Accuracy of telephone triage for predicting adverse outcomes in suspected COVID-19: an observational cohort study

OBJECTIVE

To assess accuracy of telephone triage in identifying need for emergency care among those with suspected COVID-19 infection and identify factors which affect triage accuracy.

DESIGN

Observational cohort study.

SETTING

Community telephone triage provided in the UK by Yorkshire Ambulance Service NHS Trust (YAS).

PARTICIPANTS

40 261 adults who contacted National Health Service (NHS) 111 telephone triage services provided by YAS between 18 March 2020 and 29 June 2020 with symptoms indicating COVID-19 infection were linked to Office for National Statistics death registrations and healthcare data collected by NHS Digital.

OUTCOME

Accuracy of triage disposition was assessed in terms of death or need for organ support up to 30 days from first contact.

RESULTS

Callers had a 3% (1200/40 261) risk of serious adverse outcomes (death or organ support). Telephone triage recommended self-care or non-urgent assessment for 60% (24 335/40 261), with a 1.3% (310/24 335) risk of adverse outcomes. Telephone triage had 74.2% sensitivity (95% CI: 71.6 to 76.6%) and 61.5% specificity (95% CI: 61% to 62%) for the primary outcome. Multivariable analysis suggested respiratory comorbidities may be overappreciated, and diabetes underappreciated as predictors of deterioration. Repeat contact with triage service appears to be an important under-recognised predictor of deterioration with 2 contacts (OR 1.77, 95% CI: 1.14 to 2.75) and 3 or more contacts (OR 4.02, 95% CI: 1.68 to 9.65) associated with false negative triage.

CONCLUSION

Patients advised to self-care or receive non-urgent clinical assessment had a small but non-negligible risk of serious clinical deterioration. Repeat contact with telephone services needs recognition as an important predictor of subsequent adverse outcomes.

Prognostic accuracy of eight triage scores in suspected COVID-19 in an Emergency Department low-income setting: An observational cohort study

Introduction

Previous studies deriving and validating triage scores for patients with suspected COVID-19 in Emergency Department settings have been conducted in high- or middle-income settings. We assessed eight triage scores' accuracy for death or organ support in patients with suspected COVID-19 in Sudan.

Methods

We conducted an observational cohort study using Covid-19 registry data from eight emergency unit isolation centres in Khartoum State, Sudan. We assessed performance of eight triage scores including: PRIEST, LMIC-PRIEST, NEWS2, TEWS, the WHO algorithm, CRB-65, Quick COVID-19 Severity Index and PMEWS in suspected COVID-19. A composite primary outcome included death, ventilation or ICU admission.

Results

In total 874 (33.84 %, 95 % CI:32.04 % to 35.69 %) of 2,583 patients died, required intubation/non-invasive ventilation or HDU/ICU admission . All risk-stratification scores assessed had worse estimated discrimination in this setting, compared to studies conducted in higher-income settings: C-statistic range for primary outcome: 0.56-0.64. At previously recommended thresholds NEWS2, PRIEST and LMIC-PRIEST had high estimated sensitivities (≥0.95) for the primary outcome. However, the high baseline risk meant that low-risk patients identified at these thresholds still had a between 8 % and 17 % risk of death, ventilation or ICU admission.

Conclusion

None of the triage scores assessed demonstrated sufficient accuracy to be used clinically. This is likely due to differences in the health care system and population (23 % of patients died) compared to higher-income settings in which the scores were developed. Risk-stratification scores developed in this setting are needed to provide the necessary accuracy to aid triage of patients with suspected COVID-19.

Reducing unplanned hospital admissions from care homes: a systematic review

Background

Care homes predominantly care for older people with complex health and care needs, who are at high risk of unplanned hospital admissions. While often necessary, such admissions can be distressing and provide an opportunity cost as well as a financial cost.

Objectives

Our objective was to update a 2014 evidence review of interventions to reduce unplanned admissions of care home residents. We carried out a systematic review of interventions used in the UK and other high-income countries by synthesising evidence of effects of these interventions on hospital admissions; feasibility and acceptability; costs and value for money; and factors affecting applicability of international evidence to UK settings.

Data sources

We searched the following databases in December 2021 for studies published since 2014: Cochrane Central Register of Controlled Trials and Cochrane Database of Systematic Reviews; Cumulative Index to Nursing and Allied Health Literature; Health Management Information Consortium; Medline; PsycINFO; Science and Social Sciences Citation Indexes; Social Care Online; and Social Service Abstracts. 'Grey' literature (January 2022) and citations were searched and reference lists were checked.

Methods

We included studies of any design reporting interventions delivered in care homes (with or without nursing) or hospitals to reduce unplanned hospital admissions. A taxonomy of interventions was developed from an initial scoping search. Outcomes of interest included measures of effect on unplanned admissions among care home residents; barriers/facilitators to implementation in a UK setting and acceptability to care home residents, their families and staff. Study selection, data extraction and risk of bias assessment were performed by two independent reviewers. We used published frameworks to extract data on intervention characteristics, implementation barriers/facilitators and applicability of international evidence. We performed a narrative synthesis grouped by intervention type and setting. Overall strength of evidence for admission reduction was assessed using a framework based on study design, study numbers and direction of effect.

Results

We included 124 publications/reports (30 from the UK). Integrated care and quality improvement programmes providing additional support to care homes (e.g. the English Care Homes Vanguard initiatives and hospital-based services in Australia) appeared to reduce unplanned admissions relative to usual care. Simpler training and staff development initiatives showed mixed results, as did interventions aimed at tackling specific problems (e.g. medication review). Advance care planning was key to the success of most quality improvement programmes but do-not-hospitalise orders were problematic. Qualitative research identified tensions affecting decision-making involving paramedics, care home staff and residents/family carers. The best way to reduce end-of-life admissions through access to palliative care was unclear in the face of inconsistent and generally low-quality evidence.

Conclusions

Effective implementation of interventions at various stages of residents' care pathways may reduce unplanned admissions. Most interventions are complex and require adaptation to local contexts. Work at the interface between health and social care is key to successful implementation.

Limitations

Much of the evidence identified was of low quality because of factors such as uncontrolled study designs and small sample size. Meta-analysis was not possible.

Future work

We identified a need for improved economic evidence and the evaluation of integrated care models of the type delivered by hospital-based teams. Researchers should carefully consider what is realistic in terms of study design and data collection given the current context of extreme pressure on care homes.

Study registration

This study is registered as PROSPERO database CRD42021289418.

Funding

This project was funded by the National Institute for Health and Care Research (NIHR) Health and Social Care Delivery Research programme (award number NIHR133884) and will be published in full in Health and Social Care Delivery Research; Vol. 11, No. 18. See the NIHR Journals Library website for further project information.

Training and testing of a gradient boosted machine learning model to predict adverse outcome in patients presenting to emergency departments with suspected covid-19 infection in a middle-income setting

COVID-19 infection rates remain high in South Africa. Clinical prediction models may be helpful for rapid triage, and supporting clinical decision making, for patients with suspected COVID-19 infection. The Western Cape, South Africa, has integrated electronic health care data facilitating large-scale linked routine datasets. The aim of this study was to develop a machine learning model to predict adverse outcome in patients presenting with suspected COVID-19 suitable for use in a middle-income setting. A retrospective cohort study was conducted using linked, routine data, from patients presenting with suspected COVID-19 infection to public-sector emergency departments (EDs) in the Western Cape, South Africa between 27th August 2020 and 31st October 2021. The primary outcome was death or critical care admission at 30 days. An XGBoost machine learning model was trained and internally tested using split-sample validation. External validation was performed in 3 test cohorts: Western Cape patients presenting during the Omicron COVID-19 wave, a UK cohort during the ancestral COVID-19 wave, and a Sudanese cohort during ancestral and Eta waves. A total of 282,051 cases were included in a complete case training dataset. The prevalence of 30-day adverse outcome was 4.0%. The most important features for predicting adverse outcome were the requirement for supplemental oxygen, peripheral oxygen saturations, level of consciousness and age. Internal validation using split-sample test data revealed excellent discrimination (C-statistic 0.91, 95% CI 0.90 to 0.91) and calibration (CITL of 1.05). The model achieved C-statistics of 0.84 (95% CI 0.84 to 0.85), 0.72 (95% CI 0.71 to 0.73), and 0.62, (95% CI 0.59 to 0.65) in the Omicron, UK, and Sudanese test cohorts. Results were materially unchanged in sensitivity analyses examining missing data. An XGBoost machine learning model achieved good discrimination and calibration in prediction of adverse outcome in patients presenting with suspected COVID19 to Western Cape EDs. Performance was reduced in temporal and geographical external validation.

External validation of triage tools for adults with suspected COVID-19 in a middle-income setting: an observational cohort study

BACKGROUND

Tools proposed to triage ED acuity in suspected COVID-19 were derived and validated in higher income settings during early waves of the pandemic. We estimated the accuracy of seven risk-stratification tools recommended to predict severe illness in the Western Cape, South Africa.

METHODS

An observational cohort study using routinely collected data from EDs across the Western Cape, from 27 August 2020 to 11 March 2022, was conducted to assess the performance of the PRIEST (Pandemic Respiratory Infection Emergency System Triage) tool, NEWS2 (National Early Warning Score, version 2), TEWS (Triage Early Warning Score), the WHO algorithm, CRB-65, Quick COVID-19 Severity Index and PMEWS (Pandemic Medical Early Warning Score) in suspected COVID-19. The primary outcome was intubation or non-invasive ventilation, death or intensive care unit admission at 30 days.

RESULTS

Of the 446 084 patients, 15 397 (3.45%, 95% CI 34% to 35.1%) experienced the primary outcome. Clinical decision-making for inpatient admission achieved a sensitivity of 0.77 (95% CI 0.76 to 0.78), specificity of 0.88 (95% CI 0.87 to 0.88) and the negative predictive value (NPV) of 0.99 (95% CI 0.99 to 0.99). NEWS2, PMEWS and PRIEST scores achieved good estimated discrimination (C-statistic 0.79 to 0.82) and identified patients at risk of adverse outcomes at recommended cut-offs with moderate sensitivity (>0.8) and specificity ranging from 0.41 to 0.64. Use of the tools at recommended thresholds would have more than doubled admissions, with only a 0.01% reduction in false negative triage.

CONCLUSION

No risk score outperformed existing clinical decision-making in determining the need for inpatient admission based on prediction of the primary outcome in this setting. Use of the PRIEST score at a threshold of one point higher than the previously recommended best approximated existing clinical accuracy.

Major trauma presentations and patient outcomes in English hospitals during the COVID-19 pandemic: An observational cohort study

BACKGROUND

Single-centre studies suggest that successive Coronavirus Disease 2019 (COVID-19)-related "lockdown" restrictions in England may have led to significant changes in the characteristics of major trauma patients. There is also evidence from other countries that diversion of intensive care capacity and other healthcare resources to treating patients with COVID-19 may have impacted on outcomes for major trauma patients. We aimed to assess the impact of the COVID-19 pandemic on the number, characteristics, care pathways, and outcomes of major trauma patients presenting to hospitals in England.

METHODS AND FINDINGS

We completed an observational cohort study and interrupted time series analysis including all patients eligible for inclusion in England in the national clinical audit for major trauma presenting between 1 January 2017 and 31 of August 2021 (354,202 patients). Demographic characteristics (age, sex, physiology, and injury severity) and clinical pathways of major trauma patients in the first lockdown (17,510 patients) and second lockdown (38,262 patients) were compared to pre-COVID-19 periods in 2018 to 2019 (comparator period 1: 22,243 patients; comparator period 2: 18,099 patients). Discontinuities in trends for weekly estimated excess survival rate were estimated when lockdown measures were introduced using segmented linear regression. The first lockdown had a larger associated reduction in numbers of major trauma patients (-4,733 (21%)) compared to the pre-COVID period than the second lockdown (-2,754 (6.7%)). The largest reductions observed were in numbers of people injured in road traffic collisions excepting cyclists where numbers increased. During the second lockdown, there were increases in the numbers of people injured aged 65 and over (665 (3%)) and 85 and over (828 (9.3%)). In the second week of March 2020, there was a reduction in level of major trauma excess survival rate (-1.71%; 95% CI: -2.76% to -0.66%) associated with the first lockdown. This was followed by a weekly trend of improving survival until the lifting of restrictions in July 2020 (0.25; 95% CI: 0.14 to 0.35). Limitations include eligibility criteria for inclusion to the audit and COVID status of patients not being recorded.

CONCLUSIONS

This national evaluation of the impact of COVID on major trauma presentations to English hospitals has observed important public health findings: The large reduction in overall numbers injured has been primarily driven by reductions in road traffic collisions, while numbers of older people injured at home increased over the second lockdown. Future research is needed to better understand the initial reduction in likelihood of survival after major trauma observed with the implementation of the first lockdown.

LMIC-PRIEST: Derivation and validation of a clinical severity score for acutely ill adults with suspected COVID-19 in a middle-income setting

BACKGROUND

Uneven vaccination and less resilient health care systems mean hospitals in LMICs are at risk of being overwhelmed during periods of increased COVID-19 infection. Risk-scores proposed for rapid triage of need for admission from the emergency department (ED) have been developed in higher-income settings during initial waves of the pandemic.

METHODS

Routinely collected data for public hospitals in the Western Cape, South Africa from the 27th August 2020 to 11th March 2022 were used to derive a cohort of 446,084 ED patients with suspected COVID-19. The primary outcome was death or ICU admission at 30 days. The cohort was divided into derivation and Omicron variant validation sets. We developed the LMIC-PRIEST score based on the coefficients from multivariable analysis in the derivation cohort and existing triage practices. We externally validated accuracy in the Omicron period and a UK cohort.

RESULTS

We analysed 305,564 derivation, 140,520 Omicron and 12,610 UK validation cases. Over 100 events per predictor parameter were modelled. Multivariable analyses identified eight predictor variables retained across models. We used these findings and clinical judgement to develop a score based on South African Triage Early Warning Scores and also included age, sex, oxygen saturation, inspired oxygen, diabetes and heart disease. The LMIC-PRIEST score achieved C-statistics: 0.82 (95% CI: 0.82 to 0.83) development cohort; 0.79 (95% CI: 0.78 to 0.80) Omicron cohort; and 0.79 (95% CI: 0.79 to 0.80) UK cohort. Differences in prevalence of outcomes led to imperfect calibration in external validation. However, use of the score at thresholds of three or less would allow identification of very low-risk patients (NPV ≥0.99) who could be rapidly discharged using information collected at initial assessment.

CONCLUSION

The LMIC-PRIEST score shows good discrimination and high sensitivity at lower thresholds and can be used to rapidly identify low-risk patients in LMIC ED settings.

An Unusual cause of Endocarditis

A man in his 20s attended the emergency department with three days of fever, headache, reduced appetite and a sore throat. COVID-19 point-of-care test was negative. Blood cultures grew a gram-negative coccobacillus, Neisseria elongata Following an episode of confusion, MRI head revealed septic emboli. Prolapse of the mitral valve with regurgitation was noted on echocardiography. Infection was found to have originated from multiple dental caries and treatment required a combination of dental extraction, prolonged antibiotic therapy and surgery for mitral valve repair.N. elongata is part of the normal oropharyngeal flora but is also a rare cause of endocarditis. There are no established treatment guidelines for endocarditis of this aetiology. N. elongata endocarditis may present atypically, with a murmur only developing several days later. 'Classical' stigmata should not be relied on to make a diagnosis. N. elongata predominantly affects the left side of the heart and predisposes to embolic events.

1482 Prognostic accuracy of triage tools for adults with suspected COVID-19 in a middle-income setting

Aims, Objectives and Background

Uneven vaccination in low- and middle-income settings and less resilient health care provision mean that emergency health care systems may still be at risk of being overwhelmed during periods of increased COVID-19 infection. Risk stratification tools proposed to allow rapid triage of need for admission in ED settings have almost exclusively been developed and validated in high-income settings during early waves of the pandemic.

Our study aimed to estimate the accuracy of risk-stratification tools recommended to predict severe illness in adults with suspected COVID-19 infection in the Western Cape of South Africa.

Method and Design

An observational cohort study using routinely electronically collected clinical information in all state-run hospitals in the Western Cape between 27th August 2020 and 11th March 2022 was conducted to assess performance of the PRIEST tool, NEWS2, the WHO algorithm, CRB-65, TEWS, Quick Covid Severity Index and PMEWS in patients with suspected COVID-19. The primary outcome was death, respiratory support or ICU admission.

Abstract 1482 Figure 1

Performance of tools predicting composite primary outcome for total study period

Abstract 1482 Figure 2

Performance of tools predicting composite primary outcome for the Omicron period

Abstract 1482 Table 1

Triage tool diagnostic accuracy statistics (95% CI) for predicting any adverse outcome (entire study period)

ToolN*C-statisticThresholdN (%) above thresholdSensitivitySpecificityPPVNPVCRB-65432,5840.70(0.70, 0.71)>0102,964 (23.8%)0.61(0.61, 0.61)0.78(0.77, 0.78)0.09(0.09, 0.09)0.98(0.98, 0.98)NEWS2433,1010.80(0.79, 0.80)>1178835 (41.3%)0.83(0.83, 0.83)0.6(0.6,0.6)0.07(0.07–0.07)0.99(0.99, 0.99)PMEWS438,8100.79(0.79, 0.79)>2199,386 (45.4%)0.85(0.85, 0.85)0.56(0.56, 0.56)0.06(0.06, 0.07)0.99 (0.99,0.99)PRIEST438,8800.82(0.82, 0.82)>4158,893 (36.2%)0.83(0.83, 0.83)0.65 (0.65,0.66)0.08(0.08, 0.08)0.99(0.99, 0.99)WHO437,8500.71(0.71, 0.72)>0235,775 (53.8%)0.82(0.81, 0.82)0.47(0.47, 0.47)0.05(0.05, 0.05)0.99(0.99, 0.99)TEWS432,6120.72(0.71, 0.72)>2134,097 (31%)0.62(0.62, 0.62)0.70(0.70, 0.70)0.07(0.07, 0.07)0.98(0.98, 0.98)Quick COVID446,0880.70(0.69, 0.70)>335,145 (7.9%)0.33(0.33, 0.33)0.93(0.93, 0.93)0.14(0.14, 0.14)0.98(0.98, 0.98)

*Patients with <3 parameters were excluded from analysis when estimating performance

Abstract 1482 Table 2

Triage tool diagnostic accuracy statistics (95% CI) for predicting any adverse outcome (Omicron period)

ToolN*C-statisticThresholdN (%) above thresholdSensitivitySpecificityPPVNPVCRB-65136,9610.69(0.68, 0.70)>031,373 (22.9%)0.59(0.59, 0.59)0.78(0.78, 0.78)0.05(0.05, 0.05)0.99(0.99, 0.99)NEWS2137,1250.77(0.76, 0.78)>176,183 (55.6%)0.87(0.87, 0.87)0.45(0.45, 0.45)0.03(0.03, 0.03)0.99(0.99, 0.99)PMEWS138,9540.76(0.75, 0.76)>259,876 (43.1%)0.80(0.80, 0.80)0.58(0.58, 0.58)0.04(0.04, 0.04)0.99(0.99, 0.99)PRIEST158,8930.78(0.77, 0.79)>446,529 (33.5%)0.75(0.75, 0.75)0.67(0.67, 0.67)0.04(0.04, 0.04)0.99(0.99, 0.99)WHO138,6660.62(0.61, 0.63)>072,599 (52.4%)0.70(0.70, 0.70)0.48(0.48, 0.48)0.03(0.03, 0.03)0.99(0.99, 0.99)TEWS136,9670.73(0.72, 0.74)>239,509 (28.8%)0.64(0.64, 0.64)0.72(0.72, 0.72)0.04(0.04, 0.04)0.99(0.99, 0.99)Quick COVID1405200.61(0.60, 0.63)>38,210(6.4%)0.17(0.17, 0.17)0.94(0.94, 0.94)0.06(0.06, 0.06)0.98(0.98, 0.98)

*Patients with <3 parameters were excluded from analysis when estimating performance

Results and Conclusion

Of the 446,084 patients, 15,397 patients (3.45%, 95% CI:34% to 35.1%) experienced the primary outcome. Figure 1 presents the ROC curves for the triage tools for the total study period and figure 2 for the period of the Omicron wave. NEWS2, PMEWS, PRIEST tool and WHO algorithm identified patients at risk of adverse outcomes at recommended cut-offs with moderate sensitivity (>0.8) and specificity ranging from 0.47 (NEWS2) to 0.65 (PRIEST tool). The low prevalence of the primary outcome, especially in the Omicron period, meant use of these tools would have more than doubled admissions with only a small reduction in risk of false negative triage.

Triage tools developed specifically in low- and middle-income settings may be needed to provide accurate risk prediction. Existing triage tools may need to be used at varying thresholds to reflect different baseline-line risks of adverse outcomes in different settings.

1428 What influences decisions and predicts transfer of older care-home residents to the emergency department? A synthesis of qualitative reviews and systematic review

Aims, Objectives and Background

The proportion of adults aged over 65 is rapidly increasing in developed countries. Care home residents have disproportionate rates of transfer to the ED. An estimated 40% of emergency admissions for care home residents may be for avoidable conditions and up to 8-fold variation in hospitalisations has been identified between care homes

We aimed to synthesise the qualitative research collated in existing reviews relating to the experience of residents, family members and professionals in decisions to transfer care home residents to the ED and identify known factors which predict ED transfer from care homes.

Method and Design

Two systematic reviews were conducted simultaneously. The first identified and synthesised the qualitative evidence presented in existing systematic reviews regarding decisions to transfer residents to the ED. The second identified quantitative factors found to affect likelihood of transfer of residents. Five electronic databases were searched, including: MEDLINE, EMBASE, CINAHL, PsychINFO, Web of Science and Scopus.

Results and Conclusion

In the qualitative component, six previous reviews met the inclusion criteria. Three syntheses were formed : (i) Transfer decisions involve negotiation with unequal power dynamics between residents, family members, care home staff and clinical practitioners (ii) Some transfers occur with the expectation that treatment in hospital will improve outcomes (iii) Some transfers occur due to factors external to the resident with no expectation that hospitalisation will be beneficial.

Twenty-six primary studies met the inclusion criteria for the quantitative component. Seven common domains of factors associated with ED transfer were identified: demographics, co-morbidities, medication use, frailty, permanent indwelling devices, advanced directives and care home organisation. Within these domains, male sex, age, presence of specific comorbidities, polypharmacy and quality rating were associated with ED transfer across studies.

This provides context for policy makers and researchers „developing interventions to reduce hospitalisations or use adjusted rates of hospitalisations as a care home quality indicator.

1427 The effect of the COVID-19 pandemic on major trauma presentations and patient outcomes in English hospitals

Aims, Objectives and Background

There is evidence that COVID-19 ‘lockdowns’ may have contributed to increased non-accidental injury, domestic violence and self-harm related to deteriorating mental health. Internationally, there is also evidence that the diversion of health care resources may led to worse outcomes for patients presenting with major trauma. There has been no previous national evaluation of ‘lockdown’ measures impact on the characteristics, treatment pathways and outcomes of trauma patients in England

We aimed to assess the impact of successive lockdowns on the volume, demographics, injury mechanism, severity, treatment and outcomes of major trauma in England.

Method and Design

Demographic characteristics and clinical pathways of TARN eligible patients in the first lockdown (24th March to 3rd July 2020 inclusive) and second lock down (1st November 2020 to 16th May 2021 inclusive) were compared to equivalent pre-COVID-19 periods in 2018–2019.

A segmented regression model predicting the weekly risk adjusted survival was estimated and a discontinuity in the gradient (trend) or intercept (level) of the fitted model was tested for at the weekly time point of implementation of each lockdown.

Abstract 1427 Figure 1

Strobe diagram for inclusion of study population

Abstract 1427 Figure 2

Interrupted time series analysis assessing the impact of COVID restrictions on likelihood of survival (red horizontal lines indicate introduction and relaxation of ‘lockdown’ measures)

Abstract 1427 Table 1

Comparison of demographics ‘lockdown’ and pre-COVID periods

PeriodPeriod24Mar19 – 03Jul19 (comparator)24Mar20 – 03Jul20 (lockdown 1)Absolute change [percentage point change (95%CI)] p-value01Nov18 – 16May19 (comparator)01Nov20 – 16May21 (lockdown 1)Absolute change [percentage point change (95%CI)] p-valueTotal2224317510-4733 (-21%)p<0.0001‡41016382622754 (–6.7%)p<0.0001‡Age (years), Median (IQR)67.6 (46.5–83.1)70.9 (50.3–84.2)3.3 (2.4 to 4.2)p<0.000169.1 (48.7–83.6)73.1 (53.3–85.1)4 (3.5 to 4.2)<0.0001Age bands, n(%)Age< 1138 (0.6%)130 (0.7%)-8 [0.1(-0.04 to 0.030)] p=0.14281 (0.7%)234 (0.6%)-47 [0.1 (-0.2 to 0.04)]p=0.1979Age <16942 (4.2%)674 (3.8%)-268 [-0.4 (-0.8 to 0]p=0.05311444 (3.5%)1218 (3.2%)-226 [-0.3(-0.6 to – 0.1)p=0.0084Age 16 – 649561 (43%)6974 (39.8%)-2587 [-3.2(-4.1 to -2.2)P<0.000117173 (41.9%)13980 (36.5%)-3193 [-5.3(-6 to -5)]p<0.0001Age 65 and over11740 (52.8%)9862 (56.3%)-1878 [3.5 (2.5 to 4.5)]p<0.000122399 (54.6%)23064 (60.3%)665 [5.7(5 to 6.3)]P<0.0001Age 85 and over4610 (20.7%)4047 (23.1%)-563 [2.4(1.6 to 3.2)]p<0.00018903 (21.7%)9731 (25.4%)828 [3.7 (3.1 to 4.3)]p<0.0001Male, n(%)12316 (55.4%)9512 (54.3%)-2804 [-1 (-2 to -0.6)]p=0.037322146 (54%)19769 (51.7%)-2377 [-2.3 (-3 to -1.6)]<0.0001CCI*, n(%)CCI 09359 (42.1%)6220 (35.5%)-3139 [ -6.5 (-7.5 to -5.6)] p<0.000116665 (40.6%)12806 (33.5%)-3859 [-7.1(-7.8 to -6.5)]p<0.0001CCI 1 – 58538 (38.4%)6896 (39.4%)-1642 [1 (0.3 to 2)]p=0.042615899 (38.8%)15667 (40.9%)-232 [2.2 (1.5 to 2.9)]p<0.0001CCI 6 – 103032 (13.6%)3061 (17.5%)29 [3.8 (3.2 to 4.6)]p<0.00015987 (14.6%)6863 (17.9%)876 [3.3(2.8 to 3.8)]p<0.0001CCI > 10927 (4.2%)1024 (5.8%)97 [1.7(1.2 to 2.1)]p<0.00011648 (4%)2410 (6.3%)762 [2.3(2 to 2.6)]p<0.0001Not recorded387 (1.7%)309 (1.8%)-88 [0.2 (-0.2 to 0.3)]p=0.8513817 (2%)516 (1.3%)-301 [-0.6(-0.8 to -0.5)]p<0.0001MOI**: RTC, n(%)Car occupant1247 (30.7%)551 (20.4%)-696 [-10.4(-12.4 to -8.2)]p<0.00012485 (35.2%)1551 (31.3%)-934 [-3.9(-5.6 to -2.2)]p<0.0001Pedestrian661 (16.3%)288 (10.6%)-373 [-5.6 (-7.2 to -4)]p<0.00011629 (23.1%)962 (19.4%)-667 [-3.7(-5.1 to -2.2)]p<0.0001Motorcycles1196 (29.4%)711 (26.3%)-485 [-3.2(-5.3 to -1)]p<0.00011524 (21.6%)976 (19.7%)-548[ -1.9(-3.3 to -0.4)]p<0.0001Cyclist912 (22.4%)1139 (42.1%)227 [19.6(17.4 to 21.9)]p<0.00011315 (18.6%)1396 (28.2%)81 [9.5(8 to 11.1)]p<0.0001Other11 (0.3%)<9 ()-10 [ -0.2(-0.4 to -0.06)p=0.025131 (0.4%)10 (0.2%)-21 [-0.23(-0.4 to -0.04)]p=0.0281MOI: Intentional, n(%)Intentional assault130 (0.6%)88 (0.5%)-42 [-0.08 (-0.2 to 0.06)]p=0.2724227 (0.6%)175 (0.5%)-52 [-0.1(-0.2 to 0.002)]P=0.0570Self harm276 (1.2%)284 (1.6%)8 [0.4 (0.1 to 0.6)]p=0.0014525 (1.3%)562 (1.5%)37 [0.2 (0.02 to 0.3)]p=0.0223NAI63 (0.3%)27 (0.2%)-36 [-0.1(-0.2 to -0.03)]p=0.007297 (0.2%)90 (0.2%)-7 [-0.001(-0.07 to 0.07)]p=0.9701Shooting34 (0.2%)40 (0.2%)6 [0.08(-0.01 to 0.2)]p=0.082680 (0.2%)56 (0.1%)-24 [ -0.05(-0.1 to 0.001)]p=0.0979Stabbing450 (2%)312 (1.8%)-138 [-0.2(-0.5 to 0.03)]p=0.0816791 (1.9%)589 (1.5%)-202 [-0.4 (-0.6 to -0.2)]p<0.0001Blows1174 (5.3%)647 (3.7%)-527 [-1.6(-1.9 to -1.2)]p<0.00012059 (5%)1299 (3.4%)-760 [-1.6(-1.9 to -1.3)]p<0.0001Unintentional, n(%)Falls>2m2055 (9.2%)1757 (10%)-298 [0.8(0.2 to 1.4)]P=0.00753740 (9,1%)3528 (9.2%)-212 [0.1(-0.3 to 0.5)]p=0.6181Falls<2m13384 (60.2%)11314 (64.6%)-2070 [4.4 (3.5 to 5.4)]p<0.000125505 (62.2%)26203 (65.8%)698 [6.3 (5.6 to 6.9)]p<0.0001Sport449 (2%)320 (1.8%)-129 [-0.2 (-0.5 to 0.01]p=0.1697615 (1.5%)489 (1.3%)-126 [-0.2 (-0.4 to -0.006)]p=0.0079GCS bands , n(%)Mild19609 (88.2%)15449 (88.2%)4160 [0.1 (-0.6 to 0.7)]p=0.826435831 (87.4%)34051 (89%)-1780 [1.6 (1.2 to 2.1)]p<0.0001Moderate689 (3.1%)625 (3.6%)-64 [0.5(0.1 to 0.8)]p=0.00901333 (3.2%)1127 (2.9%)-206 [-0.3 (-0.5 to -0.06)]p=0.0135Severe955 (4.3%)765 (4.4%)-190 [0.1 (-0.3 to 0.5)]p=0.71361886 (4.6%)1464 (3.8%)-422 [-0.8(-1 to -0.5)]p<0.0001Not recorded990 (4.5%)671 (3.8%)-319 [ -0.6(-1 to -0.2)]p=0.00221966 (4.8%)1620 (4.2%)-346 [-0.6(-0.8 to -0.3)]p=0.0002ISS***, median (IQR)9 (9–18)9 (9–18)09 (9–18)9 (9–17)0ISS bands, n(%)ISS 1 – 84545 (20.4%)3062 (17.5%)-1483 [-3 (-4 to -2)]p=<0.00018266 (20.2%)7838 (20.5%)-428 [0.3(-0.2 to 0.9)]p=0.2457ISS 9 – 159290 (41.8%)7728 (44.1%)-1562 [2.4(1.4 to 3.3)]p<0.000117207 (42%)16969 (44.3%)-233 [2.4(1.7 to 3.1)]p<0.0001ISS >158408 (37.8%)6720 (38.4%)-1688 [5.6(-0.4 to 1.5)]p=0.239115543 (37.9%)13455 (35.2%)-2088 [-2.7 (-3.4 to -2)]p<0.0001ISS >253995 (18%)3127 (17.9%)-868 [-0.1(-0.9 to 0.7 )]p=0.79217521 (18.3%)6201 (16.2%)-1320 [-2.1(-2.6 to -1.6)]p<0.0001Body regions, n(%)Head AIS 3+5911 (26.6%)4670 (26.7%)-1241 [0.1 (-0.8 to 1)]p=0.830111128 (27.1%)9629 (25.2%)-1499 [ -2(-2.6 to -1.3)]p<0.0001Face AIS 3+63 (0.3%)41 (0.2%)-22 [-0.05 (-0.1 to 0.05)]p=0.341699 (0.2%)69 (0.2%)-30 [-0.06 (-0.1 to 0)]p=0.0618Chest AIS 3+4787 (21.5%)3915 (22.4%)-872 [8.3 (0.2 to 1.6)]<0.04508515 (20.8%)8075 (21.1%)-440 [0.3 (-0.2 to 0.9)]p=0.2337Abdomen AIS 3+872 (3.9%)690 (3.9%)-182 [0.02 (-0.3 to 0.4)]p=0.91771465 (3.6%)1179 (3.1%)-286 [-0.5 (-0.7 to -0.2)]p=0.0001Spine AIS 3+1985 (8.9%)1561 (8.9%)-424 [-0.01(-0.6 to 0.5)]p=0.97443784 (9.2%)3459 (9%)-325 [-0.2(-0.6 to 0.2)]p=0.3654Pelvis AIS 3+758 (3.4%)600 (3.4%)-158 [0.02(-0.3 to 0.4)]p=0.91841501 (3.7%)1386 (3.6%)-115 [-0.04(-0.3 to 0.2)]p=0.7802Limb AIS 3+5707 (25.7%)4892 (27.9%)-815 [2.3 (1.4 to 3.2)]p<0.000110719 (26.1%)10122 (26.5%)-597 [0.3(-0.3 to 0.9)]p=0.3053Other AIS 3+217 (1%)199 (1.1%)-18 [0.2 (-0.04 to 0.3)]p=0.1176375 (0.9%)396 (1%)21 [0.1 (-0.01 to 0.2]p=0.0836Polytrauma1622 (7.3%)1350 (7.7%)-272 [0.4 (-0.1 to 0.9)]p=0.11602984 (7.3%)2429 (6.3%)-555 [-0.9(-1.2 to 0.6)]p<0.0001

*CCI Charlson Comorbidity Index

**MOI Mechanism of injury

***ISS Injury Severity Score

‡chi square test for uniform distribution

Abstract 1427 Table 2

Comparison care pathways ‘lockdown’ and pre-COVID periods

PeriodPeriod24Mar19 – 03Jul19 (comparator)24Mar20 – 03Jul20 (lockdown 1)Absolute Change01Nov18 – 16May19 (comparator)01Nov20 – 16May21 (lockdown 2)Absolute Change1stHospital MTC9908 (44.5%)7376 (42.1%)-2532 [-2.4 (-3.4 to -1.4)]p<0.000118099 (44.1%)15928 (41.6%)-2171 [-2.5 (-3.2 to -1.8)]p<0.0001Treated at MTC11176 (50.2%)8256 (47.2%)-2920 [-3 (-4 to -2)]p<0.000120395 (49.7%)17852 (46.7%)-2543[-3 (-4 to -2.4)]p<0.0001Consultant ED8140 (36.6%)5562 (31.8%)-2578 [-4.8(-5.8 to -3.9)]p<0.000114779 (36%)12577 (32.9%)-2202 [-3.2 (-3.8 to -2.5)]p<0.0001CT within 1 hr5062 (31.9%)3992 (30.9%)-1070 [-0.9(-2 to 0.1)]p=0.09449203 (31.6%)7776 (27.1%)-1427 [-4(-5 to -3.7)]p<0.0001Whole body CT3348 (15.1%)3210 (18.3%)-138 [3 (2 to 4)]p<0.00016040 (14.7%)6417 (16.8%)377 [2 (1.5 to 2.5)]p<0.0001ICU stay3092 (13.9%)2208 (12.6%)-884 [-1.3(-1.9 to -0.6) ]p=0.00025591 (13.6%)3850 (10.1%)-1741 [-3.6(-4 to -3)]p<0.0001Mortality*1417 (7.1%)1316 (8.3%)-101 [1.2 (0.6 to 1.7)]p<0.00012916 (7.9%)2858 (8.1%)-58 [0.2 (-0.1 to 0.6)] p=0.2040Discharge destination, n(%)Home (own)13800 (62%)10484 (59.9%)-3316 [-2(-3.1 to -1.2)]p<0.000124961 (60.9%)23368 (61.1%)-1593 [-0.7 (-1.4 to -0.05)]p=0.0340Home (relative/carer)473 (2.1%)372 (2.1%)-101 [0 (-0.3 to 0.3)]p=0.9890974 (2.4%)852 (2.2%)-122 [-0.1(-0.4 to 0.06)]p=0.1653Mortuary*1501 (6.7%)1323 (7.6%)-178 [0.8(0.3 to 1.3)]p=0.00193086 (7.5%)2977 (7.8%)-109 [0.1 (-0.3 to 0.5)]p=0.5113No fixed abode75 (0.3%)47 (0.3%)-28 (-37.3%)107 (0.3%)87 (0.2%)-20 (-18.7%)Not Known87 (0.4%)39 (0.2%)-48 (-55.2%)101 (0.2%)95 (0.2%)-6 (-5.9%)Nursing Home1190 (5.3%)1063 (6.1%)-127 [0.7(0.3 to 1.2)]p=0.00202448 (6%)2231 (5.8%)-217 [-0.2(-0.6 to 0.1)]p=0.1620Other Acute hospital2425 (10.9%)1736 (9.9%)-689 [-0.1(-1.6 to -0.4)]p=0.00144346 (10.6%)3313 (8.7%)-1033 [-0.1(-0.5 to 0.2)]p=0.4115Other institution526 (2.4%)516 (2.9%)-10 [0.6 (0.3 to 0.9)]p=0.0003980 (2.4%)870 (2.3%)-110 [-0.1 (-0.3 to 0.1)]p=0.2817Rehabilitation2077 (9.3%)1871 (10.7%)-206 [1.3(0.7 to 1.9)]p<0.00013851 (9.4%)4274 (11.2%)423 [ 1.7(1.3 to 2.2)]p<0.0001Social care63 (0.3%)50 (0.3%)-13 [0 (-0.1 to 0.1)]p=0.9657121 (0.3%)103 (0.3%)-18 [-0.2(-0.1 to 0.5)]p=0.4939

*These totals do not correspond as mortality includes deaths in the community and is censored at 30 days

Results and Conclusion

The first ‘lockdown’ had a larger associated reduction in total trauma volume (-21%) compared to the pre-COVID period than the second ‘lockdown’ (-6.7%). Trauma volume increased for those 65 and over (3%) and 85 and over (9.3%) during the second ‘lockdown’.

There was a reduction in likelihood of survival (-1.71; 95% CI:-2.76 to -0.66) associated with the immediate introduction of the first ‘lockdown’. However, this was followed by a trend of improving survival (0.25; 95% CI: 0.14 to 0.35) and likelihood of survival returned to pre-pandemic levels by the end of the first ‘lockdown’ period.

Future research is needed understand the initial reduction in likelihood of survival after major trauma observed with the implementation of the first ‘lockdown’ to prevent this occurring if measures re-introduced.

What influences decisions to transfer older care-home residents to the emergency department? A synthesis of qualitative reviews

BACKGROUND

care home residents aged over 65 have disproportionate rates of emergency department (ED) attendance and hospitalisation. Around 40% attendances may be avoidable, and hospitalisation is associated with harms. We synthesised the evidence available in qualitative systematic reviews of different stakeholders' experiences of decisions to transfer residents to the ED.

METHODS

six electronic databases, references and citations of included reviews and relevant policy documents were searched. Reviews of qualitative studies exploring factors that influenced care home staff, medical practitioners, residents' family or residents' experiences and factors influencing decisions to transfer residents to the ED were included. Thematic analysis was used to synthesise findings.

RESULTS

six previous reviews were included, which synthesised the findings of 34 primary studies encompassing 152 care home residents, 283 resident family members or carers and 447 care home staff. Of the primary studies, 19 were conducted in the North America, seven in Australia, five were conducted in Scandinavia, two in the United Kingdom and one in Holland. Three themes were identified: (i) power dynamics between residents, family members, care home staff and health care professionals (external to the care home) influence decisions; (ii) admission can be necessary; however, (iii) some decisions may be driven by factors other than clinical need.

CONCLUSION

transfer decisions are complex and are determined not just by changes in health status interventions aimed at reducing avoidable transfers need to address the key role family members have in transfer decisions, the medical legal fears of care home staff and barriers to accessing community services.

LMIC-PRIEST: Derivation and validation of a clinical severity score for acutely ill adults with suspected COVID-19 in a middle-income setting

Background

Uneven vaccination and less resilient health care systems mean hospitals in LMICs are at risk of being overwhelmed during periods of increased COVID-19 infection. Risk-scores proposed for rapid triage of need for admission from the emergency department (ED) have been developed in higher-income settings during initial waves of the pandemic.

Methods

Routinely collected data for public hospitals in the Western Cape, South Africa from the 27 th August 2020 to 11 th March 2022 were used to derive a cohort of 446,084 ED patients with suspected COVID-19. The primary outcome was death or ICU admission at 30 days. The cohort was divided into derivation and Omicron variant validation sets. We developed the LMIC-PRIEST score based on the coefficients from multivariable analysis in the derivation cohort and existing triage practices. We externally validated accuracy in the Omicron period and a UK cohort.

Results

We analysed 305,564, derivation 140,520 Omicron and 12,610 UK validation cases. Over 100 events per predictor parameter were modelled. Multivariable analyses identified eight predictor variables retained across models. We used these findings and clinical judgement to develop a score based on South African Triage Early Warning Scores and also included age, sex, oxygen saturation, inspired oxygen, diabetes and heart disease. The LMIC-PRIEST score achieved C-statistics: 0.82 (95% CI: 0.82 to 0.83) development cohort; 0.79 (95% CI: 0.78 to 0.80) Omicron cohort; and 0.79 (95% CI: 0.79 to 0.80) UK cohort. Differences in prevalence of outcomes led to imperfect calibration in external validation. However, use of the score at thresholds of three or less would allow identification of very low-risk patients (NPV ≥0.99) who could be rapidly discharged using information collected at initial assessment.

Conclusion

The LMIC-PRIEST score shows good discrimination and high sensitivity at lower thresholds and can be used to rapidly identify low-risk patients in LMIC ED settings.

What is already known on this subject

Uneven vaccination in low- and middle-income countries (LMICs) coupled with less resilient health care provision mean that emergency health care systems in LMICs may still be at risk of being overwhelmed during periods of increased COVID-19 infection.Risk-stratification scores may help rapidly triage need for hospitalisation. However, those proposed for use in the ED for patients with suspected COVID-19 have been developed and validated in high-income settings.

What this study adds

The LMIC-PRIEST score has been robustly developed using a large routine dataset from the Western Cape, South Africa and is directly applicable to existing triage practices in LMICs.External validation across both income settings and COVID-19 variants showed good discrimination and high sensitivity (at lower thresholds) to a composite outcome indicating need for inpatient admission from the ED.

How this study might affect research practice or policy

Use of the LMIC-PRIEST score at thresholds of three or less would allow identification of very low-risk patients (negative predictive value ≥0.99) across all settings assessedDuring periods of increased demand, this could allow the rapid identification and discharge of patients from the ED using information collected at initial assessment.

Accuracy of telephone triage for predicting adverse outcomes in suspected COVID-19: An observational cohort study linking NHS 111 telephone triage, primary and secondary healthcare and mortality records

ObjectivesSettings in identifying need for emergency care amongst those with suspected COVID-19 infection and identify factors which affect triage accuracy. ApproachAn observational cohort study of adults who contacted the NHS 111 telephone triage service provided by Yorkshire Ambulance Service between March and June 2020 with symptoms indicating possible COVID-19 infection. Patient-level data encompassing  triage call, primary care, hospital care and death registration records relating to 40,261 adults were linked. The accuracy of triage outcome (self-care/non-urgent assessment versus ambulance/urgent assessment) was assessed for death or organ support 30 days from first contact. Multivariable logistic regression was used to identify factors associated with risk of false negative or false positive triage. ResultsCallers had a 3% (1,200/40,261) risk of serious adverse outcomes. Telephone triage recommended self-care or non-urgent assessment for 60% (24,335/40,261), with a 1.3% (310/24,335) risk of adverse outcomes 30 days from first contact. Telephone triage had 74.2% sensitivity (95% CI: 71.6 to 76.6%) and 61.5% specificity (61% to 62%) for the primary outcome. Analysis suggested respiratory comorbidities may be over-appreciated and diabetes under-appreciated as predictors of deterioration. Repeat contact with triage service appears to be an important under-recognised predictor of deterioration. ConclusionPatients advised to self-care or receive non-urgent clinical assessment had a small but non-negligible risk of serious clinical deterioration. Repeat contact with telephone services needs recognition as an important predictor of subsequent adverse outcomes.

Factors associated with increased Emergency Department transfer in older long-term care residents: a systematic review

The proportion of adults older than 65 years is rapidly increasing. Care home residents in this age group have disproportionate rates of transfer to the Emergency Department (ED) and around 40% of attendances might be avoidable. We did a systematic review to identify factors that predict ED transfer from care homes. Six electronic databases were searched. Observational studies that provided estimates of association between ED attendance and variables at a resident or care home level were included. 26 primary studies met the inclusion criteria. Seven common domains of factors assessed for association with ED transfer were identified and within these domains, male sex, age, presence of specific comorbidities, polypharmacy, rural location, and care home quality rating were associated with likelihood of ED transfer. The identification of these factors provides useful information for policy makers and researchers intending to either develop interventions to reduce hospitalisations or use adjusted rates of hospitalisation as a care home quality indicator.

Accuracy of emergency medical service telephone triage of need for an ambulance response in suspected COVID-19: an observational cohort study

OBJECTIVE

To assess accuracy of emergency medical service (EMS) telephone triage in identifying patients who need an EMS response and identify factors which affect triage accuracy.

DESIGN

Observational cohort study.

SETTING

Emergency telephone triage provided by Yorkshire Ambulance Service (YAS) National Health Service (NHS) Trust.

PARTICIPANTS

12 653 adults who contacted EMS telephone triage services provided by YAS between 2 April 2020 and 29 June 2020 assessed by COVID-19 telephone triage pathways were included.

OUTCOME

Accuracy of call handler decision to dispatch an ambulance was assessed in terms of death or need for organ support at 30 days from first contact with the telephone triage service.

RESULTS

Callers contacting EMS dispatch services had an 11.1% (1405/12 653) risk of death or needing organ support. In total, 2000/12 653 (16%) of callers did not receive an emergency response and they had a 70/2000 (3.5%) risk of death or organ support. Ambulances were dispatched to 4230 callers (33.4%) who were not conveyed to hospital and did not deteriorate. Multivariable modelling found variables of older age (1 year increase, OR: 1.05, 95% CI: 1.04 to 1.05) and presence of pre-existing respiratory disease (OR: 1.35, 95% CI: 1.13 to 1.60) to be predictors of false positive triage.

CONCLUSION

Telephone triage can reduce ambulance responses but, with low specificity. A small but significant proportion of patients who do not receive an initial emergency response deteriorated. Research to improve accuracy of EMS telephone triage is needed and, due to limitations of routinely collected data, this is likely to require prospective data collection.

798 Accuracy of telephone triage for predicting adverse outcomes in suspected COVID-19: an observational cohort study

Aims/Objectives/Background

To reduce the risk of spreading infection and hospitals being overwhelmed, on the 18th February 2020, NHS England advised patients with suspected COVID infection to contact NHS 111 instead of attending health care providers. In March 2020, 3 million NHS 111 calls were made; a record number and double the number of the previous year. Concerns have been raised that telephone triage may not be sufficiently accurate in identifying need for emergency care.

We aim to assess accuracy of telephone triage in identifying patients who need emergency care amongst those with suspected COVID-19 and identify factors which affect triage accuracy.

Methods/Design

A cohort study of adults who contacted NHS 111 services provided by Yorkshire Ambulance Service between the 18thMarch 2020 and 29th June 2020 with symptoms indicating possible COVID-19 infection was completed. Callers were linked to ONS death registrations and routine health care data collected by NHS Digital.

The accuracy of triage outcome (self-care/non-urgent assessment versus ambulance/urgent assessment) was assessed for death or organ support 30 days from first contact. Multi-variable logistic regression was used to identify factors associated with risk of false negative or false positive triage.

Abstract 798 Figure 1

STROBE flow diagram of study population selection

Abstract 798 Table 1

Performance of binary NHS 111 triage (ambulance or urgent assessment 4 hours or less) for composite outcome (death or organ support)

Adverse outcome up to 30 days (3%, 2.8-3.2%)N=40, 261 Adverse OutcomeNo Adverse OutcomeAmbulance/urgent assessment 890 15, 035 Sensitivity 74.2% (71.6- 76.6%)Positive Predictive Value5.6% (5.2 - 6%) Self-care/non-urgent assessment 310 24, 025 Specificity 61.5% (61% - 62%)Negative Predictive Value98.7% (98.6 - 98.9%)

Results/Conclusions

3% of the 40,261 callers experienced an adverse outcome. Self-care/non-urgent assessment was recommended for 60%, with a small but non-negligible (1.3%) risk of subsequent deterioration. Triage achieved 74.2% sensitivity (95% CI: 71.6 to 76.6%) and 61.5% specificity (61% to 62%) for the primary outcome. Multivariable analysis suggested some co-morbidities (e.g. respiratory disease) may be over-estimated, and others (e.g. diabetes) underestimated, as predictors of deterioration. Repeat contact with services appears to be an important under recognised predictor of adverse outcomes with 2 contacts (OR 1.77 95% CI: 1.14 to 2.75) and 3+ contacts (OR 4.02 95% CI: 1.68 to 9.65) associated with clinical deterioration when not provided with an ambulance/urgent clinical assessment.

965 Accuracy of emergency (999) telephone triage for predicting adverse outcomes in suspected COVID-19: an observational cohort study

Aims/Objectives/Background

In the first wave of the pandemic some ambulance services received three times their usual number of 999 calls. The increase was mostly due to calls from patients with respiratory symptoms. Call handlers must rapidly decide whether patients need an emergency face-to-face assessment or could access non-emergency services.

We assess accuracy of emergency telephone triage in identifying patients with suspected COVID-19 infection who need an ambulance response and identify factors which affect triage accuracy.

Methods/Design

An observational cohort study of adults who contacted 999 emergency telephone services provided by Yorkshire Ambulance Service between the 18thMarch 2020 and 29th June 2020 with symptoms indicating possible COVID-19 infection was completed. Callers were linked to ONS death registrations and routine health care data collected by NHS Digital.

The accuracy of triage outcome (ambulance dispatch versus telephone advice) was assessed for death or organ support 30 days from first contact. Multi-variable logistic regression was used to identify factors associated with risk of false negative or false positive triage.

Results/Conclusions

Of the 12, 655 callers, 11.1% experienced the primary outcomes. An ambulance was dispatched to 84.2% of callers. The decision to dispatch an ambulance achieved 95% sensitivity (95% CI: 93.7 to 96.1%) and 17.2% specificity (95% 16.5% to 17.9%) for adverse outcomes. Where an ambulance was not dispatched, patients had a 3.5% (2.8 to 4.4%) of subsequent deterioration. Of patients that received an ambulance only 57% were subsequently conveyed to hospital. Multivariable logistic regression modelling found false negative assessment was associated with younger age and female sex and false positive assessment was associated with malignancy, immunosuppression, respiratory and cardiovascular comorbidities.

Emergency telephone triage of patients with suspected COVID-19 achieved a high sensitivity to serious adverse outcomes. Further research is required to identify ways specificity of triage could be improved to reduce unnecessary ambulance dispatch.

867 Prognostic accuracy of triage tools for adults with suspected COVID-19 in a pre-hospital setting: an observational cohort study

Aims/Objectives/Background

Emergency Medical Service (EMS) and other practitioners assessing patients with suspected COVID-19 in the community must rapidly determine whether patients need treatment in hospital or can self-care. Tools to triage patient acuity have only been validated in hospital populations.

We aimed to estimate the accuracy of five risk-stratification tools recommended to predict severe illness and compare accuracy to existing clinical decision-making in a pre-hospital setting.

Methods/Design

An observational cohort study using linked ambulance service data for patients assessed by EMS crews in the Yorkshire and Humber region of England between 18th March 2020 and 29th June 2020 was conducted to assess performance of the PRIEST tool, NEWS2, the WHO algorithm, CRB-65 and PMEWS in patients with suspected COVID-19 infection. The primary outcome was death or need for organ support.

Abstract 867 Table 1

Triage tool diagnostic accuracy statistics (95% CI) for predicting any adverse outcome

ToolN*C-statisticThresholdProportion with scoreSensitivitySpecificityPPVNPVCRB-6574700.79(0.78, 0.80)>00.540.89(0.88, 0.89)0.54(0.53, 0.54)0.29 (0.29, 0.30)0.96 (0.95, 0.96)NEWS274350.80(0.78, 0.81)>10.750.96(0.96, 0.96)0.30(0.29, 0.30)0.23 (0.22, 0.23)0.97 (0.97, 0.97)PMEWS74600.81(0.80, 0.83)>20.720.98(0.97, 0.98)0.34(0.33, 0.34)0.24 (0.24, 0.24)0.99 (0.98, 0.99)PRIEST74700.83(0.82, 0.84)>40.660.97(0.97, 0.97)0.41(0.40, 0.41)0.26 (0.25, 0.26)0.98 (0.98, 0.99)WHO74700.64(0.64, 0.65)>00.740.98(0.97, 0.98)0.31(0.30, 0.31)0.23 (0.23, 0.24)0.98(0.98, 0.99)

*Totals rounded to nearest 5

Abstract 867 Figure 1

ROC curves showing triage tool performance for predicting any adverse outcome

Results/Conclusions

Of 7550 patients in our cohort, 17.6% (95% CI:16.8% to 18.5%) experienced the primary outcome. The NEWS2, PMEWS, PRIEST tool and WHO algorithm identified patients at risk of adverse outcomes with a high sensitivity (>0.95) and specificity ranging between 0.3 (NEWS2) and 0.41 (PRIEST tool). The high sensitivity of NEWS2 and PMEWS was achieved by using lower thresholds than previously recommended (NEWS2; 0–1 vs 2+ and PMEWS; 0–2 vs 3+).

On index (first) assessment, 65% of patients were transported to hospital and EMS decision to transfer patients achieved a sensitivity of 0.84 (95% CI 0.83 to 0.85) and specificity of 0.39 (95% CI 0.39 to 0.40) to the primary outcome. This does not account for clinical reasons not to convey patients to hospital who subsequently deteriorated.

Use of NEWS2, PMEWS, PRIEST tool and WHO algorithm could therefore potentially improve EMS triage of patients with suspected COVID-19 infection. Use of the PRIEST tool could significantly increase the sensitivity of triage without increasing the number of patients conveyed to hospital.

Prognostic accuracy of triage tools for adults with suspected COVID-19 in a prehospital setting: an observational cohort study

Background

Tools proposed to triage patient acuity in COVID-19 infection have only been validated in hospital populations. We estimated the accuracy of five risk-stratification tools recommended to predict severe illness and compared accuracy to existing clinical decision making in a prehospital setting.

Methods

An observational cohort study using linked ambulance service data for patients attended by Emergency Medical Service (EMS) crews in the Yorkshire and Humber region of England between 26 March 2020 and 25 June 2020 was conducted to assess performance of the Pandemic Respiratory Infection Emergency System Triage (PRIEST) tool, National Early Warning Score (NEWS2), WHO algorithm, CRB-65 and Pandemic Medical Early Warning Score (PMEWS) in patients with suspected COVID-19 infection. The primary outcome was death or need for organ support.

Results

Of the 7549 patients in our cohort, 17.6% (95% CI 16.8% to 18.5%) experienced the primary outcome. The NEWS2 (National Early Warning Score, version 2), PMEWS, PRIEST tool and WHO algorithm identified patients at risk of adverse outcomes with a high sensitivity (>0.95) and specificity ranging from 0.3 (NEWS2) to 0.41 (PRIEST tool). The high sensitivity of NEWS2 and PMEWS was achieved by using lower thresholds than previously recommended. On index assessment, 65% of patients were transported to hospital and EMS decision to transfer patients achieved a sensitivity of 0.84 (95% CI 0.83 to 0.85) and specificity of 0.39 (95% CI 0.39 to 0.40).

Conclusion

Use of NEWS2, PMEWS, PRIEST tool and WHO algorithm could improve sensitivity of EMS triage of patients with suspected COVID-19 infection. Use of the PRIEST tool would improve sensitivity of triage without increasing the number of patients conveyed to hospital.

The burden of traumatic brain injury from low-energy falls among patients from 18 countries in the CENTER-TBI Registry: A comparative cohort study

BACKGROUND

Traumatic brain injury (TBI) is an important global public health burden, where those injured by high-energy transfer (e.g., road traffic collisions) are assumed to have more severe injury and are prioritised by emergency medical service trauma triage tools. However recent studies suggest an increasing TBI disease burden in older people injured through low-energy falls. We aimed to assess the prevalence of low-energy falls among patients presenting to hospital with TBI, and to compare their characteristics, care pathways, and outcomes to TBI caused by high-energy trauma.

METHODS AND FINDINGS

We conducted a comparative cohort study utilising the CENTER-TBI (Collaborative European NeuroTrauma Effectiveness Research in TBI) Registry, which recorded patient demographics, injury, care pathway, and acute care outcome data in 56 acute trauma receiving hospitals across 18 countries (17 countries in Europe and Israel). Patients presenting with TBI and indications for computed tomography (CT) brain scan between 2014 to 2018 were purposively sampled. The main study outcomes were (i) the prevalence of low-energy falls causing TBI within the overall cohort and (ii) comparisons of TBI patients injured by low-energy falls to TBI patients injured by high-energy transfer-in terms of demographic and injury characteristics, care pathways, and hospital mortality. In total, 22,782 eligible patients were enrolled, and study outcomes were analysed for 21,681 TBI patients with known injury mechanism; 40% (95% CI 39% to 41%) (8,622/21,681) of patients with TBI were injured by low-energy falls. Compared to 13,059 patients injured by high-energy transfer (HE cohort), the those injured through low-energy falls (LE cohort) were older (LE cohort, median 74 [IQR 56 to 84] years, versus HE cohort, median 42 [IQR 25 to 60] years; p < 0.001), more often female (LE cohort, 50% [95% CI 48% to 51%], versus HE cohort, 32% [95% CI 31% to 34%]; p < 0.001), more frequently taking pre-injury anticoagulants or/and platelet aggregation inhibitors (LE cohort, 44% [95% CI 42% to 45%], versus HE cohort, 13% [95% CI 11% to 14%]; p < 0.001), and less often presenting with moderately or severely impaired conscious level (LE cohort, 7.8% [95% CI 5.6% to 9.8%], versus HE cohort, 10% [95% CI 8.7% to 12%]; p < 0.001), but had similar in-hospital mortality (LE cohort, 6.3% [95% CI 4.2% to 8.3%], versus HE cohort, 7.0% [95% CI 5.3% to 8.6%]; p = 0.83). The CT brain scan traumatic abnormality rate was 3% lower in the LE cohort (LE cohort, 29% [95% CI 27% to 31%], versus HE cohort, 32% [95% CI 31% to 34%]; p < 0.001); individuals in the LE cohort were 50% less likely to receive critical care (LE cohort, 12% [95% CI 9.5% to 13%], versus HE cohort, 24% [95% CI 23% to 26%]; p < 0.001) or emergency interventions (LE cohort, 7.5% [95% CI 5.4% to 9.5%], versus HE cohort, 13% [95% CI 12% to 15%]; p < 0.001) than patients injured by high-energy transfer. The purposive sampling strategy and censorship of patient outcomes beyond hospital discharge are the main study limitations.

CONCLUSIONS

We observed that patients sustaining TBI from low-energy falls are an important component of the TBI disease burden and a distinct demographic cohort; further, our findings suggest that energy transfer may not predict intracranial injury or acute care mortality in patients with TBI presenting to hospital. This suggests that factors beyond energy transfer level may be more relevant to prehospital and emergency department TBI triage in older people. A specific focus to improve prevention and care for patients sustaining TBI from low-energy falls is required.

Prognostic accuracy of emergency department triage tools for adults with suspected COVID-19: the PRIEST observational cohort study

BACKGROUND

The WHO and National Institute for Health and Care Excellence recommend various triage tools to assist decision-making for patients with suspected COVID-19. We aimed to compare the accuracy of triage tools for predicting severe illness in adults presenting to the ED with suspected COVID-19.

METHODS

We undertook a mixed prospective and retrospective observational cohort study in 70 EDs across the UK. We collected data from people attending with suspected COVID-19 and used presenting data to determine the results of assessment with the WHO algorithm, National Early Warning Score version 2 (NEWS2), CURB-65, CRB-65, Pandemic Modified Early Warning Score (PMEWS) and the swine flu adult hospital pathway (SFAHP). We used 30-day outcome data (death or receipt of respiratory, cardiovascular or renal support) to determine prognostic accuracy for adverse outcome.

RESULTS

We analysed data from 20 891 adults, of whom 4611 (22.1%) died or received organ support (primary outcome), with 2058 (9.9%) receiving organ support and 2553 (12.2%) dying without organ support (secondary outcomes). C-statistics for the primary outcome were: CURB-65 0.75; CRB-65 0.70; PMEWS 0.77; NEWS2 (score) 0.77; NEWS2 (rule) 0.69; SFAHP (6-point rule) 0.70; SFAHP (7-point rule) 0.68; WHO algorithm 0.61. All triage tools showed worse prediction for receipt of organ support and better prediction for death without organ support. At the recommended threshold, PMEWS and the WHO criteria showed good sensitivity (0.97 and 0.95, respectively) at the expense of specificity (0.30 and 0.27, respectively). The NEWS2 score showed similar sensitivity (0.96) and specificity (0.28) when a lower threshold than recommended was used.

CONCLUSION

CURB-65, PMEWS and the NEWS2 score provide good but not excellent prediction for adverse outcome in suspected COVID-19, and predicted death without organ support better than receipt of organ support. PMEWS, the WHO criteria and NEWS2 (using a lower threshold than usually recommended) provide good sensitivity at the expense of specificity.

TRIAL REGISTRATION NUMBER

ISRCTN56149622.

Performance of the Hull Salford Cambridge Decision Rule (HSC DR) for early discharge of patients with findings on CT scan of the brain: a CENTER-TBI validation study

BACKGROUND

There is international variation in hospital admission practices for patients with mild traumatic brain injury (TBI) and injuries on CT scan. Only a small proportion of patients require neurosurgical intervention, while many guidelines recommend routine admission of all patients. We aim to validate the Hull Salford Cambridge Decision Rule (HSC DR) and the Brain Injury Guidelines (BIG) criteria to select low-risk patients for discharge from the emergency department.

METHOD

A cohort from 18 countries of Glasgow Coma Scale 13-15 patients with injuries on CT imaging was identified from the multicentre Collaborative European NeuroTrauma Effectiveness Research in TBI (CENTER-TBI) Study (conducted from 2014 to 2017) for secondary analysis. A composite outcome measure encompassing need for ongoing hospital admission was used, including seizure activity, death, intubation, neurosurgical intervention and neurological deterioration. We assessed the performance of our previously derived prognostic model, the HSC DR and the BIG criteria at predicting deterioration in this validation cohort.

RESULTS

Among 1047 patients meeting the inclusion criteria, 267 (26%) deteriorated. Our prognostic model achieved a C-statistic of 0.81 (95% CI: 0.78 to 0.84). The HSC DR achieved a sensitivity of 100% (95% CI: 97% to 100%) and specificity of only 4.7% (95% CI: 3.3% to 6.5%) for deterioration. Using the BIG criteria for discharge from the ED achieved a higher specificity (13.3%, 95% CI: 10.9% to 16.1%) and lower sensitivity (94.6%, 95% CI: 90.5% to 97%), with 12/105 patients recommended for discharge subsequently deteriorating, compared with 0/34 with the HSC DR.

CONCLUSION

Our decision rule would have allowed 3.5% of patients to be discharged, none of whom would have deteriorated. Use of the BIG criteria may select patients for discharge who have too high a risk of subsequent deterioration to be used clinically. Further validation and implementation studies are required to support use in clinical practice.

External validation of the Dutch prediction model for prehospital triage of trauma patients in South West region of England, United Kingdom

IMPORTANCE

This paper investigates the use of a major trauma prediction model in the UK setting. We demonstrate that application of this model could reduce the number of patients with major trauma being incorrectly sent to non-specialist hospitals. However, more research is needed to reduce over-triage and unnecessary transfer to Major Trauma Centres.

OBJECTIVE

To externally validate the Dutch prediction model for identifying major trauma in a large unselected prehospital population of injured patients in England.

DESIGN

External validation using a retrospective cohort of injured patients who ambulance crews transported to hospitals.

SETTING

South West region of England.

PARTICIPANTS

All patients ≥16 years with a suspected injury and transported by ambulance in the year from February 1, 2017.

EXCLUSION CRITERIA

1) Patients aged ≤15 years; 2) Non-ambulance attendance at hospital with injuries; 3) Death at the scene and; 4) Patients conveyed by helicopter. This study had a census sample of cases available to us over a one year period.

INTERVENTIONS OR EXPOSURES

Tested the accuracy of the prediction model in terms of discrimination, calibration, clinical usefulness, sensitivity and specificity and under- and over triage rates compared to usual triage practices in the South West region.

MAIN OUTCOME MEASURE

Major trauma defined as an Injury Severity Score>15.

RESULTS

A total of 68799 adult patients were included in the external validation cohort. The median age of patients was 72 (i.q.r. 46-84); 55.5% were female; and 524 (0.8%) had an Injury Severity Score>15. The model achieved good discrimination with a C-Statistic 0.75 (95% CI, 0.73 - 0.78). The maximal specificity of 50% and sensitivity of 83% suggests the model could improve undertriage rates at the expense of increased overtriage rates compared with routine trauma triage methods used in the South West, England.

CONCLUSIONS AND RELEVANCE

The Dutch prediction model for identifying major trauma could lower the undertriage rate to 17%, however it would increase the overtriage rate to 50% in this United Kingdom cohort. Further prospective research is needed to determine whether the model can be practically implemented by paramedics and is cost-effective.

Predicting need for hospital admission in patients with traumatic brain injury or skull fractures identified on CT imaging: a machine learning approach

BACKGROUND

Patients with mild traumatic brain injury on CT scan are routinely admitted for inpatient observation. Only a small proportion of patients require clinical intervention. We recently developed a decision rule using traditional statistical techniques that found neurologically intact patients with isolated simple skull fractures or single bleeds <5 mm with no preinjury antiplatelet or anticoagulant use may be safely discharged from the emergency department. The decision rule achieved a sensitivity of 99.5% (95% CI 98.1% to 99.9%) and specificity of 7.4% (95% CI 6.0% to 9.1%) to clinical deterioration. We aimed to transparently report a machine learning approach to assess if predictive accuracy could be improved.

METHODS

We used data from the same retrospective cohort of 1699 initial Glasgow Coma Scale (GCS) 13-15 patients with injuries identified by CT who presented to three English Major Trauma Centres between 2010 and 2017 as in our original study. We assessed the ability of machine learning to predict the same composite outcome measure of deterioration (indicating need for hospital admission). Predictive models were built using gradient boosted decision trees which consisted of an ensemble of decision trees to optimise model performance.

RESULTS

The final algorithm reported a mean positive predictive value of 29%, mean negative predictive value of 94%, mean area under the curve (C-statistic) of 0.75, mean sensitivity of 99% and mean specificity of 7%. As with logistic regression, GCS, severity and number of brain injuries were found to be important predictors of deterioration.

CONCLUSION

We found no clear advantages over the traditional prediction methods, although the models were, effectively, developed using a smaller data set, due to the need to divide it into training, calibration and validation sets. Future research should focus on developing models that provide clear advantages over existing classical techniques in predicting outcomes in this population.

Post-exertion oxygen saturation as a prognostic factor for adverse outcome in patients attending the emergency department with suspected COVID-19: a substudy of the PRIEST observational cohort study

BACKGROUND

Measurement of post-exertion oxygen saturation has been proposed to assess illness severity in suspected COVID-19 infection. We aimed to determine the accuracy of post-exertional oxygen saturation for predicting adverse outcome in suspected COVID-19.

METHODS

We undertook a substudy of an observational cohort study across 70 emergency departments during the first wave of the COVID-19 pandemic in the UK. We collected data prospectively, using a standardised assessment form, and retrospectively, using hospital records, from patients with suspected COVID-19, and reviewed hospital records at 30 days for adverse outcome (death or receiving organ support). Patients with post-exertion oxygen saturation recorded were selected for this analysis. We constructed receiver-operating characteristic curves, calculated diagnostic parameters, and developed a multivariable model for predicting adverse outcome.

RESULTS

We analysed data from 817 patients with post-exertion oxygen saturation recorded after excluding 54 in whom measurement appeared unfeasible. The c-statistic for post-exertion change in oxygen saturation was 0.589 (95% CI 0.465 to 0.713), and the positive and negative likelihood ratios of a 3% or more desaturation were, respectively, 1.78 (1.25 to 2.53) and 0.67 (0.46 to 0.98). Multivariable analysis showed that post-exertion oxygen saturation was not a significant predictor of adverse outcome when baseline clinical assessment was taken into account (p=0.368). Secondary analysis excluding patients in whom post-exertion measurement appeared inappropriate resulted in a c-statistic of 0.699 (0.581 to 0.817), likelihood ratios of 1.98 (1.26 to 3.10) and 0.61 (0.35 to 1.07), and some evidence of additional prognostic value on multivariable analysis (p=0.019).

CONCLUSIONS

Post-exertion oxygen saturation provides modest prognostic information in the assessment of selected patients attending the emergency department with suspected COVID-19.

TRIAL REGISTRATION NUMBER

ISRCTN Registry (ISRCTN56149622) http://www.isrctn.com/ISRCTN28342533.

Derivation and validation of a clinical severity score for acutely ill adults with suspected COVID-19: The PRIEST observational cohort study

OBJECTIVES

We aimed to derive and validate a triage tool, based on clinical assessment alone, for predicting adverse outcome in acutely ill adults with suspected COVID-19 infection.

METHODS

We undertook a mixed prospective and retrospective observational cohort study in 70 emergency departments across the United Kingdom (UK). We collected presenting data from 22445 people attending with suspected COVID-19 between 26 March 2020 and 28 May 2020. The primary outcome was death or organ support (respiratory, cardiovascular, or renal) by record review at 30 days. We split the cohort into derivation and validation sets, developed a clinical score based on the coefficients from multivariable analysis using the derivation set, and the estimated discriminant performance using the validation set.

RESULTS

We analysed 11773 derivation and 9118 validation cases. Multivariable analysis identified that age, sex, respiratory rate, systolic blood pressure, oxygen saturation/inspired oxygen ratio, performance status, consciousness, history of renal impairment, and respiratory distress were retained in analyses restricted to the ten or fewer predictors. We used findings from multivariable analysis and clinical judgement to develop a score based on the NEWS2 score, age, sex, and performance status. This had a c-statistic of 0.80 (95% confidence interval 0.79-0.81) in the validation cohort and predicted adverse outcome with sensitivity 0.98 (0.97-0.98) and specificity 0.34 (0.34-0.35) for scores above four points.

CONCLUSION

A clinical score based on NEWS2, age, sex, and performance status predicts adverse outcome with good discrimination in adults with suspected COVID-19 and can be used to support decision-making in emergency care.

REGISTRATION

ISRCTN registry, ISRCTN28342533, http://www.isrctn.com/ISRCTN28342533.

59 Performance of the mTBI decision rule for early discharge of patients with findings on CT: a center-TBI validation study

Aims/Objectives/Background

GCS 13–15 patients with TBI identified by CT imaging are routinely admitted for observation in the UK. A small proportion of patients clinically deteriorates or requires intervention. We previously derived a prognostic model and decision rule to identify low-risk patients with injuries on CT who could be safely discharged from the ED. Neither has been externally validated.

We aim to externally validate our empirically derived prognostic model and decision rule.

Methods/Design

A cohort of initial GCS13-15 patients with injuries on CT was derived from the CENTER-TBI cohort study. CENTER-TBI recruited patients who underwent CT imaging for head trauma between December 2014 and 2017 at 63 centres across Europe and Israel. A composite outcome encompassing need for hospital admission was used, including: seizures, death, intubation, admission to ICU, neurosurgical intervention and neurological deterioration. Performance of the model was assessed by measures of discrimination and calibration. The sensitivity and specificity of the decision rule to the composite outcome was estimated at the discharge threshold.

Abstract 59 Figure 1

STROBE flow diagram of selection of study population

Abstract 59 Table 1

Discrimination and calibration of mTBI prognostic model in CENTER TBI cohort

Model performanceValue (averaged across imputations)Recalibrated model performanceValueC-Statistic 0.81 (95% CI, 0.78 – 0.84) C-Statistic 0.81 Calibration in the large (CITL) -5 Calibration in the large (CITL) 0.0 Calibration slope 0.5 Calibration slope 1 Expected vs. observed (E:O) 3.3 Expected vs. observed (E:O) 1.0 Abstract 59 Table 2

Performance of mTBI discharge decision rule and BIG criteria

mTBI Risk ScoreN=961 DeterioratedNo deteriorationRisk=0 (discharge ED)0 34 (3.5%) Sensitivity 100% (95% CI: 97% to 100%) Risk>0 (Inpatient admission)234 693 Specificity 4.7% (95% CI: 3.3% to 6.5%) BIG CriteriaN=921 DeterioratedNo deteriorationBIG 1 (discharge ED)12 93 Sensitivity 94% (95% CI: 90.5% to 97%) BIG 2/3 (Inpatient admission)210 606 Specificity 13.3% (95% CI: 10.9% to 16.1%)

Results/Conclusions

1047 of 4509 patients recruited to the CENTER-TBI study met the inclusion criteria. 25.5% (95% CI: 22.9% to 28.2%) clinically deteriorated and 20.2% (95% CI: 17.9% to 22.8%) underwent neurosurgery, died, or were intubated. The prognostic model had an estimated C-static of 0.81 and a calibration slope of 0.5. Our decision rule achieved 100% (95% CI: 97% to 100%) sensitivity and specificity of 4.7% (95% CI: 3.3% to 6.5%) to clinical deterioration. This would allow 3.5% of patients to be discharged- none of whom deteriorated. The decision rule outperformed the BIG criteria, which is used to triage hospital admissions in the USA.

External validation shows our decision rule may be safe for routine use in clinical practice. The inclusion of biomarkers or other novel factors may improve the calibration of the model and the specificity of the decision rule.

11 External validation of the Dutch prediction model for prehospital triage of trauma patients in South West region of England, United Kingdom

Aims/Objectives/Background

This is the first external validation of a European empirically derived prediction model for identifying major trauma in an unselected group of injured patients transported by ambulance in the United Kingdom.

Methods/Design

This study was an external validation of a Dutch prediction model for identifying major trauma using a retrospective cohort of injured patients who ambulance crews transported to hospitals in the South West region of England. Major trauma was defined as Injury Severity Score (ISS)>15.

Participants were patients ≥16 years with a suspected injury and transported by ambulance from February 1, 2017 to February 1, 2018. This study had a census sample of cases available to us over a one year period.

We tested the accuracy of the prediction model in terms of discrimination, calibration, clinical usefulness, sensitivity and specificity and under- and over triage rates compared to existing trauma triage practices in the South West region.

Results/Conclusions

A total of 68 698 adult patients were included in the final external validation cohort. The median age of patients was 72 (i.q.r. 46–84); 55.5% were female; and 524 (0.8%) had an ISS>15. In comparison the Dutch cohort was younger (45 years), more were male (58.3%) and more patients had an ISS>15. (8.8%) The model achieved good discrimination with a C-Statistic 0.75 (95% CI, 0.73 – 0.78). At a maximal specificity of 50% the model resulted in a sensitivity of 86%. The model improved undertriage rates at the expense of increased overtriage rates compared with routine trauma triage methods in the South West of England.

The Dutch prediction model for identifying major trauma can lower the undertriage rate to 17%, however it would increase the overtriage rate to 50% in this UK cohort. Further research is needed to determine whether the model can be practically implemented by paramedics and is cost-effective.

049 Development of a clinical decision rule for the early safe discharge of patients with mild traumatic brain injury and findings on CT brain scan: a retrospective cohort study

Background

International guidelines recommend routine hospital admission for all patients with mild traumatic brain injury (TBI) who have injuries on CT brain scan. Only a small proportion of these patients require neurosurgical or critical care intervention. We aimed to develop an accurate clinical decision rule to identify low risk patients safe for discharge from the emergency department (ED) and facilitate earlier referral of those requiring intervention.

Method and results

A retrospective cohort study of case-notes of patients admitted with initial GCS13-15 and injuries identified by CT was completed. Data on a primary outcome measure of clinically important deterioration (indicating need for hospital admission) and secondary outcome of neurosurgery, ICU admission or intubation (indicating need for neurosurgical admission) were collected. Multivariable logistic regression was used to derive models and a risk score predicting deterioration using routinely reported candidate variables identified in a systematic review. We compared the performance of this new risk score with the Brain Injury Guideline (BIG) criteria, derived in the USA.

Abstract 049 Figure 1

Population selection

Abstract 049 Table 1

Model performance

Abstract 049 Table 2

Performance of risk score and BIG criteria

Conclusions

1699 patients were included from 3 English Major Trauma Centres. 27.7% (95% CI: 25.5% to 29.9%) met the primary, and 13.1% (95% CI: 11.6% to 14.8%) met the secondary, outcome of deterioration. The derived clinical decision rule suggests that patients with simple skull fractures or intracranial bleeding less than 5 mm in diameter who are fully conscious could be safely discharged from the Emergency Department. The decision rule achieved a sensitivity of 99.5% (95% CI: 98.1% to 99.9%) and specificity of 7.4% (95% CI: 6% to 9.1%) to the primary outcome. The BIG criteria achieved the same sensitivity but lower specificity (5%).

Our empirical models showed good predictive performance and outperformed the BIG criteria. This would potentially allow ED discharge of one in twenty patients currently admitted for observation. However prospective external validation and economic evaluation is required.

016 An evaluation of the impact of the NICE head injury guidelines on inpatient mortality from traumatic brain injury: an interrupted time series analysis

Background

Traumatic brain injury (TBI) is the commonest cause of death and disability in UK Citizens aged 1–40. In England three (National Institute of Health and Care Excellence - NICE) guidelines have been implemented to improve TBI outcomes. All guidelines recommended increased CT imaging. The second guideline recommended the management of patients with severe TBI in specialist neuroscience centres.

This study uses national data and interrupted time series analysis to assess the impact of the NICE guidelines.

Individual level Office of National Statistics (ONS) cause of death data linked to Hospital Episode Statistics for inpatient admissions in England between 1998–2017 were used to estimate the monthly population mortality and admission rate for TBI.

An interrupted time series analysis was conducted with intervention points when each guideline was introduced. The analysis was stratified by guideline recommendation specific age groups (0–15, 16–64 and 65+).

The monthly TBI mortality and admission rate in the 65+ age group increased from 0.5 to 1.5 and 10 to 30 per 100, 000 population respectively. The increasing mortality rate was unaffected by the introduction any of the guidelines.

The introduction of the 2nd NICE Head Injury guideline was associated with a significant reduction in the monthly TBI mortality rate in 16–64 age group (−0.005; 95% CI: −0.002 to −0.007).

In the 0–15 age group the TBI mortality rate fell from around 0.05 to 0.01 per 100 000 population, the trend was unaffected by the guidelines.

Abstract 016 Figure 1

The impact of the NICE head injury guidelines on monthly TBI mortality rate per 100 000 population

Abstract 016 Figure 2

The impact of the NICE head guidelines on monthly TBI hospital admissions per 100,000 population

Conclusion

The introduction of NICE head injury guidelines was associated with reduced population based mortality rates after specialist care was recommended for severe TBI. The improvement was solely observed in 16–64 year olds.

The cause of the observed increased admission and mortality rate in those 65+ and potential treatments for TBI in this age group requires further investigation.

Development of a Clinical Decision Rule for the Early Safe Discharge of Patients with Mild Traumatic Brain Injury and Findings on Computed Tomography Brain Scan: A Retrospective Cohort Study

International guidelines recommend routine hospital admission for all patients with mild traumatic brain injury (TBI) who have injuries on computed tomography (CT) brain scan. Only a small proportion of these patients require neurosurgical or critical care intervention. We aimed to develop an accurate clinical decision rule to identify low-risk patients safe for discharge from the emergency department (ED) and facilitate earlier referral of those requiring intervention. A retrospective cohort study of case notes of patients admitted with initial Glasgow Coma Scale 13–15 and injuries identified by CT was completed. Data on a primary outcome measure of clinically important deterioration (indicating need for hospital admission) and secondary outcome of neurosurgery, intensive care unit admission, or intubation (indicating need for neurosurgical admission) were collected. Multi-variable logistic regression was used to derive models and a risk score predicting deterioration using routinely reported clinical and radiological candidate variables identified in a systematic review. We compared the performance of this new risk score with the Brain Injury Guideline (BIG) criteria, derived in the United States. A total of 1699 patients were included from three English major trauma centers. A total of 27.7% (95% confidence interval [CI], 25.5–29.9) met the primary and 13.1% (95% CI, 11.6–14.8) met the secondary outcomes of deterioration. The derived clinical decision rule suggests that patients with simple skull fractures or intracranial bleeding <5 mm in diameter who are fully conscious could be safely discharged from the ED. The decision rule achieved a sensitivity of 99.5% (95% CI, 98.1–99.9) and specificity of 7.4% (95% CI, 6.0–9.1) to the primary outcome. The BIG criteria achieved the same sensitivity, but lower specificity (5%). Our empirical models showed good predictive performance and outperformed the BIG criteria. This would potentially allow ED discharge of 1 in 20 patients currently admitted for observation. However, prospective external validation and economic evaluation are required.

Electrocardiographic characteristics in patients with heart failure and normal ejection fraction: A systematic review and meta-analysis

BACKGROUND

Little is known about ECG abnormalities in patients with heart failure and normal ejection fraction (HeFNEF) and how they relate to different etiologies or outcomes.

METHODS AND RESULTS

We searched the literature for peer-reviewed studies describing ECG abnormalities in HeFNEF other than heart rhythm alone. Thirty five studies were identified and 32,006 participants. ECG abnormalities reported in patients with HeFNEF include atrial fibrillation (prevalence 12%-46%), long PR interval (11%-20%), left ventricular hypertrophy (LVH, 10%-30%), pathological Q waves (11%-18%), RBBB (6%-16%), LBBB (0%-8%), and long JTc (3%-4%). Atrial fibrillation is more common in patients with HeFNEF compared to those with heart failure and reduced ejection fraction (HeFREF). In contrast, long PR interval, LVH, Q waves, LBBB, and long JTc are more common in patients with HeFREF. A pooled effect estimate analysis showed that QRS duration ≥120 ms, although uncommon (13%-19%), is associated with worse outcomes in patients with HeFNEF.

CONCLUSIONS

There is high variability in the prevalence of ECG abnormalities in patients with HeFNEF. Atrial fibrillation is more common in patients with HeFNEF compared to those with HeFREF. QRS duration ≥120 ms is associated with worse outcomes in patients with HeFNEF. Further studies are needed to address whether ECG abnormalities correlate with different phenotypes in HeFNEF.

Evaluation of the impact of the NICE head injury guidelines on inpatient mortality from traumatic brain injury: an interrupted time series analysis

OBJECTIVE

To evaluate the impact of National Institute for Health and Care Excellence (NICE) head injury guidelines on deaths and hospital admissions caused by traumatic brain injury (TBI).

SETTING

All hospitals in England between 1998 and 2017.

PARTICIPANTS

Patients admitted to hospital or who died up to 30 days following hospital admission with International Classification of Diseases (ICD) coding indicating the reason for admission or death was TBI.

INTERVENTION

An interrupted time series analysis was conducted with intervention points when each of the three guidelines was introduced. Analysis was stratified by guideline recommendation specific age groups (0-15, 16-64 and 65+).

OUTCOME MEASURES

The monthly population mortality and admission rates for TBI.

STUDY DESIGN

An interrupted time series analysis using complete Office of National Statistics cause of death data linked to hospital episode statistics for inpatient admissions in England.

RESULTS

The monthly TBI mortality and admission rates in the 65+ age group increased from 0.5 to 1.5 and 10 to 30 per 100 000 population, respectively. The increasing mortality rate was unaffected by the introduction of any of the guidelines.The introduction of the second NICE head injury guideline was associated with a significant reduction in the monthly TBI mortality rate in the 16-64 age group (-0.005; 95% CI: -0.002 to -0.007).In the 0-15 age group the TBI mortality rate fell from around 0.05 to 0.01 per 100 000 population and this trend was unaffected by any guideline.

CONCLUSION

The introduction of NICE head injury guidelines was associated with a reduced admitted TBI mortality rate after specialist care was recommended for severe TBI. The improvement was solely observed in patients aged 16-64 years.The cause of the observed increased admission and mortality rates in those 65+ and potential treatments for TBI in this age group require further investigation.

Impact of the SIGN head injury guidelines and NHS 4-hour emergency target on hospital admissions for head injury in Scotland: an interrupted times series

OBJECTIVES

Head injury is a common reason for emergency department (ED) attendance. Around 1% of patients have life-threatening injuries, while 80% of patients are discharged. National guidelines (Scottish Intercollegiate Guidelines Network (SIGN)) were introduced in Scotland with the aim of achieving early identification of those with acute intracranial lesions yet safely reducing hospital admissions.This study aims to assess the impact of these guidelines and any effect the national 4-hour ED performance target had on hospital admissions for head injury.

SETTING

All Scottish hospitals between April 1998 and March 2016.

PARTICIPANTS

Patients admitted to hospital for head injury or traumatic brain injury (TBI) diagnosed by CT imaging identified using administrative Scottish Information Services Division data. There are 275 hospitals in Scotland. In 2015/2016, there were 571 221 emergency hospital admissions in Scotland.

INTERVENTIONS

The SIGN head injury guidelines introduced in 2000 and 2009. The 4-hour ED target introduced in 2004.

OUTCOMES

The monthly rate of hospital admissions for head injury and traumatic brain injury.

STUDY DESIGN

An interrupted time series analysis.

RESULTS

The first guideline was associated with a reduction in monthly admissions of 0.14 (95% CI 0.09 to 4.83) per 100 000 population. The 4-hour target was associated with a monthly increase in admissions of 0.13 (95% CI 0.06 to 0.20) per 100 000 population. The second guideline reduced monthly admissions by 0.09 (95% CI-0.13 to -0.05) per 100 000 population. These effects varied between age groups.The guidelines were associated with increased admissions for patients with injuries identified by CT imaging-guideline 1: 0.06 (95% CI 0.004 to 0.12); guideline 2: 0.05 (95% CI 0.04 to 0.06) per 100 000 population.

CONCLUSION

Increased CT imaging of head injured patients recommended by SIGN guidelines reduced hospital admissions. The 4-hour ED target and the increased identification of TBI by CT imaging acted to undermine this effect.

The Risk of Deterioration in GCS13-15 Patients with Traumatic Brain Injury Identified by Computed Tomography Imaging: A Systematic Review and Meta-Analysis

The optimal management of mild traumatic brain injury (TBI) patients with injuries identified by computed tomography (CT) brain scan is unclear. Some guidelines recommend hospital admission for an observation period of at least 24 h. Others argue that selected lower-risk patients can be discharged from the Emergency Department (ED). The objective of our review and meta-analysis was to estimate the risk of death, neurosurgical intervention, and clinical deterioration in mild TBI patients with injuries identified by CT brain scan, and assess which patient factors affect the risk of these outcomes. A systematic review and meta-analysis adhering to PRISMA standards of protocol and reporting were conducted. Study selection was performed by two independent reviewers. Meta-analysis using a random effects model was undertaken to estimate pooled risks for: clinical deterioration, neurosurgical intervention, and death. Meta-regression was used to explore between-study variation in outcome estimates using study population characteristics. Forty-nine primary studies and five reviews were identified that met the inclusion criteria. The estimated pooled risk for the outcomes of interest were: clinical deterioration 11.7% (95% confidence interval [CI]: 11.7%-15.8%), neurosurgical intervention 3.5% (95% CI: 2.2%-4.9%), and death 1.4% (95% CI: 0.8%-2.2%). Twenty-one studies presented within-study estimates of the effect of patient factors. Meta-regression of study characteristics and pooling of within-study estimates of risk factor effect found the following factors significantly affected the risk for adverse outcomes: age, initial Glasgow Coma Scale (GCS), type of injury, and anti-coagulation. The generalizability of many studies was limited due to population selection. Mild TBI patients with injuries identified by CT brain scan have a small but clinically important risk for serious adverse outcomes. This review has identified several prognostic factors; research is needed to derive and validate a usable clinical decision rule so that low-risk patients can be safely discharged from the ED.

Postmortem ICD interrogation in mode of death classification

BACKGROUND

The definition of sudden death due to arrhythmia relies on the time interval between onset of symptoms and death. However, not all sudden deaths are due to arrhythmia. In patients with an implantable cardioverter defibrillator (ICD), postmortem device interrogation may help better distinguish the mode of death compared to a time-based definition alone.

OBJECTIVE

This study aims to assess the proportion of "sudden" cardiac deaths in patients with an ICD that have confirmed arrhythmia.

METHODS

We conducted a literature search for studies using postmortem ICD interrogation and a time-based classification of the mode of death. A modified QUADAS-2 checklist was used to assess risk of bias in individual studies. Outcome data were pooled where sufficient data were available.

RESULTS

Our search identified 22 studies undertaken between 1982 and 2015 with 23,600 participants. The pooled results (excluding studies with high risk of bias) suggest that ventricular arrhythmias are present at the time of death in 76% of "sudden" deaths (95% confidence interval [CI] 67-85; range 42-88).

CONCLUSION

Postmortem ICD interrogation identifies 24% of "sudden" deaths to be nonarrhythmic. Postmortem device interrogation should be considered in all cases of unexplained sudden cardiac death.

A protocol for the development of a prediction model in mild traumatic brain injury with CT scan abnormality: which patients are safe for discharge?

BACKGROUND

Head injury is an extremely common clinical presentation to hospital emergency departments (EDs). Ninety-five percent of patients present with an initial Glasgow Coma Scale (GCS) score of 13-15, indicating a normal or near-normal conscious level. In this group, around 7% of patients have brain injuries identified by CT imaging but only 1% of patients have life-threatening brain injuries. It is unclear which brain injuries are clinically significant, so all patients with brain injuries identified by CT imaging are admitted for monitoring. If risk could be accurately determined in this group, admissions for low-risk patients could be avoided and resources could be focused on those with greater need.This study aims to (a) estimate the proportion of GCS13-15 patients with traumatic brain injury identified by CT imaging admitted to hospital who clinically deteriorate and (b) develop a prognostic model highly sensitive to clinical deterioration which could help inform discharge decision making in the ED.

METHODS

A retrospective case note review of 2000 patients with an initial GCS13-15 and traumatic brain injury identified by CT imaging (2007-2017) will be completed in two English major trauma centres. The prevalence of clinically significant deterioration including death, neurosurgery, intubation, seizures or drop in GCS by more than 1 point will be estimated. Candidate prognostic factors have been identified in a previous systematic review. Multivariable logistic regression will be used to derive a prognostic model, and its sensitivity and specificity to the outcome of deterioration will be explored.

DISCUSSION

This study will potentially derive a statistical model that predicts clinically relevant deterioration and could be used to develop a clinical risk tool guiding the need for hospital admission in this group.

The risk of intra-cranial haemorrhage in those presenting late to the ED following a head injury: a systematic review

BACKGROUND

Head injury represents an extremely common presentation to emergency departments (ED), but not all patients present immediately after injury. There is evidence that clinical deterioration following head injury will usually occur within 24 h. It is unclear whether this means that head injury patients that present in a delayed manner, especially after 24 h, have a lower prevalence of significant traumatic injuries including intra-cranial haemorrhages.

METHODS

A systematic review protocol was designed with the aim of systematically identifying and evaluating studies in delayed ED presentation head injury populations in order to establish whether the prevalence of significant intra-cranial injury was affected by delay in presentation. Two independent researchers assessed retrieved studies for inclusion against pre-determined inclusion criteria. Studies had to be conducted in ED head injury populations presenting in a delayed manner, and report a measure of prevalence of traumatic CT abnormality as an outcome.

RESULTS

Three studies were eligible for inclusion. They were all of poor methodological quality, and heterogeneity prevented meta-analysis. The reported prevalence of traumatic intra-cranial injury on CT was between 2.2 and 6.3%. This is generally lower than reported in the literature for non-delayed presentation head injury populations.

CONCLUSIONS

Available evidence suggests that head injury patients who present in a delayed fashion to the ED may have lower rates of intra-cranial injury compared to non-delayed head injury patients. However, the evidence is sparse and it is of too low quality to guide clinical practice. Further research is required to help the clinical risk assessment of this group.

TRIAL REGISTRATION

PROSPERO

CRD42015016135.