Pandemic Respiratory Infection Emergency System Triage (PRIEST).
Respiratory infections, such as influenza or the coronavirus, affect the lungs and airways, causing symptoms including fever, sore throat, coughing and breathing difficulties. If a new strain of the virus becomes widespread across many countries, this can be classified as a pandemic. During a pandemic, more patients attend hospital services and require investigation or admission, which puts a huge strain on the NHS.
Patients who contact emergency services (via telephone triage, emergency ambulance services or emergency departments) with a suspected pandemic respiratory infection need to be rapidly assessed to determine the severity of their illness, and whether they need to attend or be admitted to the hospital. This process is called "triage".
Triage often uses methods such as scores or decision rules. In a pandemic, we need to know how well these scores or rules predict the risk of adverse outcomes and thus how useful they are in supporting decision-making.
Pandemic Respiratory Infection Emergency System Triage (PRIEST) consisted of two studies:
- Emergency Department (ED) PRIEST, involving patients attending the emergency department.
- Prehospital PRIEST, involving patients contacting the NHS 111 telephone triage service or emergency 999 ambulance service.
ED PRIEST aimed to evaluate the accuracy of existing triage scores or rules and then develop a new score for predicting adverse outcomes, defined as death or need for life-saving major organ support, in patients attending the emergency department with suspected COVID-19. We subsequently used data from the study to describe the use of Do Not Resuscitate decisions during the COVID-19 pandemic.
Prehospital PRIEST aimed to evaluate the accuracy of existing triage scores or rules, the new score developed in ED PRIEST, and decision-making during the pandemic, in terms of whether patients were provided with an urgent response or transport to a hospital.
The ED PRIEST study collected data from 22,445 people with suspected COVID-19 in the first wave of the pandemic across 70 emergency departments in the UK. Research nurses reviewed hospital records to determine whether patients had died or received major organ support (heart, lung or kidneys) by 30 days after their initial hospital attendance.
The prehospital PRIEST study used routine data from telephone triage contacts, emergency ambulance calls, and ambulance patient report forms relating to cases contacting the Yorkshire Ambulance Service with suspected COVID-19 during the first wave of the pandemic. These data were linked to Office for National Statistics death registration data, NHS Digital hospital and general practice electronic health care data to determine whether patients had died or received major organ support by 20 days after their initial ambulance service contact.
Among existing rules or scores, CURB-65 (c-statistic 0.75), PMEWS (0.77) and NEWS2 (0.77) provide good, but not excellent, predictions for adverse outcomes in patients presenting to the emergency department with suspected COVID-19.
The PRIEST clinical severity score (consisting of the NEWS2 score, with added points for older age, male sex, and performance status below unrestricted normal activity) provides good accuracy for predicting adverse outcomes (c-statistic 0.8). A score of four or less could help to identify people with a low risk of adverse outcomes who do not need admission to a hospital (sensitivity 0.98, specificity 0.34).
Post-exertion oxygen saturation provides modest additional prognostic information in the assessment of selected patients attending the emergency department with suspected COVID-19.
Among adults admitted to a hospital with suspected COVID-19 during the first wave, 31% had a Do Not Attempt Resuscitation (DNAR) order recorded on or before their day of admission. These DNAR decisions were associated with recognised predictors of adverse outcomes and were inversely associated with Asian ethnicity. Most people with an early DNAR decision survived to 30 days and many received potentially life-saving interventions.
The decision to transport patients with suspected COVID-19 to a hospital after attendance by emergency ambulance had a sensitivity of 0.84 and specificity of 0.39 for predicting adverse outcomes. The PRIEST clinical severity score had sensitivity and specificity of 0.97 and 0.41 respectively, and thus had the potential to improve decision-making.
Patients contacting NHS111 with suspected COVID-19 had a 3% risk of adverse outcomes. Of these, 60% were advised to self-care or receive non-urgent clinical assessment, with a 1.3% risk of adverse outcomes. Repeat contact with telephone services was an important but under-recognised predictor of subsequent adverse outcomes.
Patients contacting 999 with suspected COVID-19 had an 11.1% risk of adverse outcomes. Of these, 16% were triaged to a non-urgent response, with a 3.5% risk of adverse outcomes.
A living systematic review of prediction models for COVID-19 (https://www.bmj.com/content/369/bmj.m1328.long) identified the PRIEST score as the only one of the eight models identified with adequate performance and low risk of bias that had been developed for emergency department triage.
The PRIEST score can be found on MDCALC, where it can be calculated and used by clinicians to support decision-making.
The PRIEST score was used by Yorkshire Ambulance Service senior clinical support cells to help assess the need for hospital conveyance in patients with suspected COVID-19.
The finding that repeated contact with NHS-111 telephone services may be associated with an unrecognised risk of deterioration was fed back to Yorkshire Ambulance Service to inform updates to NHS-111 triage pathways.
Funding from the Bill and Melinda Gates Foundation, through the International COVID-19 Data Alliance, is being used to share PRIEST data with collaborators in South Africa and Sudan, and develop triage tools that can be used in low and middle-income countries (https://icoda-research.org/project/dp-priest/).
This project is funded by the National Institute for Health Research Health Technology Assessment (NIHR HTA) Programme (project number 11/46/07). Any views or opinions expressed are those of the authors and do not necessarily reflect those of the HTA programme, NIHR, NHS or the Department of Health.
|Professor Steve Goodacre||ScHARR||Chief Investigator||+44 114 222 email@example.com|
|Carl Marincowitz||ScHARR||Chief Investigator (Pre-hospital PRIEST)||firstname.lastname@example.org|
|Dr Benjamin Thomas||ScHARR||Study Manager||+44 114 22 email@example.com|
|Katie Biggs||ScHARR||CTRU Oversight||+44 114 22 firstname.lastname@example.org|
|Laura Sutton||ScHARR||Statistician||+44 114 222 email@example.com|
|Amanda Loban||ScHARR||Data Management||+44 114 222 firstname.lastname@example.org|
|Josie Bourke||ScHARR||Trials Support Officer||+44 114 222 email@example.com|
|Dr Kirsty Challen||Lancashire Teaching Hospitals NHS Foundation Trust||Consultant in Emergency Medicinefirstname.lastname@example.org|
|Dr Andrew Bentley||University Hospital of South Manchester||Critical Care Expertise||+44 161 291 email@example.com|
|Dr Darren Walter||University Hospital of South Manchester||Expert in Pandemic Emergency Planning||+44 161 291 firstname.lastname@example.org|
|Dr Ian Maconochie||St Mary's Hospital, London||Expert in Paediatric Emergency Medicine||+44 207 886 email@example.com|
|Dr Chris Fitzsimmons||Sheffield Children's Hospital||Expert in Paediatric Emergency Medicine||+44 114 271 firstname.lastname@example.org|
|Andrew Lee||ScHARR, The University of Sheffield||Senior Clinical University Teacher||+44 114 222 email@example.com|
|Fiona Lecky||ScHARR, The University of Sheffield||Principal Investigator||+44 114 222 firstname.lastname@example.org|
|Tim Harris||Barts Health NHS Trust||Principal Investigatoremail@example.com|
The University’s four flagship institutes bring together our key strengths to tackle global issues, turning interdisciplinary and translational research into real-world solutions.