IMPACT OF HEALTH INFORMATICS TECHNOLOGY ON THE IMPLEMENTATION OF A MODIFIED EARLY WARNING SCORE (MEWS)

Health Informatics - An International Journal (HIIJ) Vol.4, No.3/4, November 2015
DOI: 10.5121/hiij.2015.4403 27
IMPACT OF HEALTH INFORMATICS TECHNOLOGY
ON THE IMPLEMENTATION OF A MODIFIED EARLY
WARNING SCORE (MEWS)
Sajeesh Kumar
Department of Health Informatics & Information Management, University of Tennessee
Health Science Center, Memphis, TN, USA
ABSTRACT
The Modified Early Warning System (MEWS) is based on a patient score that helps the medical team
monitor patients to identify a patient that may be experiencing a sudden decline in care. This study consists
of a detailed review of clinical data and patient outcomes to assess impact of technology and patient care.
There are a total of thirteen hospitals included in this review. These facilities have implemented vitals
capture and the MEWS scoring system.
KEYWORDS
Modified Early Warning System, electronic health record, Meaningful Use
1. INTRODUCTION
The adoption of Electronic Health Records has increased in recent years. As a result discrete data
elements are more widely available which can be leveraged for secondary use. Clinical decision
support and alert notification to clinicians are examples of such uses. The implementation of early
warning systems can assist in identifying patients that may be at risk for decomposition or decline
in clinical condition. Specifically, clarifications of availability in near real-time to facilitate the
calculation of the Modified Early Warning Score (MEWS).
The Modified Early Warning System is based on a patient score that helps the medical team
monitor patients to identify a patient that may be experiencing a sudden decline in care. Based on
literature review the data elements and threshold vary slightly be essentially include the
following: Respiratory Rate, Heart Rate, Systolic Blood Pressure, Consciousness/Alertness Level,
Temperature and Urine Output.
The use of technology to aid in patient care and nursing workflow has evolved significantly over
the last several years. In addition to the implementation of electronic health records,
organizations are implementing electronic vital capture devices (capsules). This equipment is
used to electronically transmit data from a device to the electronic health record (EHR). The
patients’ vitals are used in real-time to facilitate the calculation of the MEWS scores. The goal of
this research is to evaluate the use of the medical devices in facility and assess the availability of
data elements to calculate the MEWS score. This current work seeks to further explore the impact
vitals capture and MEWS implementation on patient outcomes such as cardiac arrest, length of
stay (LOS) and mortality.

28
This study consists of a detailed review of clinical data and patient outcomes to assess impact of
technology and patient care. There are a total of thirteen hospitals included in this review. These
facilities have implemented vitals capture and the MEWS scoring system. This population may be
biased if they are considered early adopters of technology. Any conclusions would need to be
interpreted with understanding of this study limitation.
2. LITERATURE
This literature search was conducted using electronic databases MEDLINE, PubMed, and
Up To Date. The search terms included but not limited to: modified warning systems,
modified early warning system, MEWS, MEWS Score, pediatric warning systems, PEWS,
PEWS Score, PEWS and Sepsis. The advanced search filters were also used to limit those
articles published from the year 1990 through the current year 2014. These techniques
resulted in almost 800 possibly relevant articles. After manual review, 66 articles appeared
relevant to thesis topic. A more detailed review of each abstract was conducted this resulted
in 5 articles that provided additional background and 6 articles with original research.
FIGURE 1: Literature Search Flow Diagram
Modified Early Warning Systems (MEWS) attempt to identify deteriorating patients early so timely
interventions can occur thus reducing serious adverse events (Hammond et al, 2012).
The studies were split based on patient age. One group consisted of two articles focused on the
pediatric population and the Pediatric Early Warning System (PEWS) score. The other group
consisted of three articles which were focused on the adult patient population and the MEWS score.
The populations studied were in the following countries: United States, United Kingdom and
Australia.
Survey Methods and Variables
The study methods varied across all articles that were reviewed. However, they all included a
retrospective review. It is important to indicate that the scoring tools in both of the pediatric studies
were generally similar. The tool from the Duncan 2012, article had a few additional evaluation
criteria in each of the areas. Additionally, this tool included a color coding system.

29
In the Sensitivity of the Pediatric Early Warning Score article, the researchers highlighted two prior
studies that addressed the Pediatric Early Warning Score (PEWS) to unplanned transfers to a higher
level of care. Additionally, the authors identified that this study “added” to the collective research.
The team had a hypothesis that at least 80% of patients had a critical PEWS score preceding the
event (code call/rapid response team notification). The study was conducted through a retrospective
chart review of 186 cases/unique patients that had a rapid response team and/or code blue event.
Based on the records identified, a nurse that specialized in critical care determined the PEWS for
each patient at various intervals in the 24 hours prior to the event. Additional data was also collected
to further aid the research. The study concluded that sensitivity of the PEWS was 85.5%, defined as
the patient having a critical score within 24 prior to the event. The team determined that the median
time for the first critical PEWS event was 696 minutes (11 hours and 36 minutes). In the other
article that addressed the pediatric population the authors focused on the experience of Stony Brook
University Medical Center (SBUMC) and their journey with implementing an early warning
system. The center started with the implementation of the Pediatric Early Warning Score (PEWS) in
2007. They provided detailed descriptions that presented an overview of the workflow and
processes that should be taken based on the patient’s score. They implemented a status board at
each nursing station providing a quick view of the unit’s general acuity level. The program has been

30
expanded to the adult population and SBUMC has implemented the modified early warning system
(MEWS). These tools continued to be refined and have been incorporated into the electronic health
record. However, the facility has identified limitations with bedside documentation. As a result,
there could be delays in the scoring and alerting of the patient’s condition. Even with some of the
limitations, the utilization of the warning system is another tool that staff can use to help patients
that may be at risk for a decompensating condition (Duncan, et al 2012).
The other two articles focused on the adult population. In the third article by Gardner-Thorpe, J. et
al (2006), the study included 334 elective and emergent surgical patients that were selected over a
five month period. The team conducted prospective data collection on various elements (age,
gender, American Society of Anesthesiologists (ASA) Physical Status Classification System
surgical operation, length of stay) as the MEWS was calculated for each patient daily. In the data
analysis the team selected a threshold of a MEWS greater than four or more. Of these patients, 75%
had triggered the early warning system and went to critical care. They also completed detailed
sensitivity and specificity testing to demonstrate how the MEWS could be leveraged as a predictor
of admission to a critical care unit. In another, there was a review of the frequency of vital signs
before and after the implementation of MEWS charting and staff education. The study evaluated 69
patients pre-implementation and 70 patients post-implementation. For the purpose of the study a full
set of patient vitals included blood pressure, heart rate, temperature, oxygen saturation, respiratory
rate and urine output. This study has very specific findings that indicated a 210% increase in overall
frequency of vital sign documentation. This study was limited to the ICU population and was done
within three months of the implementation of the MEWS (Hammond et al, 2012).
Modified Early Warning Score (MEWS) transforms this physiological data, and using weighted
criteria, determines an aggregate score that may be used to categorize clinical deterioration and
direct care. The score can be determined electronically or manually based on the processes of the
individual facilities. The table below illustrates the scoring system. In 2001, Subbe et al. validated
a Modified Early Warning Score (MEWS) that demonstrated a significant relationship between
the score and transfers to the ICU or death. Additionally the study indicates that a clinical
pathway or alerting system could be implemented to have clinical staff respond to those patients
with high scores to respond with medical intervention. In this study, there is a bedside evaluation
of the patient and calculation of the MEWS score with a screening tool to assist in the
identification of patients that might be at risk. The score is comprised of five measures:
1. Blood pressure
2. Heart Rate
3. Temperature
4. Level of Consciousness or (AVPU)
5. Respiratory Rate
Table 2: MEWS Scoring System
Components 3 2 1 0 1 2 3
Systolic Blood
Pressure (mmHg)
<70 71-
80
81-
100
101-
199
≥200
Pulse Rate (HR)
(bpm)
<40 41-
50
51-100 101-110 111-129 ≥130
Respiratory rate
(bpm)
<9 9-14 15-20 21-29 ≥30
Temperature (C) <35 35-38.4 ≥38.5
AVPU score Alert Reacting Reacting Unresponsive

31
to Voice to Pain
This study concluded that the MEWS was a simple bedside tool that was administered by nursing
staff. This tool helped to identify those patients that were at risk of deterioration and needed more
acute clinical intervention.
All of the above studies independently concluded value in the implementation of MEWS and/or
PEWS score. In general they found that there is a strong indicator of its potential to alert the care
providers of the patient’s clinical deterioration. Even with some of the limitations of the various
studies, the researchers determined that the utilization of the warning system is another tool that
staff can use to help patients that may be at risk for a decompensating condition. The authors
concluded that the MEWS is an important part of a risk management strategy that is simple to
implement, (Gardner-Thorpe J. et al 2006).
3. METHOD
This current study includes thirteen acute care facilities that have implemented the Vital Now
technology and actively implemented the MEWS scoring system. This is a review patient level
summary data and includes the evaluation of the availability of the key data elements to calculate
the MEWS score. Additionally, the study also includes a review of a review of key outcome
indicators. This includes the number of cardiac arrests, length of stay and mortality rate.
This data analysis is a retrospective data review from data collected in the calendar year 2013. The
data has been extracted from the Enterprise Data Warehouse and specifically the clinical domain of
vitals and internal outcomes reporting system. The facilities to be included in the evaluation would
include the following, implementation of MEWS score, implementation of vitals capture and on the
latest version of the HIS system. This would ensure that all required data elements had the potential
to be present.
A select group of thirteen facilities have been selected for study. This will be referred to as the study
cohort. Given the confidential and proprietary nature of the information, all patient identifiers have
been removed and the facilities have been blinded for external review purposes.
4. RESULTS
The summary of the study results are described in the section below. Frist is the analysis of all
facilities and their vital capture rates. The vitals capture metric includes several elements. The
electronic vitals are defined as the total number of vitals obtained electronically and submitted
through the neuron to the hospital information system (HIS). The manual vitals are defined as
the total number of vitals entered manually into the HIS. The electronic vitals capture rate is a
percentage as follows, (electronic/(manual + electronic))x 100.
Table 3: Vitals Capture

32
Facility Electronic Manual Percentage Electronic Manual Percentage Electronic Manual Percentage
A 151689 16065 90 40925 2792 94 11074 624 95
B 160886 56412 74 78239 28655 73 6737 2000 77
C 63963 31466 67 34931 17215 67 2976 1258 70
D 143938 20666 87 84304 10949 89 6814 761 90
E 123497 27834 82 64964 15175 81 4990 1181 81
F 123289 27317 82 68923 11807 85 4985 550 90
G 91692 18365 83 48256 8348 85 3520 756 82
H 76286 12820 86 38200 6581 85 3335 587 85
I 44829 9631 82 26211 5485 83 2240 332 87
J 76442 8881 90 41530 4996 89 3655 387 90
K 77434 10252 88 40617 5818 87 2739 362 88
L 49362 3187 94 24697 1617 94 2240 106 95
M 113355 22831 83 62625 9545 87 5041 485 91
Last Month6 Months 3 Months

33
In reviewing the data, Facility C has the lowest electronic capture rate with just 70% in the last
month. Facilities A and L have the highest electronic capture rate. In the last month they both
achieved 95%. Upon review of the trend over the last six month, seven of the thirteen facilities
improved their performance of the capture of vitals through the neuron to the HIS. This is a
positive trend and would allow timely calculation of the MEWS score.
The next analysis is on the MEWS score calculation. The MEWS score metric includes
several elements. The MEWS vitals are defined as the total number of vitals obtained
electronically where the five elements are all present. This includes systolic blood pressure,
heart rate, respiratory rate, body temperature and level of consciousness as measure by
AVPU. Any vitals that are entered manually into the HIS are excluded from the numerator.
The all vitals calculation is defined as the total number of vitals including both manual and
electronic capture.
Table 4: MEWS Score Calculation
Facility MEWS Vitals All Vitals Percentage MEWS Vitals All Vitals Percentage MEWS Vitals All Vitals Percentage
A 80930 105092 77 29782 37937 79 29782 37937 79
B 77494 117448 66 33932 51868 65 15207 23729 64
C 31145 48555 64 13511 21945 62 5870 9705 60
D 90074 118001 76 43510 56989 76 18908 24632 77
E 76568 91855 83 35541 42765 83 14537 17514 83
F 69377 87191 80 29306 38882 75 12375 16171 77
G 46784 75284 62 22846 34005 67 10819 14755 73
H 37008 53298 69 14914 21823 68 5865 8960 65
I 31854 37818 84 15059 18010 84 6429 7901 81
J 51996 63390 82 23944 29141 82 10124 12350 82
K 38725 55141 70 15232 22318 68 6504 9777 67
L 32850 38043 86 15271 17441 88 6753 7845 86
M 71597 80438 89 32499 36552 89 13527 15121 89
6 Months 3 Months Last Month
FIGURE 3: MEWS Score Calculation Diagram
A B C D E F G H I J K L M
6 month 77 66 64 76 83 80 62 69 84 82 70 86 89
3 month 79 65 62 76 83 75 67 68 84 82 68 88 89
Past Month 79 64 60 77 83 77 73 65 81 82 67 86 89
0
10
20
30
40
50
60
70
80
90
100
Percentage
MEWS Scores

34
In reviewing the data, Facility C has the lowest MEWS Score calculation with just 60% in the
last month. Facility A that had one of the highest electronic vital capture rate is only at 79% of
the MEWS Score calculation. Facility M has the highest calculation with 89% over the last six
months. It appears that more detailed analysis of the data would be important to understand
which elements of MEWS score are not being captured electronically. This would be important
to help improve the MEWS score calculation.
The next analysis of data is focused on outcomes in several categories. The first of which is looking
at the incidence of cardiac arrest in the facility as defined by ICD-9-CM diagnosis code in either the
primary or other diagnosis categories. This metric was defined as the total number of cardiac arrests
divided by the total number of discharges x 100 to calculate a percentage.
FIGURE 4: Percentage of Cardiac Arrests Diagram
6 months 0.43 0.50 0.36 0.36 0.29 0.37 0.20 0.16 0.19 0.39 0.35 0.50 0.45
3 month 0.41 0.54 0.32 0.35 0.32 0.39 0.17 0.22 0.18 0.39 0.40 0.55 0.45
Past Month 0.47 0.34 0.39 0.30 0.38 0.47 0.23 0.05 0.25 0.38 0.20 0.53 0.37
0.00%
0.10%
0.20%
0.30%
0.40%
0.50%
0.60%
Percentage
% of Cardiac Arrest
Upon review of the data all of the facilities had less than a 1% cardiac arrest rate. The percentages
ranged from .05% to .55%. Facility H has the lowest overall cardiac arrest rate. In general their
vitals capture was about 85% and their MEWS Score Calculation was 65-69%.
In recent literature researcher has identified that there is the potential to leverage the MEWS score
an early indication of patients that may be at risk for Sepsis. However, there does not appear any
definitive correlation. Zavatti, L. et al (2010) concluded that there was a low sensitivity in the
identification of patients with severe septic shock. However, the study was limited to the review of
51 patients in the study. The next data analysis was focused on patient outcomes (LOS and
Mortality) for sepsis patients in the thirteen facilities. For both the length of stay and mortality there
is a year over year comparison with 2013 as the current year and 2012 as the prior year. The first
metric is length of stay. This is defined as the number of days during the patients stay.

35
FIGURE 5: Sepsis Length of Stay Diagram
Sepsis - LOS CY - 13 12.410.78.7614.213.28.2610.36.119.277.968.086.4311.3
Sepsis - LOS PY - 12 9.729.798.8218.613.39.358.6911.18.8411.211.9 6.4 10.6
0
2
4
6
8
10
12
14
16
18
20Days
Sepsis Length of Stay
As noted in Figure 5, Sepsis Length of Stay Diagram, Facility D has the highest LOS in 2012 with
18.6. Facility D also had the highest LOS in 2013 with 14.2 days. However, this is a positive change
of 4.4 days. Facility H had the lowest LOS in 2013 with just 6.11 days. This is a reduction of 5 days
from the prior year. This is the same facility that had the lowest number of cardiac arrests.
The last analysis was the mortality rate. This metric is defined as the total number of patients in the
reporting period and the total number of deaths. This analysis is also limited to just those patients
with a diagnosis of sepsis.
FIGURE 6: Sepsis Mortality Rate Diagram
As noted in Figure 6, Sepsis Mortality Rate Diagram, 10 of the 13 Facility had a decrease
in the Sepsis Mortality Rate from prior year. However, Facility C has the highest mortality

36
rate in 2013 with .29%. Facility L has the lowest mortality rate with .04%. Facility D which
had the highest LOS in 2013 with 14.2 days has the second lowest mortality rate with
.06%.
6. CONCLUSIONS
In general there is data available electronically that can be leveraged for secondary use. Moreover,
this data can be leveraged in near real-time to impact the care that is delivered to patients. The
facilities in this study appear to have adopted the collection of vitals in an electronic method and
compliance ranged from 95% - 67% in the last six months. Additionally, the MEWS calculation
when all five elements were present also appeared in 89% - 60% of the cases. This provides an
opportunity for improvement if the MEWS score were to be leveraged for alerting and monitoring.
There may be an opportunity for more detailed analysis on each of the five components to
determine if there is an educational opportunity or barrier to collecting all of the data required for
the MEWS score calculation.
For the outcome measures there appeared to be a strong relationship in one facility where the low
cardiac arrest rate also resulted in a low LOS for sepsis patients. There was also a notable finding in
the facility with a higher length of stay and a lower mortality rate for 2013.
For future research, a recommendation would be to evaluate the patient outcomes prior to the
implementation of MEWS at the facility. This might provide a more accurate impact analysis of
MEWS and outcomes. Additionally study would be warranted to determine if MEWS could be an
early predictor of sepsis.
ACKNOWLEDGEMENTS
The authors would like to thank the faculty at Department of Health Informatics and Information
Management, University of Tennessee.
REFERENCES
[1] Akre, M., Finkelstein, M., Erickson, M., Liu, M., Vanderbilt, L., & Billman, G. (2010). Sensitivity of
the Pediatric Early Warning Score to identify patient deterioration. Pediatrics, 125(4), e763-e769.
[2] Duncan, K. D., McMullan, C., & Mills, B. M. (2012). Early warning systems: The next level of rapid
response. Nursing 2012, 42(2), 38-44.
[3] Gardner-Thorpe J., Love N., Wrightson J., Walsh S., Keeling N. The value of Modified Early
Warning Score (MEWS) in surgical outpatients: a prospective observational study. Annals the Royal
College of Surgeons of England. 2006;88(6):571-575.
[4] Gosfield, A. G., & Reinertsen, J. L. (2005). The 100,000 lives campaign: crystallizing standards of
care for Hospitals. Health Affairs, 24(6), 1560-1570.
[5] Hammond, N., Spooner, A., Barnett, A., Corley, A., Brown, P., Fraser, J. (2013). The effect of
implementing a modified early warning scoring (MEWS) system on the adequacy of vital sign
documentation. Australian Critical Care, 26 (2013) 18-22.
[6] Montagnolo, A. J. (2012). The Future of Connectivity. Trustee: The Journal for Hospital Governing
Boards, 65(11), 25-26.
[7] Murphy K., Packer C., Stevens A., Simpson S., Hailey D., Robert G., Chan, L. (2007). Effective early
warning systems for new and emerging health technologies: Developing an evaluation framework and
an assessment of current systems. International Journal of Technology Assessment in Health Care,
23(3), 324-330.
[8] Subbe, C.P., Kruger, M., Rutherford, P. and Gemmel, L. (2001). Validation of an Early Warning
Score in Medical Admissions. QJM 94:521-526

37
[9] Winters, B. D., Pham, J. C., Hunt, E. A., Guallar, E., Berenholtz, S., & Pronovost, P. J. (2007). Rapid
response systems: A systematic review. Critical Care Medicine, 35(5), 1238-1243.
[10] Zavatti, L., Barbieri, E., Amateis, E., & Girardis, M. Rapid response Modified Early Warning Score
and identification of patients with severe sepsis. Critical Care. 2010; 14(Suppl 1)254.
AUTHOR
Sajeesh Kumar, PhD., is Executive Director- Institute for Health Outcomes and Policy ,
Chair-Health Outcomes and Policy Research PhD program and Associate Professor,
Department of Health Informatics & Information Management, University of
Tennessee-Memphis. Dr. Kumar’s research focuses on design and development of
public health informatics, telemedicine technology, health education technologies and
health informatics technology evaluation. Dr. Kumar holds an impressive record of
competitive research funding and research awards.

IMPACT OF HEALTH INFORMATICS TECHNOLOGY ON THE IMPLEMENTATION OF A MODIFIED EARLY WARNING SCORE (MEWS)

More Related Content

What's hot (15)

Viewers also liked (14)

Similar to IMPACT OF HEALTH INFORMATICS TECHNOLOGY ON THE IMPLEMENTATION OF A MODIFIED EARLY WARNING SCORE (MEWS) (20)

More from hiij (20)

Recently uploaded (20)

IMPACT OF HEALTH INFORMATICS TECHNOLOGY ON THE IMPLEMENTATION OF A MODIFIED EARLY WARNING SCORE (MEWS)