Keywords
COVID-19 - NEWS2 - clinical deterioration - patient outcomes - early warning scores
Introduction
In numerous areas of medicine, the development of clinical prediction models is prevalent.[1] In 2000, acute hospitals implemented early warning scores (EWSs) for the first time.
Today, they are widely used in various clinical contexts as they provide extremely
valuable information to help characterize the clinical risk of deterioration more
precisely, thereby facilitating risk stratification. Although 99% of acute hospitals
utilize EWS to monitor patients[2] (the National Confidential Enquiry into Patient Outcome and Death [NCEPOD], 2015),
there has been little progress over the past decade in identifying and responding
to deteriorating patients.[3]
EWS systems have been established and widely used in developed nations to identify
clinical deterioration as soon as possible.[4] They are common in Australia, the United States, and the Netherlands.[5]
The present study aims to analyze the influence of applying the National Early Warning
Score (NEWS) on coronavirus disease 2019 (COVID-19) patient outcomes, such as patient
deterioration, admission to an intensive care unit (ICU), length of hospital stays,
and patient mortality. We hypothesized that implementing NEWS2 would significantly
improve hospitalized COVID-19 patient outcomes by decreasing the length of hospital
stay, predicting ICU admissions, reducing the mortality rate, and increasing the frequency
of vital sign measurements.
Patients and Methods
Design and Setting
Among hospitalized COVID-19 patients, a quasi-experimental research design was utilized.
This study was performed at inpatient units in the COVID-19 Martyrs Medical Military
Complex and the Palestinian Red Crescent Society Hospital, which also functioned as
a certified testing and treatment facility for COVID-19 and is located in Nablus,
Palestine.
Study Population
About 384 adult male and female patients who satisfied the inclusion criteria were
divided into two groups: the study group (post-NEWS2) (192 patients) and the control
group (pre-NEWS2) (192 patients).
By reverse transcriptase polymerase chain reaction, severe acute respiratory syndrome
coronavirus-2 was detected in all adult patients admitted to our research with a confirmed
diagnosis of COVID-19. Patients who were younger than 18 years, pregnant women, those
with spinal cord injuries, non-COVID-19 patients, and patients who were readmitted
to the unit during the trial were excluded from the study. The NEWS should not be
applied to children or pregnant women since their physiological response to acute
illness may be affected. Due to functional abnormalities in the autonomic nervous
system, spinal cord injury patients may not be able to rely on the NEWS (especially
tetraplegia and severe paralysis). These criteria were developed by the Royal College
of Physicians' (RCP) 2017 standards.[6]
All adult COVID-19 patients admitted to both COVID-19 centers and those who satisfied
the criteria for inclusion were enrolled in the study during the study periods: July
to November 2021 for the pre-NEWS phase and December 2021 to March 2022 for the post-NEWS
phase. These patients were observed daily for improvement or deterioration. The researcher
categorized patient outcomes until death or discharge. The sample was selected using
the following formula, with a 95% level of significance and a margin of error of 0.05:
n = z
2 × p (1 − p)/1 + z
2 × p (1 − p) /e
2 × N; n = (1.96) 2/4 (0.05) 2 = 384.16
Where p is the proportion of the population predicted to be 0.5, e is the margin of error, N is the size of the people, and z is the z-score associated with the selected significance level. The equation yielded
a sample size of n = 384.
Validity and Reliability
The NEWS2 is the standard track and trigger early warning system used in the United
Kingdom and provides a uniform observation system for patient care continuity. Ambulance
services, general practice, and community assessments also use it.[7] In this study, NEWS2 was superior in its area under the receiver operating characteristic
(AUROC) curve when predicting ICU admission, and the cutoff point for NEWS2 to predict
in-hospital mortality was higher.
Study Protocol
For 2 weeks, using lectures, group discussions, and clinical scenarios, the researcher
taught and monitored an educational program where nurses practiced using and implementing
the NEWS system. The researcher educated nurses for 2 weeks on how to apply the NEWS
system through lectures, group discussions, and clinical situations. The program's
goals were to create a framework for communication among health care professionals
and aid in knowing the physiological measurements, the rationale behind the size,
and vital sign abnormalities. The program covered the advantages of NEWS, how it works,
the reasons for its use, the score of physiological parameters, its thresholds and
triggers, and its clinical response. The score consists of six physiological measures,
including respiratory rate, oxygen saturation, pulse rate, systolic blood pressure,
body temperature, and consciousness or new confusion. The study instruments are detailed
in [Appendix 1].
Appendix 1
Description of the study instruments
|
Part 1. The patient's medical record, age, gender, admitting diagnosis, prior medical history,
patient score, hospital stay duration, and admission date were extracted for the medical
and demographic data sheets.
|
|
Part 2. NEWS2 is a scale that can “monitor and trigger.” It is an aggregate scoring system
with multiple factors that measure, from 0 to 3, the deviation from expected normal
levels of respiratory rate, oxygen saturation, systolic blood pressure, heart rate,
body temperature, and degree of awareness. This is how the AVPU concept evaluates
awareness: A = alert, V = verbal stimulus reaction, P = pain stimulus response, and
U = unresponsive. Three points are awarded for the possibility of consciousness-level
alteration. Each of these features is assigned a score; the scores are tallied together,
and the eventual addition of additional oxygen increases the score by two points.
The correlation between the patient's clinical risk and the total score points is
as follows: low-risk (aggregate scores 1–4): immediate assessment by the ward nurse
within 6 h to determine a change in frequency of monitoring or escalation of clinical
treatment; moderate-risk (score of more than 5 or 3 in any single parameter): urgent
evaluation by ward-based nurse hourly and recheck with other nurse or doctor to discover
the cause and to decide on modification in monitoring frequency or escalation of clinical
care; and high-risk score (7 or more): indicates that the NEWS escalation protocol
will be started when the nurse contacts the physician in charge of a patient who requires
an immediate assessment by the critical care team, typically resulting in patient
transfer to higher dependency care settings. The escalation procedure assists clinicians
in making clinical decisions. Recommendations are forwarded regarding the level of
qualifications required for the condition, the evaluation interval, the level of treatment,
and the activation of a medical emergency team.
|
|
Part 3. A patient outcome observational checklist was subdivided into primary outcomes. Patient
deterioration represented the physiological parameters that were included in the NEWS
measures. The use of oxygen support therapy, vaccination status, and the frequency
of monitoring or any added special concerns or medical reviews by the health care
providers were secondary outcomes, including ICU admission, length of hospital stay,
and mortality. Each patient in the trial group received it. These data were used daily
after utilizing the NEWS compared with the documented patient's clinical condition,
use of oxygen support therapy, vaccination status, frequency of monitoring, length
of hospital stay, admission to the ICU, and discharge status of the patient (alive
or dead) for the control pre-NEWS implementation phase group.
|
Abbreviations: ICU, intensive care unit; NEWS, National Early Warning Score.
A score is assigned to each observation. When all six scores are combined, the overall
NEWS is generated and configured to activate when a patient is seriously ill or has
abnormal physiology. Each physiological parameter is assigned a score, with the value
of the score representing how much the parameter deviates from the norm. A NEWS2 score
of 5 or 6 is regarded as a critical threshold that may signal clinical worsening and
should necessitate an immediate reaction by a physician or team with expertise in
assessing and treating severely ill patients. An increased NEWS does not establish
a diagnosis, but it does assist in identifying the risky patient who systematically
requires immediate clinical assessment. The risk levels are classified as low, low-medium,
medium, or high, with suggestions for appropriate clinical responses and more intensive
treatment. It also provides a consistent monitoring system and continuity of treatment
for patients between wards. Trends in a patient's clinical response can be monitored
by regularly recording the score and providing early notifications of clinical deterioration.
When early warning systems are used in close collaboration with the medical team,
the patients are identified as needing special consideration. A specialist team can
treat them immediately, lowering the potential adverse outcomes.[8]
Statistical Analysis
Application version 25 of the Statistical Package for the Social Sciences (SPSS) was
used to tabulate and analyze the data after data collection. The quantitative data
were reported using means and standard deviations. The independent sample t-test was used to assess group homogeneity and compare the mean frequency of patient
monitoring in both groups. The significance level was established at 5% (p = 0.05). The homogeneity of the two groups in terms of patient outcomes and NEWS
escalation categories was assessed using the chi-square proportional test. Using a
post hoc chi-square test, it was determined which groups were substantially distinct.
Results
Characteristics of Patients
Patients in the control (pre-NEWS2) group were slightly but not significantly older
than those in the study (post-NEWS2) group, as shown in [Tables 1] and [2]. The mean age of patients in the pre-NEWS phase was 62.1 ± 13.9 years, and the mean
age in the post-NEWS phase was 58.1 ± 17.7 years (p = 0.015). On the other hand, more patients were 60 years and older (p = 0.040) in the pre-NEWS phase compared with those admitted during the post-NEWS
phase. Patients in both groups did not differ significantly (p > 0.05) in terms of gender, chronic kidney diseases, chronic liver diseases, chronic
lung diseases, or cardiovascular diseases.
Table 1
Characteristics of patients included in the pre- and post-NEWS phases of the study
|
Variable
|
Pre-NEWS[a]
|
Post-NEWS[a]
|
Chi-square test
|
p-Value
|
|
Number
|
192
|
192
|
–
|
–
|
|
Age, y: < 60/≥ 60
|
78 (40.6%)/114 (59.4%)
|
99 (51.6%)/93 (48.4%)
|
4.62
|
0.04
|
|
Sex: female/male
|
98 (51%)/94 (50%)
|
105 (54.7%)/87 (45.3%)
|
0.51
|
0.54
|
|
Vaccination: none
|
101 (52.6%)
|
101 (52.6%)
|
0.0
|
1.0
|
|
Vaccination: one dose
|
67 (34.9%)
|
67 (34.9%)
|
|
Vaccination: two doses
|
24 (12.5%)
|
24 (12.5%)
|
|
Vaccination: two doses + a booster
|
0 (0.0%)
|
0 (0.0%)
|
|
Diabetes mellitus: no/yes
|
89 (46.4%)/103 (53.6%)
|
115 (59.9%)/77 (40.1%)
|
7.07
|
0.01
|
|
Hypertension: no/yes
|
71 (37%)/121 (63%)
|
119 (62%)/73 (38%)
|
24.0
|
0.01
|
|
CKD: no/yes
|
172 (89.6%)/20 (10.4%)
|
183 (95.3%)/9 (4.7%)
|
4.51
|
0.052
|
|
Cancer: no/yes
|
180(93.8%)/12 (6.3%)
|
166 (86.5%)/26 (13.5%)
|
5.72
|
0.025
|
|
Chronic liver disease: no/yes
|
174 (90.6%)/18 (9.4%)
|
174 (90.6%)/18 (9.4%)
|
0
|
1.0
|
|
Autoimmune disease: no/yes
|
155 (80.7%)/37 (19.3%)
|
174 (90.6%)/18 (9.4%)
|
7.66
|
0.008
|
|
Chronic lung disease: no/yes
|
176 (91.7%)/16 (8.3%)
|
178 (92.7%)/14 (7.3%)
|
0.14
|
0.85
|
|
Cardiovascular disease: no/yes
|
132 (68.8%)/60 (31.3%)
|
130(67.7%)/62 (32.3%)
|
0.05
|
0.913
|
Abbreviations: CKD, chronic kidney disease; NEWS, National Early Warning Score.
a Values are expressed as number (percentage).
Table 2
COVID-19 features included in the pre- and post-NEWS phases of the study
|
Variable
|
Pre-NEWS
|
Post-NEWS
|
Chi-square test
|
p-Value
|
|
Number
|
192
|
192
|
–
|
–
|
|
Low-flow oxygen[a]: no/yes
|
10 (5.2%)/182 (94.8%)
|
42 (21.9%)/150 (78.1%)
|
22.78
|
< 0.001
|
|
High-flow nasal cannula no/yes
|
152 (79.2%)/40 (20.8%)
|
135 (70.3%)/57 (29.7%)
|
3.99
|
0.06
|
|
NIPPV: no/yes
|
166 (86.5%)/26 (13.5%)
|
167 (87.0%)/25 (13.0%)
|
0.02
|
1
|
|
Hospital LOS (d): < 6/≥ 6
|
80 (41.7%)/112 (58.3%)
|
112 (58.3%)/80 (41.7%)
|
10.67
|
0.002
|
|
Mechanical ventilation: no/yes
|
178 (92.7%)/14 (7.3%)
|
175 (91.1%)/17 (8.9%)
|
0.32
|
0.354
|
|
ICU admission: no/yes
|
143 (74.5%)/49 (25.5%)
|
135 (70.3%)/57 (29.7%)
|
0.83
|
0.424
|
|
Discharged alive: no/yes
|
38 (19.8%)/154 (80.2%)
|
24 (12.5%)/168 (87.5%)
|
3.77
|
0.071
|
Abbreviations: COVID-19, coronavirus disease 2019; ICU, intensive care unit; LOS,
length of stay; NEWS, National Early Warning Score; NIPPV, noninvasive positive-pressure
ventilation.
Note: Values are expressed as number (percentage).
a SPO2 < 94%.
The distribution of vaccinated and unvaccinated patients was also similar in both
phases. However, more patients were 60 years and older (p = 0.040), had diabetes mellitus (p = 0.010), hypertension (p = 0.010), and autoimmune diseases (p = 0.008) in the pre-NEWS phase compared with those admitted during the post-NEWS
phase.
During the pre-NEWS phase, 38 (19.8%) patients died, compared with 24 (12.5%) during
the post-NEWS phase (p = 0.071). Of the patients admitted during the pre-NEWS phase, 49 (25.5%) required
admission to the ICU, compared with 57 (29.7%) during the post-NEWS phase (p = 0.424). In the pre-NEWS phase, more patients received low-flow oxygen (LFO) than
those admitted during the post-NEWS phase (p < 0.001). However, patients in both phases did not differ in terms of receiving a
high-flow nasal cannula (p = 0.060), noninvasive positive pressure ventilation (NIPPV) (p = 1.000), or mechanical ventilation (p = 0.354). Nonetheless, more patients stayed in the hospital for 6 days or more in
the pre-NEWS phase than those admitted during the post-NEWS phase (p = 0.002). The mean duration of hospital stay in the pre-NEWS phase was 8.1 ± 5.5
days, and the mean hospital stay in the post-NEWS phase was 6.4 ± 5.3 days (p = 0.002).
Association between Patients Outcomes and Different Variables in the Pre-NEWS Phase
[Table 3] shows associations between patient deterioration based on oxygen usage, admission
to the ICU, and outcomes based on the length of hospitalization and hospital discharge
status in the pre-NEWS phase. Receiving a high-flow nasal cannula, mechanical ventilation,
and staying in the hospital for 6 days or more were not significantly associated with
any of the variables such as age, gender, vaccination status, diabetes, cancer, hypertension,
chronic liver disease, chronic kidney disease, autoimmune disease, or cardiovascular
disease. However, receiving LFO was associated with having cancer (p = 0.002) and chronic liver disease (p = 0.008). Receiving NIPPV was not associated with chronic lung disease (p = 0.047). Admission to the ICU was significantly associated with having cancer (p = 0.002) and chronic liver disease (p = 0.021). Discharged dead were associated with older age (p = 0.009), autoimmune disease (p = 0.020), and cardiovascular disease (p = 0.020).
Table 3
Associations between patient deterioration and outcomes with variables in the pre-NEWS
phase (N = 185)
|
Variable
|
|
Low-flow oxygen
|
High-flow nasal cannula
|
NIPPV
|
Mechanical ventilation
|
Length of stay ≥ 6 d
|
ICU
|
Discharged alive
|
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
|
Age (y)
|
< 60
|
75
|
0.49
|
0.533
|
17
|
0.07
|
0.857
|
8
|
1.21
|
0.293
|
3
|
2.31
|
0.163
|
39
|
3.75
|
0.073
|
14
|
3.96
|
0.063
|
70
|
7.52
|
0.009
|
|
≥ 60
|
107
|
|
|
23
|
|
|
18
|
|
|
11
|
|
|
73
|
|
|
35
|
|
|
84
|
|
|
|
Gender
|
Female
|
95
|
1.86
|
0.206
|
17
|
1.48
|
0.286
|
13
|
0.01
|
1.000
|
8
|
0.23
|
0.783
|
53
|
1.49
|
0.244
|
21
|
1.76
|
0.191
|
80
|
0.26
|
0.718
|
|
Male
|
87
|
|
|
23
|
|
|
13
|
|
|
6
|
|
|
59
|
|
|
28
|
|
|
74
|
|
|
|
Vaccine dose
|
None
|
93
|
3.50
|
0.171
|
21
|
1.36
|
0.490
|
16
|
1.15
|
0.531
|
7
|
0.06
|
0.971
|
64
|
3.50
|
0.168
|
27
|
0.18
|
0.945
|
81
|
3.65
|
0.160
|
|
One
|
65
|
|
|
12
|
|
|
8
|
|
|
5
|
|
|
33
|
|
|
16
|
|
|
57
|
|
|
|
Two
|
24
|
|
|
7
|
|
|
2
|
|
|
2
|
|
|
15
|
|
|
6
|
|
|
16
|
|
|
|
DM
|
Yes
|
96
|
1.17
|
0.344
|
21
|
0.03
|
1.000
|
13
|
0.16
|
0.833
|
9
|
0.69
|
0.579
|
64
|
1.32
|
0.304
|
31
|
2.45
|
0.137
|
78
|
2.81
|
0.105
|
|
Hypertension
|
Yes
|
115
|
0.04
|
1.000
|
24
|
0.20
|
0.714
|
19
|
1.31
|
0.283
|
8
|
0.22
|
0.775
|
70
|
0.03
|
0.881
|
34
|
1.14
|
0.309
|
96
|
0.16
|
0.713
|
|
ESRD
|
Yes
|
18
|
1.03
|
0.309
|
6
|
1.13
|
0.380
|
0
|
3.48
|
0.081
|
0
|
1.75
|
0.369
|
9
|
1.63
|
0.235
|
7
|
1.06
|
0.415
|
15
|
0.38
|
0.556
|
|
Cancer
|
Yes
|
8
|
20.40
|
0.002
|
3
|
0.13
|
1.000
|
2
|
0.11
|
1.000
|
1
|
0.02
|
1.000
|
12
|
9.14
|
0.004
|
8
|
11.34
|
0.002
|
8
|
1.47
|
0.259
|
|
Chronic liver
|
Yes
|
14
|
11.59
|
0.008
|
7
|
3.91
|
0.065
|
2
|
0.10
|
1.000
|
1
|
0.09
|
1.000
|
15
|
5.11
|
0.042
|
9
|
6.23
|
0.021
|
12
|
2.28
|
0.209
|
|
Autoimmune
|
Yes
|
35
|
0.00
|
1.000
|
7
|
0.10
|
0.826
|
3
|
1.16
|
0.309
|
1
|
1.42
|
0.312
|
20
|
0.35
|
0.582
|
5
|
3.48
|
0.091
|
35
|
5.98
|
0.020
|
|
Chronic lung
|
Yes
|
14
|
1.87
|
0.198
|
5
|
1.14
|
0.333
|
5
|
4.65
|
0.047
|
1
|
0.03
|
1.000
|
10
|
0.13
|
0.796
|
5
|
0.30
|
0.765
|
14
|
0.58
|
0.537
|
|
CVD
|
Yes
|
58
|
0.62
|
0.511
|
15
|
0.92
|
0.444
|
10
|
0.73
|
0.495
|
6
|
0.94
|
0.373
|
38
|
0.90
|
0.430
|
18
|
0.92
|
0.374
|
42
|
5.73
|
0.020
|
Abbreviations: CVD, cardiovascular disease; DM, diabetes mellitus; ESRD, end-stage
renal disease; ICU, intensive care unit; NEWS, National Early Warning Score; NIPPV,
noninvasive positive pressure ventilation.
Relationship between Patients' Outcomes and Variables in the Post-NEWS Phase
[Table 4] shows associations between patient deterioration based on oxygen usage, admission
to the ICU, outcomes based on length of stay in the hospital, and hospital discharge
status with different variables in the post-NEWS phase. Receiving LFO was associated
with older age (p = 0.001), diabetes (p = 0.001), and hypertension (p = 0.006). Receiving a high-flow nasal cannula was significantly associated with older
age (p = 0.011), diabetes (p = 0.004), and hypertension (p < 0.001). Furthermore, receiving NIPPV was significantly associated with older patients
(p < 0.001), male patients (p = 0.018), diabetes (p = 0.047), and hypertension (p = 0.025). Receiving mechanical ventilation was significantly associated with older
age (p < 0.001). Staying 6 days or more was significantly associated with older age (p = 0.041), diabetes (p = 0.004), hypertension (p < 0.001), and cardiovascular disease (p = 0.029). Admission to the hospital was significantly associated with older age (p < 0.001), male patients (p = 0.027), diabetes (p < 0.001), hypertension (p < 0.001), chronic liver disease (p = 0.005), and cardiovascular disease (p = 0.012). Discharged dead were associated with older age (p < 0.001) and cancer (p = 0.026).
Table 4
Associations between patient deterioration and outcomes with variables in the post-NEWS
phase
|
Variable
|
|
Low-flow oxygen
|
High-flow nasal cannula
|
NIPPV
|
Mechanical ventilation
|
Length of stay ≥ 6 d
|
ICU
|
Discharged alive
|
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
N
|
Chi-square test
|
p-Value
|
|
Age (y)
|
< 60
|
68
|
10.65
|
0.001
|
21
|
7.03
|
0.011
|
4
|
14.55
|
0.000
|
0
|
19.86
|
0.000
|
34
|
4.51
|
0.041
|
19
|
10.79
|
0.001
|
98
|
24.67
|
0.000
|
|
≥ 60
|
82
|
|
|
36
|
|
|
21
|
|
|
17
|
|
|
46
|
|
|
38
|
|
|
70
|
|
|
|
Gender
|
Female
|
79
|
1.13
|
1.130
|
27
|
1.75
|
0.206
|
8
|
5.97
|
0.018
|
6
|
2.83
|
0.126
|
41
|
0.65
|
0.463
|
24
|
5.18
|
0.027
|
96
|
3.27
|
0.082
|
|
Male
|
71
|
|
|
30
|
|
|
17
|
|
|
11
|
|
|
39
|
|
|
33
|
|
|
72
|
|
|
|
Vaccine dose
|
None
|
84
|
3.83
|
0.162
|
30
|
0.00
|
1.000
|
15
|
1.69
|
0.469
|
11
|
1.16
|
0.588
|
44
|
0.37
|
0.856
|
32
|
0.45
|
0.814
|
84
|
4.42
|
0.110
|
|
One
|
50
|
|
|
20
|
|
|
6
|
|
|
4
|
|
|
27
|
|
|
18
|
|
|
63
|
|
|
|
Two
|
16
|
|
|
7
|
|
|
4
|
|
|
2
|
|
|
9
|
|
|
7
|
|
|
21
|
|
|
|
DM
|
Yes
|
70
|
12.29
|
0.001
|
32
|
8.68
|
0.004
|
15
|
4.74
|
0.047
|
10
|
2.72
|
0.122
|
42
|
8.77
|
0.004
|
37
|
20.77
|
0.000
|
63
|
3.79
|
0.073
|
|
Hypertension
|
Yes
|
65
|
8.21
|
0.006
|
34
|
16.09
|
0.000
|
15
|
5.89
|
0.025
|
9
|
1.76
|
0.200
|
45
|
19.34
|
0.000
|
38
|
28.23
|
0.000
|
60
|
3.03
|
0.114
|
|
ESRD
|
Yes
|
8
|
0.64
|
0.687
|
3
|
0.06
|
1.000
|
1
|
0.03
|
1.000
|
0
|
0.91
|
0.610
|
4
|
0.03
|
1.000
|
4
|
0.98
|
0.454
|
7
|
0.81
|
0.605
|
|
Cancer
|
Yes
|
22
|
0.74
|
0.457
|
8
|
0.02
|
1.000
|
5
|
1.02
|
0.345
|
4
|
1.58
|
0.256
|
12
|
0.25
|
0.672
|
8
|
0.02
|
1.000
|
19
|
5.69
|
0.026
|
|
Chronic liver
|
Yes
|
15
|
0.31
|
0.768
|
2
|
3.28
|
0.102
|
0
|
2.96
|
0.136
|
0
|
1.92
|
0.233
|
7
|
0.06
|
1.000
|
0
|
8.39
|
0.005
|
18
|
2.82
|
0.135
|
|
Autoimmune
|
Yes
|
16
|
1.34
|
0.371
|
4
|
0.53
|
0.593
|
2
|
0.06
|
1.000
|
2
|
0.13
|
1.000
|
8
|
0.06
|
1.000
|
3
|
1.61
|
0.282
|
16
|
0.04
|
1.000
|
|
Chronic lung
|
Yes
|
13
|
1.91
|
0.203
|
3
|
0.49
|
0.561
|
1
|
0.46
|
0.699
|
1
|
0.06
|
1.000
|
4
|
1.07
|
0.403
|
2
|
1.71
|
0.238
|
13
|
0.39
|
0.702
|
|
CVD
|
Yes
|
53
|
2.90
|
0.096
|
22
|
1.47
|
0.240
|
9
|
0.18
|
0.819
|
6
|
0.08
|
0.790
|
33
|
5.03
|
0.029
|
26
|
6.58
|
0.012
|
54
|
0.01
|
1.000
|
Abbreviations: CVD, cardiovascular diseases; DM, diabetes mellitus; ESRD, end-stage
renal disease; ICU, intensive care unit; NEWS, National Early Warning Score; NIPPV,
noninvasive positive pressure ventilation.
Relationship between NEWSs and Variables
Student's t-test and analysis of variance showed that the mean NEWSs were higher for patients
who were 60 years and older (p < 0.001), had diabetes (p < 0.001), hypertension (p < 0.001), had cardiovascular disease (p = 0.011), received LFO (p = 0.001), NIPPV (p < 0.001), mechanical ventilation (p < 0.001), were admitted to the ICU (p < 0.001), and discharged dead (p < 0.001). There was a positive correlation between age and the NEWS (Pearson's r = 0.473, p < 0.001). Additionally, there was a positive correlation between NEWSs and length
of hospital stay (Pearson's r = 0.184, p < 0.001) ([Table 5]). A multiple linear regression model included the variables significantly associated
with the student's t-tests to identify the significant predictors of high and low NEWSs. [Tables 6] and [7] show that high NEWSs were significantly predicted by the need to receive NIPPV (p = 0.013), stay in the hospital for 6 or more days (p = 0.013), be admitted to the ICU (p = 0.006), and be discharged dead (p < 0.001).
Table 5
Relationship between mean NEWSs and variables
|
Variables
|
N (%)
|
Mean (SD)
|
p-Value
|
|
Age (y): < 60/≥ 60
|
99 (51.6%)/93 (48.4%)
|
3.3 (1.5)/5.0 (2.7)
|
< 0.001
|
|
Sex: Female/male
|
105 (54.7%)/87 (45.3)
|
3.9 (2.2)/4.5 (2.4)
|
0.083
|
|
Vaccination status: none
|
101 (52.6%)
|
4.5 (0.2)
|
0.052
|
|
Vaccination status: one dose
|
67 (34.9%)
|
3.6 (0.2)
|
|
Vaccination status: two doses
|
24 (12.5%)
|
4.2 (0.5)
|
|
Vaccination status: two doses + a booster
|
0 (0.0%)
|
NA
|
|
Diabetes mellitus: no/yes
|
115 (59.9%)/77 (40.1%)
|
3.6 (2.0)/4.9 (2.5)
|
< 0.001
|
|
Hypertension: no/yes
|
119 (62.0%)/73 (38.0%)
|
3.6 (2.1)/5.0 (2.4)
|
< 0.001
|
|
CKD: no/yes
|
183 (95.3%)/9 (4.7%)
|
4.1 (2.3)/5.2 (3.0)
|
0.145
|
|
Cancer: no/yes
|
166 (86.5%)/26 (13.5%)
|
4.1 (2.3)/4.6 (2.3)
|
0.332
|
|
Chronic liver disease: no/yes
|
174 (90.6%)/18 (9.4%)
|
4.3 (2.4)/3.1 (1.0)
|
0.043
|
|
Autoimmune disease: no/yes
|
174 (90.6%)/18(9.4%
|
4.2 (2.3)/3.7 (2.4)
|
0.418
|
|
Chronic lung disease: no/yes
|
178 (92.7%)/14 (7.3%)
|
4.1 (2.4)/4.6 (1.5)
|
0.466
|
|
Cardiovascular disease: no/yes
|
130 (67.7%)/62 (32.3%)
|
3.9 (2.2)/4.8 (2.4)
|
0.011
|
|
Use of low-flow oxygen[a]: no/yes
|
42 (21.9%)/150 (78.1%)
|
3.2 (2.7)/4/4 (2.1)
|
0.001
|
|
Use of high-flow nasal cannula: no/yes
|
135 (70.3%)/57 (29.7%)
|
3.3 (1.7)/6.2 (2.4)
|
< 0.001
|
|
Use of NIPPV: no/yes
|
167 (87.0%)/25 (13.0%)
|
3.6(1.6)/8.1/2.4)
|
< 0.001
|
|
Hospital LOS (d): < 6/≥ 6
|
112 (58.3%)/80 (41.7%)
|
3.2(1.5)/5.4 (2.6)
|
< 0.001
|
|
Mechanical ventilation: no/yes
|
175 (91.0%)/17 (8.9%)
|
3.7 (1.8)/8.6 (2.5)
|
< 0.001
|
|
ICU admission: no/yes
|
135 (70.3%)/57 (29.7%)
|
3.2 (1.3)/6.5 (2.4)
|
< 0.001
|
|
Discharged alive: no/yes
|
24 (12.5%)/168 (87.5%)
|
3.2 (1.3)/6.5 (2.4)
|
< 0.001
|
Abbreviations: CKD, chronic kidney disease; ICU, intensive care unit; LOS, length
of stay; NEWS, National Early Warning Score; NIPPV, noninvasive positive-pressure
ventilation; SD, standard deviation.
a SPO2 < 94%.
Table 6
Use of NEWSs to predict deterioration and outcomes of patients using both multiple
linear regression and ROC curve analysis
|
Variable
|
Unstandardized coefficients
|
SE
|
Standardized coefficients
|
T
|
p-Value
|
|
Age
|
0.25
|
0.24
|
0.06
|
1.07
|
0.287
|
|
Diabetes
|
0.14
|
0.25
|
0.03
|
0.58
|
0.565
|
|
Hypertension
|
−0.06
|
0.26
|
−0.01
|
−0.24
|
0.813
|
|
Chronic liver disease
|
−0.15
|
0.36
|
−0.02
|
−0.41
|
0.681
|
|
Cardiovascular disease
|
0.34
|
0.25
|
0.07
|
1.39
|
0.167
|
|
Low-flow oxygen
|
0.46
|
0.25
|
0.08
|
1.84
|
0.068
|
|
High-flow nasal cannula
|
0.29
|
0.33
|
0.06
|
0.90
|
0.368
|
|
NIPPV
|
1.35
|
0.54
|
0.20
|
2.52
|
0.013
|
|
Mechanical ventilation
|
−0.79
|
0.66
|
−0.10
|
−1.20
|
0.231
|
|
Length of hospital stay
|
0.60
|
0.24
|
0.13
|
2.51
|
0.013
|
|
ICU
|
1.03
|
0.38
|
0.21
|
2.75
|
0.006
|
|
Discharged alive
|
−3.20
|
0.48
|
−0.46
|
−6.73
|
< 0.001
|
Abbreviations: ICU, intensive care unit; NEWS, National Early Warning Score; NIPPV,
noninvasive positive pressure ventilation; ROC, receiver operating characteristic;
SE, standard error.
Table 7
Use of NEWSs to predict deterioration and outcomes of patients using ROC curve analysis
|
Variable
|
AUC
|
p-Value
|
Lower
|
Upper
|
Sensitivity
|
Specificity
|
|
Low-flow oxygen
|
0.74
|
< 0.001
|
0.64
|
0.84
|
0.527
|
0.810
|
|
High-flow nasal cannula
|
0.87
|
< 0.001
|
0.82
|
0.92
|
0.561
|
0.904
|
|
NIPPV
|
0.95
|
< 0.001
|
0.90
|
0.99
|
0.920
|
0.904
|
|
Hospital stay ≥ 6 d
|
0.78
|
< 0.001
|
0.72
|
0.85
|
0.550
|
0.812
|
|
ICU
|
0.91
|
< 0.001
|
0.87
|
0.96
|
0.702
|
0.896
|
|
Mechanical ventilation
|
0.94
|
< 0.001
|
0.88
|
1.00
|
0.882
|
0.897
|
|
Mortality
|
0.96
|
< 0.001
|
0.92
|
1.00
|
0.875
|
0.911
|
Abbreviations: AUC, area under the curve; ICU, intensive care unit; NEWS, National
Early Warning Score; NIPPV, noninvasive positive pressure ventilation; ROC, receiver
operating characteristic.
Discussion
To the best of our knowledge, most studies have yet to directly address the impact
of NEWS systems in identifying clinical deterioration and outcomes in COVID-19 hospitalized
patients.
In all fields of practice, it is vital to identify patients whose conditions are deteriorating.
In this investigation, the NEWS2 showed strong discrimination in predicting the combined
outcome of the patient's admission to the ICU, the patient's need for intensive respiratory
support as an enormous sign of deterioration, a lengthening hospital stay, and in-hospital
death.
The overall mean length of hospital stay after implementing the score was decreased,
from 8.1 ± 5.5 days in pre-NEWS phase to 6.4 ± 5.3 days in post-NEWS phase which in
line with those obtained in studies in countries other than China, which revealed
that the median length of stay in hospitals varied from 4 to 21 days, but in China,
the median length of stay in hospitals ranged from 4 to 53 days. However, an Egyptian
quasi-experimental study found that there were no statistically significant changes
in the total duration of hospital stays before or after the adoption of NEWS.[9]
Furthermore, the present study not only found a nonsignificant decrease in mortality
rate in the study group compared with the control groups but also less than the reported
mortality for all hospitalized COVID-19 patients in some European, American, and Chinese
studies, with a percentage ranging between 17 and 23.4%.[10]
[11]
[12]
[13]
[14]
[15]
Additionally, the present study's conclusion supports Farenden et al, who found nonsignificant
differences between deaths and outcomes before and after NEWS implementation.[16] But in contrast, Moon et al showed a substantial reduction in in-hospital mortality
following the deployment of the NEWS system.[17]
Although the NEWS2 method assesses supplemental oxygen as a binary variable, it cannot
distinguish between oxygen supply rates. The RCP has verified this issue, stating
that any increase in oxygen therapy among COVID-19 patients should prompt a physician
assessment and increased monitoring.[17] While it has been established that oxygen supplementation is an independent risk
factor for new coronavirus pneumonia escalating to a serious situation, no previous
studies have shown an association between NEWS and supplemental oxygen therapy. Some
studies have established a link between the oxygen saturation measures in the score
and death as a predictive physiological parameter in patients with COVID-19 infection.
Liu et al found that the oxygen saturation level had a good predictive performance
for predicting death.[18]
Despite the lack of published statistics on NEWS2's effectiveness in controlling COVID-19
patients' clinical deterioration, this study showed that NEWS2 was good at predicting
deteriorating patients based on their oxygen requirement, with an acceptable level
of sensitivity and specificity for LFO, high-flow oxygen (HFO), and mechanical ventilation
(intermittent mandatory ventilation [IMV]).
In El-Hamshari et al's study,[18] ICU admission was recommended if the patient met ICU admission criteria: hemodynamic
instability or an increase in acute respiratory failure demanding an increase in oxygen
therapy from LFO to a nasal cannula, a simple facemask, a nonrebreathe mask (NRB),
or mechanical ventilation. Receiving ICU-level care depends mainly on patients' advanced
requirements for oxygen, such as HFO therapy, NIPPV, or IMV, but some patients were
monitored outside the ICU on medical floors due to a lack of ICU beds. Patients' oxygen
treatment increased, from NRB to HFO therapy to NIPPV. With the patients' or family's
consent, intubation and IMV were considered if none of these interventions raised
oxygenation to > 90% and if the patient continued worsening breathing difficulties.
No reliable prognostic score can predict clinical outcomes in patients with COVID-19,
regardless of whether they are treated in an inpatient or outpatient setting.[19] The present study found that the mean NEWS during the patient's hospital stay was
a good predictor of staying in the hospital for 6 days or more and an excellent predictor
of admission to the ICU. Like worldwide hospital policies, most ICU admissions were
estimated to take 2.0 days from hospital admission. In this study, the mean NEWS during
the patient's hospital stay was an excellent predictor of mortality. These findings
aligned with a few small cohort studies that looked at the predictive value of baseline
NEWS2 and other clinical scoring systems in predicting clinical outcomes in COVID-19
patients based on a single measurement at the time of hospital admission. In a Chinese
study of 654 COVID-19 admissions, the baseline NEWS2 had a better prediction of mortality
than CURB-65 and performed better than qSOFA.[20] In a Korean study of 110 COVID-19 inpatients, using a baseline threshold of NEWS
more than or equal to 5 resulted in a poor predictive value of 0.98 but a positive
predictive value of 0.59 for a future event.[21] The baseline (original) NEWS anticipated an event (defined as ICU admission and
death) for qSOFA. The baseline NEWS2 in a Norwegian study of 66 inpatients predicted
a composite adverse outcome of inpatient death and ICU admission. This was in comparison
with qSOFA's AUROC of 0.62 and CURB-65's AUROC of 0.58.[22] The prognostic qualities of NEWS2 for admission to the ICU were similarly established
in 68 patients with severe COVID-19 in another study.[23] “Serious events” were classified as any of the following occurring while the patient
was hospitalized: death, an unexpected transfer to an ICU, or the start of noninvasive
ventilation. With equivalent AUROC scores for predicting serious events of 0.837 for
admission NEWS and 0.846 for admission, NEWS is modified by age.[24]
According to Covino et al (2020), NEWS was the most reliable indicator of ICU admission
within 48 hours and 7 days of emergency department entrance. The greatest AUROC values
for predicting in-hospital death, according to the findings of Liu et al.[3] However, Knight et al. (2020)[25] found NEWS with in-hospital mortality as an outcome has a good, modest predictive
value for ICU admissions, which would help prioritize a huge attribution number of
patients with COVID-19 who subsequently deteriorate.
The NEWS risk categorization may exist because it provides a quick and easy approach
to spotting deteriorating patients using a united surveillance system that directs
medical personnel to patients according to priority. This opinion is consistent with
Ludikhuize et al[26] who found that NEWS, based on the decline in vital signs noted at least once in
48 hours before an adverse event, might have identified more than 80% of patients
with urgent cases.
The results of this study showed that there were statistically significant differences
between the study post-NEWS group and the control pre-NEWS group for the frequency
of vital sign measurements as the number of vital sign measurements varied, much like
in the control group. In other words, before implementing the NEWS, the staff could
take vital signs twice or once during a shift. They typically only measured two parameters.
However, after implementing NEWS, the mean number of vital signs monitored in the
study group increased noticeably as they measured all vital sign parameters by NEWS
standards. Nevertheless, the fact that there is little or no high-quality data confirming
trends in the myriads of scoring systems established to predict clinical deterioration
should be an important contribution to evidence-based practice.
During the COVID-19 pandemic, the researcher included all patients diagnosed with
COVID-19 and admitted to the hospital, except those who met RCP-recommended exclusion
criteria. Since hospital admission, we have collected comprehensive data on vital
signs, covering every aspect of the NEWS2. Also, the commitment from the medical team
to using the score was high, which indicates that the implementation of the score
was easy and applicable based on a physiological variable and enhanced the flow of
patient monitoring in a cost-effective manner. However, it is important to spell out
some of this study's limitations. First, because the sample was taken from a single
geographic region in Palestine in a relatively short review time and the post-NEWS
data were gathered only 3 months after the score's introduction, the research findings
are less likely to be generalized. Also, this was a double-center study in the same
geographical area, which limits its generalization. Second, not all ward patients
required to be admitted to the ICU based on their NEWS were immediately admitted to
the ICU. Third, the patient's nursing staff was called first, delaying some ICU admissions
when the ICU team was not available or there was a shortage of ICU beds. Fourth, because
patients were not followed up after being discharged from the hospital and the study's
outcomes were restricted to a hospital stay, 28- and 90-day mortality rates were not
provided. Finally, a potential problem with using a control group is the risk bias
caused by other concurrent changes. It could be helpful to improve predictive performance
by supplementing the score parameters with additional measures, such as blood tests.
Instead of creating a new EWS, which would imply a larger sample size and an external
validation cohort, the purpose was to investigate the utility of the existing NEWS2
rating system in the context of COVID-19, as it is already used in medical care. Therefore,
this study is expected to add value to the body of research.
Conclusion
Adoption of NEWS2 was associated with a significant prediction of patient deterioration
and showed a considerable improvement in hospitalized COVID-19 patient outcomes (decreased
length of hospital stay, predicted ICU admissions, and modest decrease in mortality
rate), an increase in the frequency of vital signs measurements, which indirectly
raised the number of medical reviews following the patient's clinical deterioration,
and also that the medical team's commitment to using the score was high. This indicates
that the implementation of the score was easy, applicable, and based on a physiological
variable, and it enhanced the flow of patient monitoring cost-effectively.
