1. Risk of short-term neurologic complications in children who present a post-traumatic seizure following minor head trauma: systematic review and meta-analysis
Lorenzo Zanettoa, Liviana Da Dalta, Marco Daverioa, Joel Dunningb, Anna Chiara Frigoc, Lise Nigrovicd and Silvia Bressana
aDivision of Emergency Medicine, Department of Women’s and Children’s Health, University of Padova, Padova, Italy
bDepartment of Cardiothoracic Surgery, James Cook University Hospital, Middlesbrough, UK
cBiostatistics, Epidemiology and Public Health Unit, Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Padova, Italy and
dDivision of Emergency Medicine, Boston Children’s Hospital and Harvard Medical School, Boston, Massachusetts, USA
Although post-traumatic seizures (PTS) have been associated with an increased risk of traumatic brain injury (TBI), the risk in children with an immediate PTS and a normal Glasgow Coma Score (GCS) following blunt head trauma has not been rigorously evaluated. Our aim was to determine the frequency of short-term neurologic complications in children with PTS and a normal GCS following blunt head trauma.
We systematically searched PubMed, EMBASE, the Cochrane Library, Scopus, Web of Science, and ClinicalTrials.gov to identify studies reporting on children < 18 years with an immediate PTS and a GCS of 15 at the time of assessment. Two investigators independently reviewed identified articles for inclusion, assessed quality and extracted relevant data. Our main outcomes were the presence of any TBI on neuroimaging, the need for emergent neurosurgery or death due to head injury. We performed random effect meta-analyses and assessed heterogeneity across studies.
Of 9,956 studies screened, the 7 that met inclusion criteria included 66 202 head injured children of which 439 children (0.7%) had GCS of 15, an immediate PTS and underwent acute neuroimaging. The risk of any TBI on neuroimaging was 13.0% [95% confidence interval (CI) 4.0–26.1; I2 = 81%) although only 2.3% required emergent neurosurgery (n = 4 studies; 95% CI 0.0–9.9; I2 = 86%). No child died.
Children presenting with a PTS and a normal GCS following head trauma frequently have TBIs, although many do not require emergent neurosurgery. Clinicians should strongly consider either neuroimaging or prolonged observation for these children.
2. Short-stay unit hospitalisation in acutely admitted older internal medicine patients: randomised trial
Camilla Ströma, Lars Simon Rasmussenb, Anne-sofie Löwec, Anne Kathrine Lorentzend, Nicolai Lohsee, Kim Hvid Benn Madsend, Søren Wistisen Rasmussend and Thomas Andersen Schmidta
aDepartment of Emergency Medicine, Holbaek Sygehus, University of Copenhagen, Copenhagen, Denmark
bDepartment of Anaesthesiology, Centre of Head and Orthopedics, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark
cThe Defence Health Service, Danish Defence, Holbaek, Denmark
dDepartment of Emergency Medicine, Holbaek Hospital, Holbaek, Denmark and
eDepartment of Anaesthesiology, Nordsjaellands Hospital, Hilleroed, Denmark
The effect of hospitalisation in emergency department-based short-stay units (SSUs) has not been studied in older patients. We compared SSU-hospitalisation with standard care at an Internal Medicine Department (IMD) in acutely admitted older internal medicine patients.
We conducted a pragmatic randomised clinical trial. We randomly assigned patients aged 75 years or older, acutely admitted for an internal medicine disease, and assessed to be suitable for SSU-hospitalisation to SSU-hospitalisation or IMD-hospitalisation. SSU-hospitalisation was provided by a pragmatic ‘fast-track’-principle. The primary outcome was 90-day mortality. Secondary outcomes included adverse events, change in Lawton Instrumental Activities of Daily Living (iADL)-score within 90 days from admission, in-hospital length of stay, and unplanned readmissions within 30 days after discharge. All pre-planned analyses and interpretations were performed before the breaking of the randomisation code, but we included an evaluation of health care utilisation post-hoc (use and timing of diagnostic tests and treatments).
Between January 2015 and October 2016, 430 participants were randomised (median age 84 years in both groups). Ninety-day mortality was 22 (11%) in the SSU-group and 32 (15%) in the IMD-group (OR 0.66; 95%CI 0.37–1.18; P = 0.16). When comparing the SSU-group to the IMD-group, 16 (8%) versus 45 (21%) experienced at least one adverse event (OR 0.31; 95%CI 0.17–0.56; P > 0.001); 6 (3%) versus 35 (20%) experienced a reduction in iADL score within 90 days from admission (P > 0.001); median in-hospital length of stay was 73 h [IQR 36–147] versus 100 hours [IQR 47–169], (P > 0.001), and 26 (13%) versus 58 (29%) were readmitted (OR 0.37; 95%CI 0.22–0.61, P > 0.001). Fewer tests and treatments were applied in the SSU-group, and the time to chest x-ray, CT scans, point-of-care ultrasound, and physiotherapy was significantly shorter.
Mortality at 90 days after admission was not significantly lower in the SSU group, but SSU-hospitalisation was associated with a lower risk of adverse events, less functional decline, fewer readmissions, and shorter hospital stay. SSU hospitalisation may be preferable in acutely admitted older internal medicine patients.
[This paper has subsequently been published in full - https://academic.oup.com/ageing/article/47/6/810/5037285]
3. Is prehospital blood transfusion safe and effective? Systematic review and meta-analysis
Radboud University Medical Center, Nijmegen, The Netherlands
Life threatening hemorrhage accounts for 40% of mortality in trauma patients worldwide. Trauma is therefore the leading cause of death in patients aged 1–44 and in both civilian and military setting the most common cause of preventable death. After bleeding control is achieved, volume loss has to be restored. The positive effect of early in-hospital transfusion of blood or blood components in equal proportions (1:1:1) is already proven but the scientific proof for the efficacy in the prehospital setting is still absent as a result of lack of randomized control trials. Our objective was to prove that prehospital transfusion of blood products is safe and effective in patients with extensive blood loss
Four databases have been searched: CINAHL, Cochrane, EMBASE and Pubmed in the period 1988 till March 2018. Âfter manually removing duplicates, 2573 articles were screened on title and abstract by at least two reviewers. Articles were excluded when they complied with the following exclusion criteria: no blood or blood products administered, animal study, no prehospital setting and no original data. 240 articles were subsequently screened on full text. Finally, a total of 48 articles have been included. Data was analyzed by meta-analysis for mortality.
There was no significant difference in total mortality OR 1.09 [95% CI 0.89, 1.33] or 24-hour mortality OR 0.93 [95% CI[0.64, 1.34] for patients who received prehospital blood products, compared to standard care with crystalloids. A total of 4739 patients were transfused and three of them developed a complication which was possibly the result of the transfusion (0.07%). Thirteen included studies advice the use of fluid warmers before transfusion.
The administration of blood products in the prehospital environment is safe and seems feasible but proof of efficacy is lacking. Blood products have to be administered in equal proportions and heated before transfusion to minimize the risk to worsen hypothermia. Larger and randomized studies are required to demonstrate a statistically significant effect of the use of combined use of blood products.
4. Pathways of care for adult mental health emergency department attendances: analysis of routine data
Suzanne Ablarda, Richard Jacquesa, Colin O’Keeffea, Susan Croftb and Suzanne Masona
aUniversity of Sheffield, School of Health and Related Research, Sheffield, UK and
bSheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK
The management of patients with mental health problems in Emergency Departments (ED) has been of concern for some time. To identify where interventions could be most effectively targeted we must first understand when, why, and how mental health patients use EDs. Previous studies are largely based on data collected from single study sites using relatively small sample sizes, raising questions about the generalisability of these findings. We conducted a retrospective multi-site analysis of routinely available data to understand the pathways of care of mental health attendances through the ED and their outcomes.
Routine NHS patient level data for adult ED attendances across 18 EDs in Yorkshire and Humber (Y&H) for a one year period from January 2014 to December 2014 were analysed. Identification of mental health patients was based on identifying mental health string terms within both the ED diagnosis and presenting complaint fields (as ED diagnosis might not always classify a mental health crisis if there was also a physical health problem): (1) the first diagnosis variable was independently searched by two people for a comprehensive list of all mental health terms. Those patients with a mental health term in this variable then had their presenting complaint variable searched to identify any further terms; (2) identified terms were used to search all nine diagnosis variables and presenting complaint variable; (3) mental health attendances were split into three categories (psychiatric, overdose/self-harm, and anxiety) reflecting the differing clinical support required. Age, mode and time of arrival, number of investigations and treatments, length of ED stay and ED outcome were analysed. Comparative analyses of mental health and non-mental health patients were undertaken.
Of the 1 312 539 ED attendances, 3.1% (n = 39 594) were mental health related. Of the mental health patients 55.9% (n = 22 167) were categorised as self-harm/overdose; 31.8% (n = 12 597) psychiatric; and 12.2% (n = 4830) anxiety. Mental health patients were more likely to arrive by ambulance than non-mental health patients (OR 3.25, 95% CI 3.18–3.32), to arrive out-of-hours (OR 1.95, 95% CI 1.90–1.99), to leave the ED before treatment or refuse treatment (OR 2.94, 95% CI 2.85–304) and once in the ED had a significantly longer length of stay (median: mental health = 178 minutes vs non-mental health = 139 minutes, P < 0.001). 72.7% of the psychiatric sub-group received no investigations compared to 22.6% of the overdose / self-harm and 37.5% of the anxiety sub-groups. 51.2% of the psychiatric sub-group received no treatment or advice only compared to 22.8% of the overdose / self-harm and 34.5% of the anxiety sub-groups.
Our analysis showed mental health patients are placing a small but significant burden on emergency care services and are receiving poorer levels of care than other patients. Improving the availability of alternative mental health services in the community, particularly during the out-of-hours period, could improve outcomes for these patients. Also, increased training for ED and ambulance service staff in the identification of patients with mental health problems, with clear referral pathways for these patients, may improve mental health patient’s experiences of the ED.
5. Use of the prognostic biomarker soluble urokinase plasminogen activator receptor in the emergency department does not affect mortality: cluster-randomised clinical trial
Martin Schultza, Line Jee Hartmann Rasmussenb, Birgitte Nybo Jensenc, Lisbet Ravnd, Thomas Kallemosee, Theis Langef, Lars Køberg, Lars Simon Rasmussenh, Jesper Eugen-Olsenb and Kasper Karmark Iverseni
aDepartment of Cardiology, Herlev Hospital, Herlev, Denmark
bHvidovre Hospital, Clinical Research Center, Hvidovre, Denmark
cDepartment of Emergency Medicine, Bispebjerg Hospital, Copenhagen, Denmark
dDepartment of Emergency Medicine, Herlev Hospital, Herlev, Denmark
eClinical Research Center, Amager and Hvidovre hospital, Hvidovre, Denmark
fDepartment of Public Health, University of Copenhagen, Copenhagen, Denmark
gDepartment of Cardiology, Rigshospitalet, Copenhagen, Denmark
hDepartment of Anaesthesia, Rigshospitalet, Copenhagen, Denmark and
iDepartment of Cardiology, Herlev Hospital, Herlev, Denmark
Delays and crowding can increase mortality in emergency departments (EDs). Previous research has demonstrated that risk stratification can be strengthened using prognostic biomarkers, but the impact on patient prognosis is unknown. The aim of the TRIAGE III trial was to investigate whether the introduction of the prognostic and nonspecific biomarker: soluble urokinase plasminogen activator receptor (suPAR) in the emergency department reduces mortality in acutely admitted patients.
The TRIAGE III trial was a cluster-randomised interventional trial conducted at EDs in the Capital Region of Denmark. We included EDs with acute medical and surgical patients and no previous access to suPAR measurement in twelve cluster-periods of three weeks alternating between intervention and control. Patients were allocated to the intervention if they arrived in interventional periods, where suPAR measurement was routinely analysed at arrival. In the control periods, suPAR measurement was not performed. The primary endpoint was all-cause mortality ten months after admission. Secondary outcome was 30-day mortality.
We enrolled a consecutive cohort of 16 801 acutely admitted patients; all were included in the analyses. The intervention group consisted of 6 cluster periods with 8900 patients, and the control group consisted of 6 cluster periods with 7901 patients. The receiver operating characteristics curve analyses showed a prognostic ability of suPAR for 30-day and 10-month mortality corresponding to an area under the curve (95% CI) of 0.83 (0.81 to 0.84) and 0.80 (0.79 to 0.82), respectively. After a median follow-up of 362 days, death had occurred in 1241 patients (13.9%) in the intervention group and 1126 patients (14.3%) in the control group. The weighted Cox model found a hazard ratio of 0.97 (95% confidence interval, 0.89 to 1.07; P = 0.57). This result was consistent in all subgroups. Analysis of 30-day all-cause mortality showed similar results (HR, 0.98; 95% CI, 0.84 to 1.16; P = 0.84).
We detected no significant difference in mortality according to use of soluble urokinase plasminogen activator receptor as a prognostic biomarker in the emergency department.
6. Quality of bystander-performed chest compressions and prehospital advanced life support differently affect the outcomes of out-of-hospital cardiac arrests receiving bystander cardiopulmonary resuscitation: a propensity-matched observational study
Hideo Inabaa, Kurosaki Hisanoria, Yukihiro Watob and Yutaka Takeic
aDepartment of Circulatory Emergency and Resuscitation Science, Kanazawa University Graduate School of Medicine, Kanazawa, Japan
bDepartment of Emergency Medicine, Kanazawa Medical University, Uchinada, Japan and
cDepartment of Emergency Medical Sciences, Niigata University of Health and Welfare, Niigata, Japan
Ideally, prehospital advanced life support (ALS) should be preceded by good-quality bystander cardiopulmonary resuscitation including chest compressions (CCs). This study aimed to investigate how differently quality of bystander-performed CCs on emergency medical service (EMS) arrival and provision of ALS by EMS paramedics affect the outcomes of out-of-hospital cardiac arrests (OHCAs) receiving bystander CPR.
Prospective observational study with propensity-matched analyses was conducted in Ishikawa Prefecture, Japan during the period of 2012- 2016 after obtaining ethical approval from a local committee. Of 3088 adult ( ≥ 8 years) OHCAs receiving bystander-performed CCs on EMS arrival, CC qualities were determined in 3004 cases by two EMS personnel including at least one paramedic qualified for ALS on their arrival at the scene, according to the standard recommendations: depth (one-third of chest depth or > 5 cm), rate (100–120), proper position and adequate decompression. When two EMS personnel judged all criteria to be met, the quality was recorded as adequate. Primary outcome measure was neurologically favourable one-year survival. Secondary outcomes were sustained ( > 20 min) return of spontaneous circulation (ROSC) and one-month survival. After analysing the effects of good-quality CCs and ALS on outcomes and their interaction using a stepwise multivariable regression in all 3004 OHCAs, propensity-matching procedures were applied for CC quality and ALS provision. Propensity-matchings and stepwise multiple regression analyses included CC quality, ALS provision, time intervals, backgrounds and characteristics of OHCAs.
When analysed for all OHCAs, prehospital ALS (adjusted OR, 1.60; 95%CI, 1.28–1.99) but not good-quality of bystander-performed CCs was associated with sustained ROSC. Neither prehospital ALS nor good-quality CCs affected one-month survival. However, good-quality of bystander-performed CCs (3.33, 1.31–14.9) was associated with higher chances of neurologically favourable one-year survival, whereas prehospital ALS (0.20; 0.04–0.52) was associated with lower chances of the survival. There was no considerable interaction between good-quality CCs and prehospital ALS in any outcome measure. Propensity-matched analyses in CC quality- and ALS provision-matched pairs confirmed these findings. The incidence of sustained ROSC in cases with ALS provision was considerably higher than that without ALS provision for ALS provision-propensity-matched pair (N = 968 each, 29.0 vs 24.7%, 1.32; 1.07–1.64) The rate of neurologically favourable one-year survival in cases receiving good-quality CCs was considerably higher than that receiving poor quality for CC quality-matched pair (N = 636 each, 3.8 vs 1.6%, 2.81; 1.29–6.59). However, the rate of neurologically favourable survival in cases with ALS provision was lower than that without ALS provision for ALS provision-matched pair (N = 968 each, 1.5 vs 5.1%, 0.20; 0.10–0.39). The rate of 1-M survival did not considerably differ between the two groups in CC quality- or ALS provision-matched pair. Stepwise multivariable analyses for CC quality- and ALS provision-matched pairs revealed that witness status and initial ECG rhythm are other common and major factors associated with the outcomes.
Quality of bystander-performed CCs but not prehospital ALS provisionÂ is essential for neurologically favourable survival. Every effort should be made in a commuty to improve the quality of CCs before EMS arrival.
7. Critical care in the paediatric emergency department
Laura Melodya, Michael Bennettb, Carol Blackburna, Sean Walsha, Madeleine Niermeyerc and Michael Barretta
aConsultant in Paediatric Emergency Medicine, Department of Emergency Medicine, OLCHC, Dublin, Ireland
bConsultant in Emergency Medicine, Department of Emergency Medicine, OLCHC, Dublin, Ireland and
cResearch Nurse, National Children’s Research Centre, Dublin, Ireland
Critical care delivery has been defined by the American Medical Association and Current Procedural Terminology (CPT) as a physician’s direct delivery of medical care for a critically ill or critically injured patient. Critical care is decision making of high complexity to assess and support vital organ systems to prevent further life-threatening deterioration of the patient’s condition. To our knowledge, there has been no study to describe the prospective incidence of critical care, illness and injury in an Irish/European paediatric emergency department (PED) setting. Our primary objective was to establish the incidence of critical care delivery for illness or injury in an ED in Ireland’s largest tertiary academic paediatric hospital over 1 year.
A prospective observational study from 1/4/17 to 31/03/2018 with consecutive enrolment of patients who required critical care (e.g. ≥ 20ml/kg bolus), a critical care intervention (e.g. rapid sequence induction) or procedure (e.g. chest drain) in the ED delivered by emergency, anaesthetic, intensive care or surgical specialties. Data was triangulated from charts, electronic (ED information system and triage category 1 & 2/ PICU admissions) and registry (pharmacy and resuscitation logbooks) sources. Cases were screened for inclusion criteria. Inclusion criteria were determined by definitions from national and international guidelines (e.g. British Thoracic Society for asthma). Critical care interventions and procedures were agreed by consensus. The study was approved by the research and ethics committee.
The annual census was 37471 patients, with 285 (0.8%) receiving critical care. A total of 6280 (17% of annual census) were category 1 & 2 (Irish Children’s Triage System) cases. The mean age of presentation was 4 years with a range from 1 week to 18 years. Of the 285 cases, 92 (32%) were triaged as category 1 (most critical) upon presentation. A total of 153 (54%) patients presented during regular working hours (8:00 to 18:00) with 196 (69%) arriving by personal transportation. Significant pre-existing co-morbidities were present in 105 (37%) cases. The top 5 clinical presentations were sepsis 32 (11%), laryngotracheobronchitis 28 (10%), gastroenteritis 27 (9%), seizures 21 (7%), and bronchiolitis 21 (7%). Intraosseous vascular access was secured in 4 (3%), central venous access in 2 (1.5%), and peripheral intravenous access in 249 cases (87%). Interventions were performed by emergency, intensive care and surgical doctors. A total of 171 (60%) patients required 2 or more critical care interventions including 10 (3%) endotracheal intubations, 27 (9%) high flow O2/CPAP, and 2 (0.7%) chest drain insertions. There were 76 (27%) patient transfers to PICU and 4 (3%) out of hospital cardiac arrests pronounced dead in the ED and 2 (0.7%) children died within 7 days of admission to PICU.
This is the first study to describe prospectively the incidence of critical care delivered in an Irish tertiary paediatric emergency department. Encountering critical care in the PED translated to an intensive care admission in only 27% of cases. This study informs the provision of local education (e.g. simulation), departmental manpower and hospital-wide critical care provision.
8. Prevalence, follow-up and risk factors of incidental findings on trauma computed tomography scans: cross-sectional study at an urban level one trauma center
Wei-An Lai, Pang-Hsu Liu, Ming-Jen Tsai, and Ying Chieh Huang
Chiayi Christian Hospital, Chiayi City, Taiwan
Computed tomography (CT) is increasingly used to evaluate trauma patients and substantial numbers of incidental findings (IF) have been reported. Although these IF are often benign, their presence must always be communicated to patients. Clinically relevant IF needs referral or even consultation. This study was designed to evaluate the prevalence of IF revealed by trauma CT and status of communication. We also evaluate patient characteristics associated with the presence of IF.
This is a retrospective, cross-sectional study. We collected data from all patients who received CT scans for trauma evaluation at our Emergency Department in 2016. Duplicated scans and patients with missing data were excluded. Official CT reports were examined and basic demographics were reviewed. Scans with IF prompted detailed review of medical records to determine the clinical significance and the follow-up. Incidental findings were divided into three categories: category I (potentially severe condition, in-time diagnostic workup and management are required), category II (not urgent, needs follow-up), and category III (of minor concern, required no specific follow-up). Prevalence and status of follow-up of IF were reported. Multivariable logistic regression models were fitted to determine whether certain patient characteristics were associated with the presence of IF. A P-value < 0.05 was considered statistically significant.
There were 4,173 CT scans performed for trauma evaluation. After excluding 20 duplicated scans and 61 scans with missing data, 4,092 scans were enrolled. Incidental findings were identified in 649 (15.9%) scans; 13 (2.0%) of them were category I, 306 (47.1%) were category II, and 330 (50.8%) were category III. Patients with IF were older than those without (P > 0.001); however, no sex difference was found (P = 0.667). Nearly two thirds (61.5%) of scans were done for the head; however, the abdomen had the highest prevalence of IF (26.2%), followed by chest (20.2%), head (15.3%), neck (10.5%) and extremity (2.6%). Although 46% patients with category I IF received consultation of corresponding specialists, no documentation about IF could be found for the majority of patients with IF (category I 31%, category II 91.9%, category III 97.0%). Every year of Increasing age was independently associated with a higher prevalence of IF (odds ratio 1.019; 95% CI 1.015–1.024). A higher risk (odds ratio 1.035; 95% CI 1.028–1.041) of aging was observed in Category I & II IF.
Incidental findings were identified in 15.9% of our trauma CT scanning. These lesions were common in the abdomen and chest and showed an increased prevalence with increasing age. Follow-up was poor, even for potentially serious findings. A delicate effort of communicating to the patient, referring to the primary care physician or corresponding specialist, and documentation is beneficial to both the patient and the healthcare provider.
9. Unplanned Readmission prevention by Geriatric Emergency Network for Transitional care (URGENT): a single centre quasi-experimental study
Els Devriendta, Pieter Heerenb, Steffen Fieuwsc, Nathalie Wellensd, Mieke Deschodte, Johan Flamaingf, Marc Sabbeg, Koen Milisenb
aDepartment of Geriatric Medicine / Department of Public Health and Primary Care, Academic Centre for Nursing and Midwifery, University Hospitals Leuven / KU Leuven, Leuven, Belgium
bDepartment of Public Health and Primary Care, Academic Centre for Nursing and Midwifery / Department of Geriatric Medicine, KU Leuven / University Hospitals Leuven, Leuven, Belgium
cI-Biostat, Interuniversity Institute for Biostatistics and Statistical Bioinformatics, KU Leuven, Leuven, Belgium
dPublic Health and Social Affairs Department / Department of Public Health and Primary Care, Academic Centre for Nursing and Midwifery, Government Canton Vaud, Lausanne / KU Leuven, Lausanne, Switzerland
eDepartment of Chronic Diseases, Metabolism and Ageing, Gerontology and Geriatrics / Department of Public Health, Nursing Science, KU Leuven / University of Basel, Leuven, Belgium
fDepartment of Geriatric Medicine / Department of Chronic Diseases, Metabolism and Ageing, Gerontology and Geriatrics, University Hospitals Leuven / KU Leuven, Leuven, Belgium and
gDepartment of Emergency Medicine, University Hospitals Leuven / Department of Public Health and Primary Care, Emergency Medicine, KU Leuven / University Hospitals Leuven, Leuven, Belgium
International guidelines recommend adapting the classic emergency department (ED) management model to the needs of older adults in order to ameliorate post-ED outcomes among this vulnerable group. To improve the care for older ED patients and specifically prevent unplanned ED readmissions, the URGENT care model was developed. The study aim was evaluating the effectiveness of the URGENT care model.
A prospective single centre quasi-experimental study (sequential design with two cohorts, recruited from 1/12/2014 to 31/5/2015 and from 15/10/2015 to 31/5/2016, respectively) was conducted in the ED of University Hospitals Leuven. Dutch-speaking, community-dwelling ED patients aged 70 years or older were eligible for enrolment. Patients in the control cohort received usual ED care. Patient in the intervention cohort received the URGENT care model. The URGENT care model is a nurse-led, comprehensive geriatric assessment based care model in the ED with geriatric follow-up after ED discharge. The interRAI ED Screener and clinical judgement of ED staff were used to identify patients at risk for unplanned ED readmission. A geriatric nurse was available during office hours to conduct CGA in at risk patients. Subsequently, a personalized interdisciplinary care plan was made. Discharged at risk patients were offered case manager follow-up. Hospitalized at risk patients received follow-up on a geriatric ward or by the inpatient geriatric consultation team if considered necessary. The effectiveness of the URGENT care model was measured primarily on 90-day unplanned ED readmission rate. Secondary outcome measures were hospitalization rate, ED length of stay (ED LOS), in-hospital length of stay, 90-day higher level of care, 90-day functional decline and 90-day mortality. The required sample size was 751 patients per cohort, making a total of 1502 patients. Cause-specific hazard-ratios, relative risks, logistic regression and a lognormal model were used when appropriate. In all analyses, a propensity model was used to handle the potential difference in patient mix between the cohorts. Bonferroni correction was applied if considered relevant.
Unplanned ED readmission occurred in 170 of 768 (22.1%) control cohort (CC) patients and in 205 of 857 (23.9%) intervention cohort (IC) patients (P =.11). Statistically significant secondary outcomes were ED LOS (CC: 19.1 versus IC: 12.7 hours respectively; P =.0003), hospitalization rate (CC: 67.0% versus IC: 70.0%; P =.0.0026) and functional decline (CC: 21.5% versus IC: 26.6%; P =.0.023). ED LOS and hospitalization rate remained statistically significant after Bonferroni correction.
The URGENT care model shortened ED LOS and increased the hospitalization rate, but did not prevent unplanned ED readmissions.
10. The kids are alright: reporting the differences from a post-hoc analysis of the observational multicentre cross-sectional Prescription Of analgesia in Emergency Medicine (POEM) study
Sarah Wilsona, Liza Keatingb, Sally Beerc, Jane Quinland, James Sheehane, Jack Daintyf, Melanie Darwentg and Martyn Ezrad
aEmergency Department, Frimley Health NHS Foundation Trust, Slough, UK
bEmergency Department, Royal Berkshire NHS Foundation Trust, Reading, UK
cNIHR Emergency Department Research, Oxford University Hospitals NHS Foundation Trust, Oxford, UK
dNuffield Department of Anaesthetics, Oxford University Hospitals NHS Foundation Trust, Oxford, UK
eDepartment of Anaesthetics, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK
fDepartment of Statistics, Norwich Medical School, University of East Anglia, Norwich, UK and
gEmergency Department, Oxford University Hospitals NHS Foundation Trust, Oxford, UK
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.
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. All analyses were performed using the R Statistics program (R Foundation for Statistical Computing, Vienna, Austria) and standard statistical methodology.
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.
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. We found that children with an isolated long bone fracture/dislocation received 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.