1
The hospital burden of critical illness across global settings:
a point-prevalence and cohort study in Malawi, Sri Lanka and Sweden
Carl Otto Schell
1,2,3
, Raphael Kayambankadzanja
4
, Abigail Beane
5
,Andreas Wellhagen
1,2,6
,
Chamira Kodippily
7
,
Anna Hvarfner
1,8
, Grace Banda-Katha
9,10
, Nalayini Jegathesan
11
,
Christoffer Hintze
12
, Wageesha Wijesiriwardana
7,13
, Martin Gerdin Wärnberg
1,14
,
Mtisunge Kachingwe
15
, Petronella Bjurling-Sjöberg
2,16
, Annie Kalibwe Mkandawire
17
,
Hampus Sjöstedt
2,3
, Surenthirakumaran Rajendra
18
, Cecilia Stålsby Lundborg
1
, Miklos Lipcsey
19,20
,
Lisa Kurland
21
, Rashan Haniffa
5,7,22
, Tim Baker
1,23
1. Department of Global Public Health, Karolinska Institutet, Stockholm, Sweden
2. Centre for Clinical Research Sörmland, Uppsala University, Eskilstuna, Sweden
3. Department of Medicine, Nyköping Hospital, Sörmland Region, Nyköping, Sweden
4. Market Dynamics, Global Health Program, PATH, Lilongwe, Malawi
5. Pandemic Science Hub and Institute for Regeneration and Repair, University of Edinburgh, Edinburgh, Scotland, UK.
6. Department of Anaesthesia & Intensive care, Nyköping Hospital, Sörmland Region, Nyköping, Sweden
7. National Intensive Care Surveillance-MORU, Colombo, Sri Lanka
8. Kusten Primary Health Care Center, Ytterby, Sweden
9. Adult Emergency and Trauma Centre, Queen Elizabeth Central Hospital, Blantyre, Malawi
10. Emergency Medicine Unit, Kamuzu University of Health Sciences, Blantyre, Malawi
11. Teaching Hospital Jaffna, Jaffna, Sri Lanka
12. Department of Otorhinolaryngology, Karolinska University Hospital, Stockholm, Sweden
13. Department of Allied Health Sciences, Faculty of Medicine, University of Colombo, Sri Lanka
14. Function Perioperative Medicine and Intensive Care, Karolinska University Hospital, Stockholm, Sweden
15. Department of Anesthesia and Intensive care, Queen Elizabeth Central Hospital, Blantyre, Malawi
16. Department of Public Health and Caring Sciences, Uppsala University, Uppsala, Sweden
17. Department of Medical Surgical Nursing, Malawi College of Health Sciences, Blantyre, Malawi
18. Department of Community and Family Medicine, Faculty of Medicine, University of Jaffna, Jaffna, Sri Lanka
19. Anesthesiology and Intensive Care, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden
20. Hedenstierna laboratory, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden
21. School of Medical Sciences, Örebro University, Örebro, Sweden
22. University College Hospital London, UK
23. Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania
Key Points
Question: What is the burden of critical illness in hospitals in different global settings, and where are
critically ill patients being cared for?
Findings: Among 3652 hospitalized patients in countries of different socio-economic levels (Malawi,
Sri Lanka, and Sweden) we found a point-prevalence of critical illness of 12.0% (95% CI, 11.0-13.1),
with a hospital mortality of 18.7% (95% CI, 15.3-22.6). The odds ratio of death of critically ill
compared to non-critically ill patients was 7.5 (95% CI, 5.4-10.2). Of the critically ill patients 3.9 %
(95% CI, 2.4-6.1) were cared for in Intensive Care Units (ICUs).
Meaning: Critical illness is common in hospitals and has a high mortality. Ensuring that feasible
critical care interventions are implemented throughout hospitals including in general wards where
more than nine in ten critically ill patients are cared for, has potential to improve outcomes across all
medical specialties.
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
2
Abstract
Importance: Large unmet needs of critical care have been identified globally, but evidence to guide
policy priorities is scarce. Available studies into the burden of critical illness have important
limitations.
Objective: To assess the adult burden of critical illness in hospitals across global settings.
Design, Setting, and Participants: This was a prospective, observational, international, hospital-
based, point-prevalence and cohort study in Malawi, Sri Lanka, and Sweden. On specific days, all
adult in-patients in the eight study hospitals were examined for the presence of critical illness and
followed up for hospital mortality.
Exposure: Patients with one or more severely deranged vital sign were classified as critically ill.
Main Outcomes and Measures: The primary study outcomes were the point-prevalence of critical
illness and 30-day in-hospital mortality. In addition, we assessed the proportion of critically ill
patients who were cared for in Intensive Care Units (ICU)s, and the association between critical illness
and 30-day in-hospital mortality.
Results: Among 3652 hospitalized patients in countries of different socio-economic levels we found a
point-prevalence of critical illness of 12.0% (95% CI, 11.0-13.1), with a hospital mortality of 18.7%
(95% CI, 15.3-22.6). The odds ratio of death of critically ill compared to non-critically ill patients was
7.5 (95% CI, 5.4-10.2). Of the critically ill patients 3.9 % (95% CI, 2.4-6.1) were cared for in ICUs.
Conclusions and Relevance: The study has revealed a substantial burden of critical illness in hospitals
from different global settings. One in eight hospital in-patients were critically ill, 19% of them died in
hospital, and 96% of the critically ill patients were cared for outside ICUs. Implementing feasible, low-
cost, critical care in general wards and units throughout hospitals would impact all critically ill
patients and has potential to improve outcomes across all acute care specialties.
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
3
Introduction
Critical illness is as a ‘state of ill health with vital organ dysfunction, a high risk of imminent death if
care is not provided and the potential for reversibility.
1
Regardless of underlying diagnosis, critically ill
patients require similar initial actions to stabilize vital organ functions and prevent death. Such critical
care interventions are needed wherever a critically ill patient is located.
1,2
Although many effective
critical care interventions are low-cost and feasible throughout hospitals
3,4
there is alarming evidence
from different settings that they are frequently not provided.
5-9
Improving critical care has the
potential to increase survival across medical disciplines.
5,6,10-13
The evidence to guide policy makers about critical illness and critical care is scarce. Decisions around
priorities and investments in health care and research are often grounded in sources based on
patients’ diagnoses where information about critical illness have not been captured.
14,15
Most
research into critical illness outcomes is confined to Intensive Care Units (ICUs), where advanced and
high-cost critical care is provided. ICUs are sparce in rural and low resource settings where care needs
are high.
13,16-18
Per 100,000 population, ICU beds vary – from 0.1 in a Malawi, (low-income country)
and 2.3 in Sri Lanka, (middle-income county) to 5.8 in Sweden and 35 in the USA (high-income
countries).
19-21
It has been estimated that the global incidence of critical illness among adults is 30-45 million per
year, based on extrapolation from a North American ICU registry.
22
This may be an underestimation as
the adult incidence of sepsis alone is 24 million per year.
23
Additionally, there are indications that
critically ill patients may often be cared for outside ICUs.
24-30
To guide policy making, there is a need
for investigations that include critically ill patients throughout hospitals, across ward types,
specialties, diagnoses, and socioeconomic levels. In this multi-centre global study, we aimed to assess
the adult burden of critical illness in hospitals.
Methods
Study design and settings
This was a prospective, observational, international, hospital-based, point-prevalence and cohort
study in Malawi, Sri Lanka, and Sweden. The study countries were chosen to include a low-income,
middle-income and high-income country: the annual health expenditure (USD) per capita ranges from
33 in Malawi, to 151 in Sri Lanka and 6915 in Sweden.
31
The study took place in eight public hospitals
including first-line and referral hospitals in each country (Table 1). Each hospital was assessed at least
twice to control for seasonal variation. The principles of Good Clinical Practice were followed. Ethical
permissions were provided in all settings - Malawi: College of Medicine (P.08/16/2007); Sri Lanka:
University of Kelaniya (P/111/04/2018) and University of Jaffna (J/ERC/19/102/NDR/0205); Sweden:
Ethical Review Board Stockholm (2017-1907-31-1).
Participants and outcomes
In each hospital on the days of point-prevalence assessment, all patients in all somatic wards above
18 years of age were included in the study. Participants who were able to, provided informed
consent. For the validity of the study, we included patients with reduced consciousness in absence of
objection from the patient (verbal or non-verbal) or from the next of kin. We could not include
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
4
patients who were in operating theaters or were absent and could not be found later in the day. The
study excluded women in active labor, patients who were not admitted to hospital (neither had
stayed, or planned to stay overnight), and moribund patients identified as “dying” by the attending
nurse. All participants had their vital signs examined for presence of critical illness, and they were
followed up for hospital mortality, censored at 30 days. We used the term burden of critical illness
when referring to the impact of critical illness - both occurrence (prevalence) and consequence
(mortality). The primary study outcomes were the point-prevalence of critical illness and the 30-day
in-hospital mortality of patients with critical illness. In addition, we assessed the proportion of
critically ill patients who were cared for in ICUs, and the association between critical illness and 30-
day in-hospital mortality.
Table 1. Country and hospital information
Variables
Critical illness was defined as “a state of ill health with vital organ dysfunction, a high risk of imminent
death if care is not provided and a potential for reversibility”
1
and operationalized by classifying a
critically ill patient as someone with one or more severely deranged vital sign at the point prevalence
examination. Such criteria are independent of ward type and specialty and are pragmatic for use in
clinical practice. The a priori decided cutoffs for critical illness are based on triggers for clinical
intervention used at Karolinska University Hospital (Sweden) and in Tanzania (Table 2).
27,30,32,33
The patient’s hospital records were used for clinical information about age, sex, diagnosis, specialty,
and decision to not resuscitate in case of cardiac arrest (DNR). ICU-beds were classified per hospital
definition. All other patients were classified as located in general wards. Some hospitals described
some ward beds as providing a higher care intensity, “high care beds”, or “high dependency unit”
beds. However, the care and interventions available in such locations varied substantially between
settings and precluded a formal analysis.
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
5
Table 2. Parameters and cutoffs for danger signs
Data collection
Data collections took place on individual days between 2017 and 2019 in the study hospitals. All
hospital wards and units were visited, regardless of admitting specialty. Teams of nurses and students
of health care professions went from ward to ward to include all the hospital in-patients and assess
their vital signs. A senior health worker or researcher supervised in each ward to ensure quality data
collection. Prior to this, the data collectors had a day of practice and training on research methods,
ethics, study methods, equipment usage, and standardized vital signs assessments. The equipment
was quality tested before each data collection and included automatic blood-pressure monitors,
pulse oximeters, thermometers, and clocks. Abnormal vital signs were re-checked, and alternative
methods were used if a vital sign could not be assessed (e.g. using manual blood pressure
measurement). The nurse-in-charge of the ward was notified immediately when a patient was
identified as critically ill. The research team offered to document all vital signs collected in the clinical
records for use in patient care.
In Malawi and Sri Lanka, clinical information was extracted from the paper-based patient records and
outcomes were collected through follow-up of records and hospital administrative data. In Sweden,
these data were collected from electronic medical records.
Statistical methods
We used percentages to present point-prevalence, in-hospital mortality, and location of the critically
ill. The association of critical illness and in-hospital mortality was assessed using odds ratios (OR) in
crude logistic regression models and adjusted odds ratios (aOR) in models including prespecified
potential confounders: for age, sex and country. Exact logistic regression was used for analyse at
country level. Missing data for a single vital sign were classified as not being a danger sign (the most
common value), enabling use of the other vital signs to classify the patient’s critical illness status.
Participants who were lost to follow up or had missing data for age, sex or country were excluded
from analyses. A 95% confidence interval was used in reporting findings. Stata/IC 15 (Stata Corp,
College Station, TX) was used for analyses.
Danger Signs
A
Airway sounds
Stridor or Gurgling or Snoring
B
Respiratory rate (per minute)
Oxygen saturation (%)
<8 or >30
<90
C
Heart rate (per minute)
Systolic blood pressure (mmHg)
<40 or >130
<90
D
Conscious level
Glasgow Coma Scale <9
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
6
Results
A total of 3682 participants were initially included. A final cohort of 3652 participants was used for
analyses, after exclusion of twenty patients who were lost to follow-up (18 from Malawi and 2 from
Sri Lanka) and ten patients from Malawi with missing data for age. Out of 21912 expected data points
for vital signs (6 per participant), 72 (0.03%) were missing.
Women comprised 2015 (55%) of all participants and 224 (51%) of the critically ill. The median age of
the cohort was 58 years [IQR34-75] and 61 years [IQR 37-76] among critically ill patients. A majority
of patients were admitted to a medical department: 1846 (51%) of all patients and 327(74%) of the
critically ill.
In the study countries, there were 653 (59%) women in Malawi, 436 (60%) in Sri Lanka, and 926 (51%)
in Sweden. The median age was 35 years [IQR 26-49] in Malawi, 41 years [IQR 30-62] in Sri Lanka and
73 years [IQR61-82] Sweden. There were two patients having a DNR in Malawi (0.2%) none in Sri
Lanka and 346 in Sweden (19%). Clinical characteristics of the cohort are presented in Table 3.
Table 3. Participant characteristics
ALL
MALAWI
SRI LANKA
SWEDEN
All
All
Critical
All
Critical
All
Critical
Participants, n
3652
1107
204
723
43
1822
192
Death in hospital, n (%)
178(4.9%)
85(7.1%)
42(21%)
8(1.1%)
6(14%)
85(4.7%)
34(18%)
Female, n (%)
2015(55%)
653(59%)
119(58%)
436(60%)
17(40%)
926(51%)
88(46%)
Age, years [IQR]
58[34-75]
35[26-49]
38[30-51]
41[30-62]
56[39-68]
73[61-82]
76[69-86]
Specialty, n (% per column)
- Medicine
- OBG
- Surgery
- Unknown
1846(51%)
653(18%)
1151(31%)
2
461(42%)
283(26%)
362(33%)
1
143(70%)
16(7.8%)
45(22%)
-
244(34%)
260(36%)
218(30%)
1
24(56%)
7(16%)
11(26%)
1(2.3%)
1141(62%)
110(6.0%)
571(31%)
-
160(83%)
1(0.5%)
31(16.2%)
-
(DNR), n (%)
348(10%)
2(0,2%)
2(1%)
0
0
346(19%)
79(41%)
Level of care,
n (% per column)
- ICU
- General Ward
52(1.4%)
3600(99%)
5(0.5%)
1102(99%)
3(1.5%)
201(98.5%)
7(1.0%)
716(99%)
3(7.0%)
40(93%)
40(2.2%)
1782(98%)
11(5.7%)
181(94%)
DNR indicates Decision to Not Resuscitate in case of cardiac arrest; ICU Intensive Care Unit; IQR indicates Inter Quartile
Range
Critical illness was present in 439 patients, corresponding to a point-prevalence of critical illness of
12.0% (95% CI 11.0-13.1). The critically ill patients had a hospital mortality of 18.7% (15.5-22.8). Of
the critically ill patients, 17 (3.8% (2.4-6.0), were cared for in an ICU and 422 (96%) in a general ward.
Outcome data are presented in Table 4.
Table 4. Critical illness: point-prevalence, mortality, and proportion in ICU
All
Malawi
Sri Lanka
Sweden
Prevalence Critical Illness
%, (95% CI)
12.0 (11.0-13.1)
18.4(16.3-20.1)
5.9 (4.4-7.9)
10.5 (9.2-12.0)
Mortality Critical Illness
%, (95% CI)
18.7(15.3-22.6)
20.6(15.6-26.7)
14.0(6.3-28.3)
17.8(12.9-23.8)
Proportion of critically ill
patients in ICU %, (95% CI)
3.9 (2.4-6.1)
1.5 (0.5-4.5)
7.0 (2.2-20.1)
5.7 (3.2-10.1)
CI indicates confidence interval; ICU Intensive Care Unit
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
7
In the whole cohort, the association between critical illness and death was OR 7.5 (5.4-10.2). In the
model adjusted for age and sex, aOR was 7.3 (5.3-10.0). In the model adjusted for age, sex and
country, aOR was 6.1 (4.4-8.4) The use of cubic splines to ensure that the association between age
and death was not underestimated did not increase the association in any model and so was not
used. (Table 5)
Table 5. The association between critical illness and 30-day in-hospital mortality
All
Malawi
Sri Lanka
Sweden
Crude models
Critical illness (0R)
7.5 (5.4-10.2)
5.2 (3.3-8.2)
54.0 (9.3-564)
6.7 (4.2-10.6)
Adjusted models
Critical illness (aOR)
Age
Sex (male)
7.3 (5.3-10.0)
1.01 (1.00-1.02)
2.0 (1.4-2.7)
5.3 (3.5-8.5)
1.01 (0.99-1.02)
2.7 (1.7-4.4)
41.1 (7.8-217)
1.02 (0.98-1.06)
2.5 (0.46-13.6)
5.5 (3.4-8.8)
1.04 (1.02-1.06)
1.6 (0.9-2.5)
Critical illness (aOR)
Age (one year)
Sex (male)
Country
- Malawi
- Sweden
- Sri Lanka
6.1 (4.4-8.4)
1.02 (1.01-1.03)
2.1 (1.5-2.8)
6.8 (3.2-14.4)
2.4 (1.1-5.2)
1 (reference)
aOR indicates Adjusted Odds ratio; OR Odds Ratio
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
8
Discussion
In this prospective point prevalence and cohort study of all in-patients in eight hospitals from
Malawi, Sri Lanka and Sweden, we found a substantial burden of critical illness. The point-prevalence
of critical illness was 12% and the critically ill had a hospital mortality of 19%. Critically ill patients had
a significantly higher odds of in-hospital mortality than non-critically (OR 7.5) and of the critically ill,
96% were cared for outside ICUs.
The point-prevalence found is consistent with data that could be extracted from previous single-
centre studies with other aims. Among hospital patients in Finland and Sweden, 8% and 12-14%
respectively had a severely deranged vital sign
25-27
. In medical and surgical wards in Uganda, 12% of
patients had a “critical” modified early waring score of more than 5.
28
Our results support previous
indirect estimations of a substantial global burden of critical illness.
22
The mortality of critically ill patients in our study is high compared to other patient groups and
diagnoses of public interest. For example, patients admitted for care in Swedish ICUs had a hospital
mortality of 14%.
34
Among patients with COVID-19 during the first wave in USA, 20% died in
hospital.
35
Acute myocardial infarction with ST-elevation had 30-day mortalities of 2.4-5.0%, 1.0%
and 0.4% in Sub-Saharan Africa, South Asia and North Europe respectively.
28-30
Our findings confirm
that critical illness, as identified in a pragmatic way using deranged vital signs is a high-risk condition
with very high rates of mortality.
Most critically ill patients are cared for outside ICUs, and this is not limited to low resource settings.
As countries have such large differences in the number of ICU beds – 350 times more beds per
100,000 population in the USA than in Malawi
21
– the presence of critically ill patients in general
wards might be thought to be specific to low-income countries. This is neither supported by our
results nor by previous research.
25-27,29
There are likely explanations behind specific findings in each of the study countries. Sri Lanka, the
middle-income country, had the lowest critical illness prevalence and lowest critical-illness mortality
but also the lowest mortality of non-critically ill patients (0.3%). One explanation for this may be a
higher number of hospital beds (420) per 100,000 population than Sweden (210) and Malawi (130).
This could lead to a lower threshold for admitting patients to hospital in Sri Lanka than the other
countries, thus ”diluting” the proportion of critically ill patients.
31
Finding the highest critical illness
mortality in Malawi (21%), was expected, as resources are far scarcer than in the other countries,
affecting the determinants of health and the resources available for health care.
31
In low-income
countries, critical illness outside hospital may also be common, since limited access may delay or
preclude care.
36
The high mortality of critically ill patients in Sweden of 18% was an interesting
finding, and is likely explained by the high median age of the patients (73 years), above which frailty
and multimorbidity are common.
37
Implications
The findings suggest a need for health systems to recognize and prioritize critical illness throughout
hospitals. This would not be an unachievable goal. In fact the unmet needs of care for most critically
ill patients
5-9
can be provided outside ICUs.
4,11,13,38
The recently defined Essential Emergency and
Critical Care (EECC) includes 40 foundational interventions selected for clinical effectiveness and
feasibility in all hospital settings, such as triage, airway protection and oxygen therapy.
3,12
EECC aligns
with the WHO’s Fair Priorities framework to maximize the population impact of care interventions.
39
Ward-based critical care has lower costs and has been shown to be more cost-effective than ICU care
for many patients groups
40-43
. In Tanzania, EECC has been estimated to be highly cost-effective at 14
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
9
USD per healthy life-year gained.
44
Ensuring EECC is provided to all patients who need it throughout
hospitals and health systems would seem to be the rational first step when improving critical care
services.
11
In cases where the fundamental critical care is not enough to stabilize organ functions,
high dependency units (HDUs) may be a reasonable subsequent step and may be more equitable and
effective than an expansion of resource-intensive ICUs.
45,46
Governments that use a strategy to
improve critical care by starting from the fundamental level could reach all critically ill patients, be
cost-effective, and have impact at population level.
10,11,47
Strengths and limitations
The prospective examination of all in-patients in this study, regardless of diagnosis and location in
hospital, minimized the risk for selection bias and misclassification. The quality of the data collection
increased internal validity through high inclusion rates, few missing data points, and accuracy of the
data. Studying hospitals in a low-, a middle- and a high-income country enabled the inclusion of
patients from settings with a large global variation. The feasible clinical criteria for the identification
of critical illness and the pragmatic data collection methods that were used enable replication in
health facility audits and larger studies.
There are limitations. First, the pragmatic criteria for critical illness may have missed high-risk
patients whose vital signs were insufficiently deranged or who had been stabilized by healthcare
interventions. Conversely, some patients with adapted physiology due to chronic disease may have
been misclassified as critically ill. Second, we could not include patients in operating theatres, some
of whom may have been critically ill. Third, data were collected during working hours and the burden
of critical illness may be altered during weekends and nights.
48
Fourth, the ethical imperative to
inform the ward nurses about the patients who were critically ill may have led to improved care and
reduced critical illness mortality in the cohort. Last, the number of countries and hospitals included
limit generalization to all other settings, but we do not have reasons to think that the burden of
critical illness would be markedly different in other hospitals.
Conclusion
The study has revealed a substantial burden of critical illness in hospitals from different global
settings. One in eight hospital in-patients were critically ill, 19% died in hospital, and 96% of the
critically ill patients were cared for outside ICUs. Implementing feasible, low-cost, critical care in
general wards and units throughout hospitals would impact all critically ill patients and has potential
to improve outcomes across all acute care specialties.
ABBREVIATIONS
aOR Adjusted Odds Ratio
DNR Decision to Not Resuscitate in case of cardiac arrest
EECC Essential Emergency and Critical Care
HDU High Dependency Unit
ICU Intensive Care Unit
IQR Inter Quartile Range
OR Odds Ratio
USA United States of America
USD United States Dollar
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
10
REFERENCES
1. Kayambankadzanja RK, Schell CO, Wärnberg MG, et al. Towards definitions of critical illness
and critical care using concept analysis. BMJ Open. 2022;12(9):e060972.
2. Bartlett ES, Lim A, Kivlehan S, et al. Critical care delivery across health care systems in low-
income and low-middle-income country settings: A systematic review. Journal of global
health. 2023;13:04141.
3. Schell CO, Khalid K, Wharton-Smith A, et al. Essential Emergency and Critical Care: a
consensus among global clinical experts. BMJ Global Health. 2021;6(9):e006585.
4. Guinness L, Kairu A, Kuwawenaruwa A, et al. Essential emergency and critical care as a health
system response to critical illness and the COVID19 pandemic: what does it cost? Cost
Effectiveness and Resource Allocation. 2023;21(1):15.
5. Reynolds TA, Sawe H, Rubiano AM, Shin SD, Wallis L, Mock CN. Strengthening Health Systems
to Provide Emergency Care. In: Jamison DT, Gelband H, Horton S, et al., eds. Disease Control
Priorities: Improving Health and Reducing Poverty. Washington (DC): The International Bank
for Reconstruction and Development / The World Bank
© 2018 International Bank for Reconstruction and Development / The World Bank.; 2017.
6. Losonczy LI, Papali A, Kivlehan S, et al. White Paper on Early Critical Care Services in Low
Resource Settings. Ann Glob Health. 2021;87(1):105.
7. King C, Banda M, Bar-Zeev N, et al. Care-seeking patterns amongst suspected paediatric
pneumonia deaths in rural Malawi. Gates Open Res. 2020;4:178.
8. Hogan H, Healey F, Neale G, Thomson R, Vincent C, Black N. Preventable deaths due to
problems in care in English acute hospitals: a retrospective case record review study. BMJ
Quality &amp;amp; Safety. 2012;21(9):737.
9. Barasa EW, Ouma PO, Okiro EA. Assessing the hospital surge capacity of the Kenyan health
system in the face of the COVID-19 pandemic. PLoS One. 2020;15(7):e0236308.
10. Owoo C, Adhikari NKJ, Akinola O, et al. The World Health Assembly resolution on integrated
emergency, critical, and operative care for universal health coverage and protection from
health emergencies: a golden opportunity to attenuate the global burden of acute and critical
illness. Intensive Care Med. 2023;49(10):1223-1225.
11. Buowari DY, Owoo C, Gupta L, Schell CO, Baker T. Essential Emergency and Critical Care: A
Priority for Health Systems Globally. Crit Care Clin. 2022;38(4):639-656.
12. Schell CO, Gerdin Warnberg M, Hvarfner A, et al. The global need for essential emergency
and critical care. Crit Care. 2018;22(1):284.
13. Crawford AM, Shiferaw AA, Ntambwe P, et al. Global critical care: a call to action. Critical
Care. 2023;27(1):28.
14. Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240
causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study
2013. Lancet. 2014.
15. Wunsch H, Rowan KM, Angus DC. International comparisons in critical care: a necessity and
challenge. Curr Opin Crit Care. 2007;13(6):725-731.
16. Murthy S, Wunsch H. Clinical review: International comparisons in critical care - lessons
learned. Crit Care. 2012;16(2):218.
17. Murthy S, Adhikari NK. Global health care of the critically ill in low-resource settings. Ann Am
Thorac Soc. 2013;10(5):509-513.
18. Dünser MW, Towey RM, Amito J, Mer M. Intensive care medicine in rural sub-Saharan Africa.
Anaesthesia. 2017;72(2):181-189.
19. Sonenthal PD, Kasomekera N, Connolly E, et al. Critical Care Units in Malawi: A Cross-
Sectional Study. Ann Glob Health. 2023;89(1):51.
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
11
20. Phua J, Faruq MO, Kulkarni AP, et al. Critical Care Bed Capacity in Asian Countries and
Regions. Crit Care Med. 2020;48(5):654-662.
21. Ma X, Vervoort D. Critical care capacity during the COVID-19 pandemic: Global availability of
intensive care beds. J Crit Care. 2020;58:96-97.
22. Adhikari NK, Fowler RA, Bhagwanjee S, Rubenfeld GD. Critical care and the global burden of
critical illness in adults. Lancet. 2010;376(9749):1339-1346.
23. Rudd KE, Johnson SC, Agesa KM, et al. Global, regional, and national sepsis incidence and
mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet.
2020;395(10219):200-211.
24. Dart PJ, Kinnear J, Bould MD, Mwansa SL, Rakhda Z, Snell D. An evaluation of inpatient
morbidity and critical care provision in Zambia. Anaesthesia. 2017;72(2):172-180.
25. Tirkkonen J, Olkkola KT, Huhtala H, Tenhunen J, Hoppu S. Medical emergency team activation:
performance of conventional dichotomised criteria versus national early warning score. Acta
Anaesthesiol Scand. 2014;58(4):411-419.
26. Friman O, Bell M, Djarv T, Hvarfner A, Jaderling G. National Early Warning Score vs Rapid
Response Team criteria-Prevalence, misclassification, and outcome. Acta Anaesthesiol Scand.
2019;63(2):215-221.
27. Bell MB, Konrad D, Granath F, Ekbom A, Martling CR. Prevalence and sensitivity of MET-
criteria in a Scandinavian University Hospital. Resuscitation. 2006;70(1):66-73.
28. Kruisselbrink R, Kwizera A, Crowther M, et al. Modified Early Warning Score (MEWS)
Identifies Critical Illness among Ward Patients in a Resource Restricted Setting in Kampala,
Uganda: A Prospective Observational Study. PLoS One. 2016;11(3):e0151408.
29. Prytherch DR, Smith GB, Schmidt PE, Featherstone PI. ViEWS--Towards a national early
warning score for detecting adult inpatient deterioration. Resuscitation. 2010;81(8):932-937.
30. Mboya EA, Ndumwa HP, Amani DE, et al. Critical illness at the emergency department of a
Tanzanian national hospital in a three-year period 2019–2021. BMC Emerg Med.
2023;23(1):86.
31. World Bank. World Bank Open Data. 2024; https://data.worldbank.org/indicator. Accessed 23
January, 2024.
32. Baker T, Blixt J, Lugazia E, et al. Single Deranged Physiologic Parameters Are Associated With
Mortality in a Low-Income Country. Crit Care Med. 2015;43(10):2171-2179.
33. Parker RK, Mwachiro EB, Mwachiro MM, Pletcher J, Parker AS, Many HR. Mortality Prediction
in Rural Kenya: A Cohort Study of Mechanical Ventilation in Critically Ill Patients. Critical care
explorations. 2019;1(12):e0067.
34. Rydenfelt K, Engerström L, Walther S, Sjöberg F, Strömberg U, Samuelsson C. In-hospital vs.
30-day mortality in the critically ill - a 2-year Swedish intensive care cohort analysis. Acta
Anaesthesiol Scand. 2015;59(7):846-858.
35. Rosenthal N, Cao Z, Gundrum J, Sianis J, Safo S. Risk Factors Associated With In-Hospital
Mortality in a US National Sample of Patients With COVID-19. JAMA Network Open.
2020;3(12):e2029058-e2029058.
36. Rutherford ME, Mulholland K, Hill PC. How access to health care relates to under-five
mortality in sub-Saharan Africa: systematic review. Trop Med Int Health. 2010;15(5):508-519.
37. Cunha AIL, Veronese N, de Melo Borges S, Ricci NA. Frailty as a predictor of adverse outcomes
in hospitalized older adults: A systematic review and meta-analysis. Ageing Research Reviews.
2019;56:100960.
38. Werner K, Risko N, Kalanzi J, Wallis LA, Reynolds TA. Cost-effectiveness analysis of the multi-
strategy WHO emergency care toolkit in regional referral hospitals in Uganda. PLoS One.
2022;17(12):e0279074.
39. Ottersen T, Norheim OF. Making fair choices on the path to universal health coverage. Bull
World Health Organ. 2014;92(6):389.
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint
12
40. Cleary SM, Wilkinson T, Tamandjou Tchuem CR, Docrat S, Solanki GC. Cost-effectiveness of
intensive care for hospitalized COVID-19 patients: experience from South Africa. BMC Health
Serv Res. 2021;21(1):82.
41. Kairu A, Were V, Isaaka L, Agweyu A, Aketch S, Barasa EJBgh. Modelling the cost-effectiveness
of essential and advanced critical care for COVID-19 patients in Kenya. 2021;6(12):e007168.
42. Gooch RA, Kahn JM. ICU bed supply, utilization, and health care spending: an example of
demand elasticity. JAMA. 2014;311(6):567-568.
43. Rubenfeld GD. Cost-Effective Critical Care: Cost Containment and Rationing. Semin Respir Crit
Care Med. 2012;33(4):413-420.
44. Shah HA, Baker T, Schell CO, et al. Cost Effectiveness of Strategies for Caring for Critically Ill
Patients with COVID-19 in Tanzania. PharmacoEconomics - Open. 2023;7(4):537-552.
45. Manda-Taylor L, Mndolo S, Baker TJMMJ. Critical care in Malawi: The ethics of beneficence
and justice. 2017;29(3):268-271.
46. Morton B, Banda NP, Nsomba E, et al. Establishment of a high-dependency unit in Malawi.
BMJ Glob Health. 2020;5(11).
47. World Health Organization. Resolution WHA EB152(3): Integrated emergency, critical and
operative care for universal health coverage and protection from health em. In: World Health
Organization, ed2023.
48. Alabas OA, Jernberg T, Pujades-Rodriguez M, et al. Statistics on mortality following acute
myocardial infarction in 842 897 Europeans. Cardiovasc Res. 2019.
. CC-BY 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted March 15, 2024. ; https://doi.org/10.1101/2024.03.14.24304275doi: medRxiv preprint