio and performance against the 4 hour target).

Statistical methods

All analyses were performed using the R Statistics program (R Foundation for Statistical Computing, Vienna, Austria) and standard statistical methodology, including logistic regression of those variables described above.

Results

A total of 861 of the 8146 patients in the POEM study had a diagnosis of fNOF, of whom the majority were female (72%) and the average age was 85 years (range 19-104). For patients with a fNOF when 4-hour performance was better, initial assessment of pain was more likely to be within 20 minutes (Odds Ratio (OR) 1.057, p<0.001). In addition, there was better documentation of pain scores with increasing staff:patient ratios (OR 1.052 p=0.0152). A higher IMD score (more socially deprived) was associated with better documentation of pain scores (odds ratio 1.041, p<0.001), but a lower likelihood of receiving analgesia (odds ratio 0.982, p=0.006). Non-white patients were more likely than white patients to receive any analgesia (OR 4.127, p=0.027)

As social deprivation (IMD) increases, there is less chance of a satisfactory outcome for overall pain management in ED, based upon RCEM guidance for a timely pain score (within 20 minutes) and provision of appropriate analgesia (OR 0.973, p<0.05)).

Conclusions

Both patient and staffing factors influenced provision of timely analgesia to patients with a fNOF. In particular, higher index of multiple deprivation had a negative impact on the provision of adequate analgesia. For patients with a fNOF, crowding was associated with better documentation of pain scores but increased likelihood of delay in initial assessment.  

Ethical approval and informed consent

Approved by the Berkshire Research Ethics committee (REC 14/SC/0167), including waiver of consent; and by the Confidential Advisory Group (CAG 3-02(c)/2014) for collection of postcodes to calculate the index of multiple deprivation.

Introduction

A key principle in the effective management of major incidents is triage, the process of prioritising patients on the basis of their clinical acuity.  Within both civilian and military practice in the UK, a two-stage approach to triage is employed, with primary triage at scene followed by a secondary triage process, allowing for a more detailed assessment of the patient using the Triage Sort.  Recent studies have demonstrated that existing methods of primary triage do not effectively identify patients in need of life-saving intervention (LSI), with high rates of under-triage, associated with increased mortality.  In order to improve the performance of the primary triage process, the MPTT-24 was developed, and this outperforms previous methods demonstrating lower rates of under-triage. 

To date, no studies have analysed the performance of the Triage Sort in the civilian setting.  The primary aim of this study was to compare the performance of the Triage Sort with the MPTT-24 and the NARU/Military Sieve at identifying patients in need of life-saving intervention.  

Methods

Retrospective database review of the Trauma Audit and Research Network (TARN) database for all adult patients (>18 years) between 2006-2014.  Patients were defined as Priority One if they received one or more LSIs from a previously defined list.  Patients were categorised using the Triage Sort, NARU/Military Sieve and the MPTT-24 using first recorded hospital physiology; only those with complete data were included.  Performance characteristics were evaluated using sensitivity and specificity and statistical analysis with a McNemar’s test.

Results

During the study period, 218,985 adult patients were included in the TARN database. 127,233 (58.1%) had complete data and were included: 55.6% male aged 61.4 (IQR 43.1-80.0 years), Injury Severity Score 9 (IQR 9-16), with 24,791 (19.5%) Priority One. The Triage Sort demonstrated the lowest performance of all triage tools at identifying need for LSI (sensitivity 15.7% (95%CIs 15.2-16.2) correlating with the highest rate of under-triage (84.3% (95%CIs 83.8-84.8), but had the greatest specificity (98.7% (95%CIs 98.6-98.8).

By comparison the NARU sieve had higher sensitivity (29.5% (95%CIs 28.9-30.1), but was outperformed by the MPTT-24 which demonstrated a statistically significant increase in performance (p<0.0001) with the greatest sensitivity (53.5% (95%CIs 52.9-54.1) and the lowest rate of under-triage (46.5% (95%CIs 45.9-47.1%). However, this increase in sensitivity comes at the expense of specificity, with the MPTT-24 having the lowest specificity (74.8% (95%CIs 74.6-75.1).

Conclusion

Within a civilian trauma registry population, the Triage Sort demonstrates the poorest performance at identifying patients in need of LSI.  Its use as a secondary triage tool should be reviewed, with an urgent need for further research to determine the optimum method of secondary triage.  

Introduction: Cardiovascular diseases are the second leading cause of death in France, and among these, the acute coronary syndromes (ACS) are the most common. To improve morbidity and mortality, the current challenge is to propose an adapted therapeutic (pre-hospital and / or in-hospital) as soon as possible. It is therefore essential to identify any ACS as early as medical regulation is in place. There is currently no ACS predictive score available for use in pre-hospital Emergency Medical Service (EMS) call center. The goal is to establish such a score and check its accuracy.

Material and method: Our prospective, observational and monocentric study was conducted from 1^st January to 31^th December 2016 at a French pre-hospital EMS call center, including any call for chest pain. Our population was randomized into two subgroups. In the derivation sample (two thirds), the univariate and multivariate analyses identified independent factors associated with ACS in regulation, and allowed the creation of a predictive score, based for each significant variable, on the beta coefficients of multivariate regression. In the validation sample (one-third), the discrimination was assessed by the area under the curve (AUC) and the calibration by the Hosmer-Lemeshow test.

Results: Of the 1367 included patients, 183 (13.4%) were diagnosed with ACS. The 7 variables significantly associated with the diagnosis of ACS in regulation are: male sex (OR 2.7, p < 0.001), age (OR 3.8 for the 43 – 57 year old and OR 4.5 for the > 57 year old, p = 0.0024), smoking (OR 2.2, p = 0.0023), typicality of the pain (OR 1.9, p = 0.0183), inaugural character of the pain (OR 1.7, p = 0.0238), presence of sweats (OR 1.9, p= 0.0057) and conviction of the physician (OR 2.9, p < 0.001). The AUC of this multivariate regression model is 0.81 and the Hosmer-Lemeshow "p" is 0.74 in the validation sample.

Conclusion: We were able to establish an ACS predictive score, usable for regulation of chest pain, the SCARE score, which has good discrimination and an excellent calibration in internal validation. This score allows stratifying risk of ACS, using epidemiological elements but also using the belief of the physician, whose Negative Predictive Value is excellent.

Background:

Ambulance services provide an increasing number of Emergency Medical Service health contacts, assessing people, treating at home or conveying them to further care. Prehospital research, needed to inform evidence based ambulance care, has unique ethical considerations due to urgency, time-limitations and the locations (home, ambulance) involved (Armstrong et al, 2017). Despite clinical research based in the ambulance setting increasing, there has been limited work assessing the impact of ethical considerations on this research. We sought to explore the ethical considerations perceived by paramedics involved with research in the prehospital ambulance setting.

Methods:

We employed a qualitative design using semi-structured interviews with paramedics who had enrolled at least one participant into a clinical trial in one large (around 750 thousand patient contacts per year) English regional ambulance service. Principlism, focussing on autonomy, beneficence and maleficence was used as a theoretical framework. Participants were asked a series of questions regarding their experiences in ambulance trials, their opinions on the impact of trials on both their own practice and on patients, and their views on ethical considerations of research more generally. The interview transcripts were digitally recorded, transcribed verbatim, coded by two researchers (SA and VHP) and analysed thematically using framework analysis.

Results:

15 interviews were completed (11 male, 4 female paramedics) during 2017 and 2018 over a period of 8 months. Participants had a range of experience as paramedic from 1 to 28 years’ service, with similar numbers trained either through a higher education route (n=8) or a more traditional ‘on the job’ training route (n=9). Initial analysis highlighted 4 main themes:

Consent

• Paramedics were comfortable with gaining consent and felt that it helped calm people during the incident.

Benefits and barriers

• Paramedics felt that research was important, helping to support improvements in patient care.
• Most felt that time was the biggest barrier to research with staff attitudes also being important.

Trial specific training

• All felt that the trial specific training helped although some mentioned that being able to see trial materials, including documentation, before attending the first patient would have been beneficial.

Reasons for participating (or not)

• The most often cited reason for participating in research was to improve evidence based practice for the good of the patient. Reasons for not participating included fears about lost skills when allocated to a non-intervention group or where there was no clear patient benefit.

Conclusion

Overall, paramedics interviewed were positive about taking part in research and were confident gaining consent. The main barriers to participation were time pressures and staff attitudes towards research. Most felt that clinical trials should continue within this field as they contributed to evidence based practice, vital for the development of ambulance services and for the benefit of patients. Paramedics that chose not to be involved in research were excluded, but for future studies it may be useful to include this group to understand why they choose to not participate.

Introduction

Paramedics make important decisions as to whether a patient requires transport to hospital, or can be discharged at scene.  Taking patients to the Emergency Department (ED) who do not require ED care or services is known as over-conveyance. Research shows that nearly 20% of patients brought to ED by ambulance, could be treated elsewhere.  In contrast, under-conveyance occurs when paramedics discharge a patient on-scene who required ED or hospital care. In 2016/17 in England, 5.2% of discharged patients re-contacted the ambulance service within 24 hours.  This study aims to investigate the accuracy of conveyance decisions made by on-scene paramedics. 

Methods

Six individual patient vignettes were created using linked ambulance, ED and GP data and used in an online survey to paramedics in Yorkshire.  Half the vignettes related to clinically necessary attendances at the ED and the other half were clinically unnecessary.  Vignettes were validated by a small expert panel.  Participants were asked to determine the appropriate conveyance decision and to explain the rationale behind their decisions using a free text box.They were also asked to predict hospital admission.

Results

143 paramedics undertook the survey and 858 vignettes were completed.  There was clear agreement between paramedics for transport decisions (kappa=0.63) and for admission prediction (kappa=0.86).  Overall accuracy was 0.69 (95% CI 0.66-0.73).  Paramedics were better at ‘ruling in’ the ED with sensitivity of 0.89 (95% CI 0.86-0.92).  The specificity of ‘ruling out’ the ED was 0.51 (95% CI 0.46-0.56).  Text comments were focussed on patient safety and risk aversion.

Discussion

 Paramedics make accurate conveyance decisions but are more likely to over-convey than under-convey, meaning that whilst decisions are safe they are not always appropriate.  Some risk-averse decisions were made due to patient and professional safety reasons. It is important that paramedics feel supported by the service to make non-conveyance decisions. Reducing over-conveyance is a potential method of reducing ED demand.