usIng News2 sCorIng System In coVid ninEteen - The INCISIVE Study  

Partners 

UHB, UoB, The Turing, UCL, UCLH, Oxford

Lay Summary: 

When patients present to hospital, there are standard scores that are used to alert medical professionals to patients who are at risk of deteriorating. These scores help identify which patients are the most unwell, who is responding to treatment, and who might need more help such as breathing support.   

 

These scores are calculated using  a number of different vital signs, such as heart rate, breathing rate, blood pressure and what the oxygen levels are in the blood.   The National Early Warning Score 2 (NEWS2) is one such score which is now used in all NHS hospitals in England, with every patient having a NEWS2 score calculated when they come into hospital.    

 

NEWS2 was developed before the current COVID-19 pandemic and it is unclear whether it is appropriate to use NEWS2 in COVID-19 patients.   

 

We will assess whether NEWS2 successfully identifies COVID-19 patients at most of risk of becoming more unwell and ensures that they receive the care they need. 

 

NEWS2 is a summary score of six physiological parameters or “vital signs”: respiratory rate, 

oxygen saturation, systolic blood pressure, heart rate, level of consciousness, temperature and supplemental oxygen dependency.   Each measure is given a score based on how abnormal the result is and that score is combined for an overall score.  NEWS2 has set thresholds, with a score of >5 requiring an urgent clinical review.  A central advantage of NEWS2 is its standardization. Medical professionals move hospitals and wards frequently so having a single score that is used everywhere minimizes the risk of confusion/errors.  

 

Unfortunately, there have been major changes in clinical practice during the pandemic which may make the interpretation of this score and the standard thresholds for increased observation or treatment less effective than when it is used for non-COVID medical emergencies.   

 

This project will assess how accurate NEWS2 is in identifying COVID-19 patients who are unwell, and how the score could be improved.    

 

Members of the DECOVID Analytics Workstream will bring expertise in statistical modelling 

together with clinical experience in how NEWS2 is used in practice. The project will answer three specific questions: 

 

1. What is the performance of NEWS2 in COVID-19 patients compared to non-COVID-19 

patients using conventional markers of patient deterioration? How are these outcomes 

different across time? How does performance change when we consider escalations in clinical care as a result of the NEWS2 score? 

2. Why is NEWS2 less effective in COVID-19?  Is this because of the types of patients admitted or how COVID-19 alters their vital signs?    How can the NEWS2 scores and thresholds be improved to support COVID-19 patients? For example by including different measures or changing the scoring system for each vital sign?  

3. Can the performance of NEWS2 be improved to better ……….,  if it is measured not just once, but a few times when the patient in in hospital?  What change in NEWS2 is associated with the best and worst clinical outcomes for patients?  

Scientific Summary: 

NEWS2 is a summary score of six physiological parameters or “vital signs”: respiratory rate, 

oxygen saturation, systolic blood pressure, heart rate, level of consciousness, temperature and supplemental oxygen dependency. It is used to alert medical professionals to patients at risk of early clinical deterioration in UK NHS hospitals. A central advantage of NEWS2 is its standardization. Having a single score that is used everywhere minimizes the risk of confusion/errors. NEWS2 comes with clinical thresholds at which clinical personnel should review the patient and amendcare if necessary. 

The performance of NEWS2 as an early warning score for COVID-19 patients, and the consequent impact on their care, is unclear. The clinical presentation of COVID-19 is unusual in that patients are relatively insensitive to hypoxaemia. There is concern that NEWS2, which was developed in the general patient population over many years, may not accurately identify hospitalized COVID-19 patients who need an escalation of their care. Verification of the performance of NEWS2 score in the COVID-19 population will be important information for hospitals nationally. 

 

Early warning scores are conventionally validated against a composite endpoint of cardiac arrest, death or intensive care unit (ICU) admission. These endpoints are suboptimal but are necessary to allow NEWS2 to be applied nationally in hospitals that use paper records. The endpoints cause additional challenges to the interpretation of NEWS2 during the pandemic as hospitals are adapting many wards into ICU units. Validation of NEWS2 against process metrics would not only improve the state-of-the-art practice in the field of EWS validation but also inform national policy. 

 

In addition to characterising the overall performance of NEWS2 it is important to identify 

situations in which inpatient deterioration may be missed and suggest strategies to improve it. Recognised situations where the NEWS2 score leads to underappreciation of disease severity are patients on a deterioration trajectory that have not yet developed extremely abnormal physiology, and patients where their physiology has been pseudo-normalised by escalating oxygen therapy.  Single centre descriptive analyses have shown these issues to be present in COVID-19 patients. Providing solutions to address these issues would improve patient safety. 

 

We aim to evaluate the performance of the National Early Warning Score NEWS2, and its clinical. thresholds, in patients hospitalised with COVID-19 compared to hospitalised patients without COVID-19. 

 

We will investigate five specific research questions: 

1. What is the performance of NEWS2 in COVID-19 patients compared to non-COVID-19 

patients using conventional markers of patient deterioration (cardiac arrest/ICU admission/ 

death) within 6h, 12h and 24h? 

2. Does our assessment of the performance difference change if we account for escalations 

in clinical care (e.g. increased monitoring or initiation of ventilation) that are a direct 

consequence of NEWS2, either as an outcome in their own right or on the causal pathway to 

an eventual outcome (e.g. death)? 

3. Can performance of NEWS2 be improved by including its components (e.g. oxygen therapy) in a more graduated manner than the current binary threshold? 

4. What are the characteristics and physiological state of patients where a poor outcome is not predicted by their NEWS2 score, for example by considering additional parameters such as C-reactive protein (CRP), a commonly used biomarker of sepsis? 

5. Can performance of NEWS2 be improved by incorporating information about trajectories, 

either of the NEWS2 score or of its vital sign components?