REVIEW 10.1111/j.1469-0691.2012.03956.

Bed occupancy rates and hospital-acquired infections—should beds be

kept empty?

K. Kaier1, N. T. Mutters2 and U. Frank3

1) Department of Environmental Health Sciences, University Medical Centre Freiburg, Freiburg, 2) Department of Infectious Diseases, Medical Microbiology
and Hygiene, Heidelberg University Hospital and 3) Division of Infection Control and Hospital Epidemiology, Department of Infectious Diseases, Heidelberg
University Hospital, Heidelberg, Germany

Abstract

There is growing evidence that bed occupancy (BO) rates, overcrowding and understaffing influence the spread of hospital-acquired
infections (HAIs). In this article, a systematic review of the literature is presented, summarizing the evidence on the adverse effects of
high BO rates and overcrowding in hospitals on the incidence of HAIs. A Pubmed database search identified 179 references, of which
44 were considered to be potentially relevant for full-text review. The majority (62.9%) focused on methicillin-resistant Staphylococcus
aureus-associated infection or colonization. Only 12 studies were found that provided a statistical analysis of the impact of BO on HAI
rates. The median BO rate of the analysed studies was 81.2%. The majority of studies (75%) indicated that BO rates and understaffing
directly influence the incidence of HAIs. Only three studies showed no significant association between BO rates and the incidence of
HAIs. Interestingly, only one of the included studies detected a seasonal trend in the BO rate. The present review shows an association
between BO rates and the spread of HAIs in various settings. Because the evidence on this topic is limited, we conclude that further
research is needed in order to analyse the rationale of a threshold BO rate, because keeping beds empty is comparatively costly.

Keywords: Bed occupancy rates, hospital-acquired infections, infectious disease epidemiology, methicillin-resistant Staphylococcus aureus,
threshold
Article published online: 18 June 2012
Clin Microbiol Infect 2012; 18: 941–945

Corresponding author: K. Kaier, Department of Environmental

Health Sciences, University Medical Centre Freiburg, Breisacher
Straße 115b, D-79106 Freiburg, Germany
E-mail: klaus.kaier@uniklinik-freiburg.de

HAIs are of great concern in European society, and may

Introduction
Hospital-acquired infections (HAIs) are the most frequent
adverse events in healthcare delivery [1]. Because of growing
awareness in recent years, the burden caused by these infec-
tions, which is often complicated by antimicrobial resistance,
has become a top priority within the European public health
agenda [2,3]. It has been estimated that, in the EU alone,
approximately 37 000 lives are lost because of HAIs each
year, with an associated monetary cost of roughly €7 billion,
which is mainly attributable to increased length of hospital
stay [1].
be considered from different viewpoints—those of the
clinician, the healthcare official, a pharmaceutical company,
the patient, and the public authority [4]. From a public
health perspective, the major focus is on hospital-wide cor-
relations between incidence rates of hospital-specific patho-
gens and a basket of hospital-specific parameters [5].
Transmission of HAIs is influenced by many different fac-
tors: several individual risk factors, such as exposure to
invasive devices [6], antimicrobial use, intensity and fre-
quency of patient contact with others or with the environ-
ment, and implementation of hygiene measures; and
organizational and institutional factors. There is growing evi-

ª2012 The Authors

Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases
942 Clinical Microbiology and Infection, Volume 18 Number 10, October 2012
dence that bed occupancy (BO) rates, overcrowding and
understaffing do influence many of these transmission-asso-
ciated factors. Unfortunately, relatively little is known about
the effects of these organizational and institutional factors
on the risk of HAIs [6]. Studies have shown that over-
crowding and understaffing lead to failure of patient safety
programmes via decreased healthcare worker hand hygiene
compliance, increased movement of patients and staff
between hospital wards, decreased levels of cohorting, and
overburdening of screening and isolation facilities [7]. Even
a BO threshold for patient safety has been discussed [8];
however, there are still insufficient data to support this.
The present article summarizes the evidence on the influ-
ence of BO rates and overcrowding on HAIs.
Methods
Studies were retrieved from PubMed (http://www.ncbi.nlm.
nih.gov/pubmed/; accessed on 6 February 2012). Search
terms included: (‘Clostridium difficile’ OR ‘VRE’ OR ‘vanco-
mycin resistant enterococci’ OR ‘ESBL’ OR ‘extended spec-
trum beta lactamase’ OR ‘MRSA’ OR ‘methicillin-resistant
Staphylococcus aureus’ OR ‘BSI’ OR ‘blood stream infections’
OR ‘central line-associated blood stream infections’ OR
‘CLA-BSI’ OR ‘VAP’ OR ‘ventilator acquired pneumonia’ OR
‘UTI’ OR ‘urinary tract infections’ OR ‘catheter-associated
urinary tract infections’ OR ‘CA-UTI’ OR ‘SSI’ OR ‘surgical
site infections’ OR ((‘nosocomial’ OR ‘hospital-associated’
OR ‘hospital-acquired’) OR ‘hospital associated’ OR ‘hospital
Titles and abstracts identified and screened
n = 179
Full copies retrieved and screened for inclusion
n = 43
Total numbers of studies included
n = 12
From initial query n = 11
From reference list n=1
FIG. 1. Flow chart of the review process.
ª2012 The Authors
Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 941–945
CMI
acquired’) AND (‘infection’ OR ‘infections’))) AND (‘bed
occupancy’ OR ‘overcrowding’ OR ‘bed utilization’ OR
‘workload’). References were also identified from the biblio-
graphies of studies. The search was performed to identify
studies that analysed the impact of BO on the spread of
HAIs. The search was restricted to studies published
between 1998 and February 2012. Language of publication
was not restricted. The abstracts of all studies were
reviewed. Two authors (K.K. and N.T.M.) independently
assessed the inclusion (presence of a statistical analysis of
association) and exclusion (reviews, comments, etc.) criteria.
K.K. and N.T.M. abstracted data systematically and indepen-
dently according to design, methods, sample size, type of
infectious agent, and setting.
Results
Our literature search identified 179 references, of which 43
were considered to be potentially relevant for full-text
review. An additional article was retrieved from the refer-
ence list in the literature [9]. Of the 44 full copies screened,
35 investigated the infectious agents associated with HAIs.
The majority (62.9%) focused on methicillin-resistant S. aur-
eus (MRSA)-associated infection or colonization. Only two
focused on viral diseases (i.e. rotavirus and norovirus), and
one on a parasitic disease (scabies).
Finally, 12 studies were identified as being valid for analy-
sis. The flow of studies through the review process and the
number of studies excluded are shown in Fig. 1.
Not relevant n = 136
Excluded n = 32
No direct statistical
Association n = 26
Reviews n=5
Comments n=1
CMI Kaier et al.
The findings from the 12 articles met the inclusion criteria
and were considered to be relevant (Table 1). Of these,
66.7% (8/12) investigated MRSA-associated infection or colo-
nization [9–16], one studied MRSA and extended-spectrum
b-lactamase-producing Gram-negatives [5], one studied C. dif-
ficile [11], one studied Enterobacter cloacae [12], and one
studied different infection types caused by varying agents
(pneumonia and bloodstream infections) [17]. The median
BO rate was 81.2%.
The majority of the studies (75%) indicated that over-
crowding of hospitals directly influences the incidence of
HAIs. Six used monthly time-series data, and found a positive
correlation for the incidence of HAIs and BO rates [5,10–
12,16,18]. Of these, three used a multivariate adjusted auto-
regressive integrated moving average model [5,12,18], and
three determined simple univariate correlation coefficients
[10,11,13]. However, one study showed limited validity, as it
was not controlled for the number of patient bed-days [10].
Two studies investigated annual cross-sectional data while
applying a statistical comparison of BO and MRSA rates of
different hospitals [9,13]. The first of these studies analysed
two time periods, but found a positive correlation for the
second period only [13]. The second study, however, applied
the very same methodology to another set of hospitals, and
found a positive correlation for the entire sample of data [9].
One study, a retrospective cohort study, found that the risk
ratio for E. cloacae infection was markedly higher in times of
overcrowding and understaffing [19].
In contrast to these studies, some (25%) found divergent
results. One study, using a multivariate logistic regression
including 61 MRSA patients (colonization and infections),
found no evidence that intensive-care unit staffing or work-
load predicts the risk of MRSA acquisition [14]. Sakamoto
et al. [15] used monthly time-series data to determine simple
univariate correlation coefficients, and showed that BO rates
(r = 0.033, p 0.199), patient/nurse ratio (r = )0.769, p 0.295)
and colonization pressure (r = 0.006, p 0.571) were not sig-
nificantly associated with MRSA incidence density rates. The
only factor that was significantly associated was the amount
of alcohol-based hand sanitizer used for one patient per day
(r = )0.769, p 0.011). The third descriptive epidemiological
study analysed a total of 1312 cases of pneumonia and 513
cases of bloodstream infection . The major finding was that
fewer nosocomial infections were associated with a high BO
rate and a high nurse/ventilated patient ratio [17]. However,
the turnover interval was not checked, which might have
impaired the results. As, by definition, no case of nosocomial
infection would appear if a patient stayed on a ward for
<24 h, a short turnover interval could lead to an underesti-
mation of incidence rates.
Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 941–945
Bed occupancy rates and hospital-acquired infections 943
Interestingly, only one study detected a seasonal trend in
the BO rates [11]. Although a seasonal trend can be
excluded for some studies [5,12,18], for the majority of all
studies (66%) no information on seasonality is available.
Discussion
BO rates have been proposed as a measure that reflects the
ability of a hospital to properly care for patients [20]. How-
ever, whether this measure can be considered useful in guid-
ing the planning and operational management of hospital
beds in order to guarantee appropriate compliance with
patient safety measures depends on the answers to three
questions: (i) to what extent do BO rates influence patient
outcomes; (ii) what would be a target BO rate to aspire to;
and (iii) would any attempt to reduce the BO rate to the
target level be cost-effective?
Patient safety measures such as hand hygiene may take
considerable time. Overworked staff may feel that they do
not have enough time to undertake these measures without
compromising patient care. In addition, understaffing is most
likely to happen when BO rates are high [6,9]. Accordingly,
high BO rates directly impact on the incidence and spread of
hospital-acquired infections, as shown in the present review
of the literature. Only two studies showed no significant
association between BO rates and the incidence of HAIs,
and one found a negative correlation. In summary, the exist-
ing studies support the conclusion that BO rates have an
important impact on patient outcomes. It is not clear to
what extent BO rates have a direct impact on HAIs or can
be considered as proxy indicators for deficiencies in hygiene
measures resulting from overcrowding and understaffing.
The implementation and the true extent of the use of patient
safety measures such as hand disinfection and barrier precau-
tions is interesting with regard to future studies. The main
research in this field has been focused on mainly one patho-
gen only (MRSA). Little attention has been drawn to infec-
tions caused by Gram-negatives or viruses, or vancomycin-
resistant enterococci. As transmission pathways differ
between pathogens, it is interesting to investigate whether
BO rates influence the incidence rates of differing pathogens
to varying extents.
As mentioned before, one study showed that simple fluc-
tuations in the BO rate did not have a direct impact on the
incidence of MRSA infection as long as the BO rate was
within designated levels. Instead, episodes of significant over-
crowding, with occupancy levels in excess of designated
numbers, triggered increases in infection incidence rates
[10]. The BO rate analysed was, however, substantially
ª2012 The Authors

TABLE 1. Characteristics of included studies
Reference Setting
ª2012 The Authors
[10] 900-bed hospital, all units
included (general wards only,
no ICUs)
[11] 900-bed hospital, 8 general
medicine wards and 8 surgical
wards, each with 30 beds
[12] 1500-bed university hospital
(including 120 ICU beds)
[13] 12 hospitals
[9] 38–40 hospitals
[19] 15-bed NICU, 60 patients
[5] 1500-bed university hospital
(including 120 ICU beds)
[18] 1500-bed university hospital
(including 120 ICU beds)
[14] 1 ICU, 61 patients
[15] 33-bed NICU (223 patients)
Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 941–945
[17] 182 ICUs; 1313 pneumonia
cases and 513 BSI cases
[16] 1 ICU, 50 patients
ARIMA, autoregressive integrated moving average; BO, bed occupancy; BSI, bloodstream infection; ESBL, extended-spectrum b-lactamase; ICU, intensive-care unit; MLRM, multivariate linear regression model; MRSA, methicillin-resistant
Staphylococcus aureus; NICU, neonatal intensive-care unit; PSM, patient safety measures; RR, risk ratio.
Investigation period/type
of study
24 months of time-series data
(surveillance data);
retrospective
65 months of time-series data
(surveillance data); MLRM
(ARIMA); prospective
67 months of time-series data
(surveillance data); MLRM
(ARIMA); retrospective
Two periods of annual
cross-sectional data; simple
correlation coefficients
(surveillance data);
retrospective
One period of annual
cross-sectional data, simple
correlation coefficients
(surveillance data);
retrospective
12 months; multivariate logistic
regression; retrospective
67 months of time-series data
(surveillance data); MLRM
(ARIMA); retrospective
67 months of time-series data
(surveillance data); MLRM
(ARIMA); retrospective
48 months; multivariate logistic
regression (surveillance data);
retrospective
73 months of time-series data
(surveillance data), MLRM
(ARIMA); prospective
12 months, linear regression
model (surveillance data);
retrospective
19 months of time-series data,
simple correlation
coefficients; retrospective
944
Type of infection/infectious Potential threats
agent BO rate (%) to validity Main findings Auxiliary findings
MRSA infections 76.3 (73–86) Not controlled for BO and MRSA incidence –
the number of positively correlated
patient bed-days
MRSA infections 110 (91–124) Episodes of overcrowding Seasonal fluctuations with
triggered increases in increase in BO and
MRSA incidence antibiotic use and
decrease in staff and PSM
during winter months
MRSA infections and 78 (64–90; general BO and MRSA incidence BO and PSM negatively
colonizations wards) positively correlated correlated
MRSA bactaeremia 81.3 (62–92; first BO and MRSA incidence Turnover interval and
period); 82.5 (68–95; positively correlated in MRSA incidence
second period) second period, but not in negatively correlated in
first period second period, but not in
first period
MRSA bactaeremia 84.9 (76–94) BO and MRSA incidence –
positively correlated
Enterobacter cloacae Alternative measures Small sample size RR for acquisition of High non-compliance rate
infections and of overcrowding/ E. cloacae markedly (37%) of staff with
colonizations understaffing higher in times of hand-washing and hand
overcrowding disinfection
MRSA infections and 78 (69–86; general BO and MRSA incidence Incidence of nosocomial
colonizations; ESBL-producing wards) and BO and ESBL MRSA and ESBL showed
Gram-negative infections and incidence positively decreasing and increasing
Clinical Microbiology and Infection, Volume 18 Number 10, October 2012
colonizations correlated trends, respectively
Clostridium difficile 78 (64–90; general BO and C. difficile Length of stay and
(toxin-positive stool samples wards) incidence positively C. difficile incidence
and/or cultured isolates) correlated negatively correlated
MRSA infections and Alternative measures Small sample size No evidence that staffing
colonizations of overcrowding/ or workload predicts risk
understaffing of MRSA acquisition
MRSA infections and 85.8 BO and MRSA incidence Use of hand sanitizers
colonizations not significantly negatively correlated with
correlated MRSA incidence
BSI; pneumonia 83 Fewer infections Nurse/patient ratio not
associated with high BO significantly associated
and higher with infection rates
nurse/ventilated patient
ratio
MRSA (not further specified) Alternative measures Small sample size Significant correlation –
of overcrowding/ between MRSA incidence
understaffing and: staff/patient ratio;
nurse/patient ratio; and
staff/workload ratio
CMI
CMI Kaier et al.
higher (up to 120% during the winter months) than in any
other study of this review (median of 81.2%).
A few studies suggested that high level of HAIs in the UK
might have been attributable to the fact that BO rates
exceeded the European average [8,21]. Consequently, it was
proposed to define a threshold BO rate to deliver effective
and safer healthcare. In detail, an absolute maximum BO rate
in the range of 82–85% was proposed to keep the level of
HAIs at the minimum possible level [8,21].
The question remains of whether the implementation of
such a threshold BO rate would be rational in comparison
with other patient safety measures. A recent study was
aimed at answering this question by comparing the impacts
of the BO rate and of hand disinfection, respectively, on the
cost of MRSA-related infections [22]. Not surprisingly, the
results indicated that hand disinfection is highly (cost-)effec-
tive. The economic interpretation of the correlation between
the BO rate and the MRSA incidence rates, however, leads
to another conclusion: One additionally occupied bed influ-
ences the BO rate and thus influences the number of HAIs,
but the overall impact per single occupied bed is extremely
low. In detail, keeping an additional bed unoccupied for 24 h
saves only €0.39 in potential costs incurred by MRSA-related
infections, an amount that is negligible [22].
These results are based on crude extrapolations, focus on
the German reimbursement scheme, and may not be gener-
alizable to the situation in other European countries. They
may, however, underline the relevance of patient safety mea-
sures such as hand disinfection. Further research is needed
to answer the question of whether there should be a thresh-
old BO rate. The implementation of such a threshold BO
rate, however, will always involve keeping beds empty, which
is a costly measure, irrespective of the national healthcare
organization and reimbursement scheme.
Transparency Declaration
The authors declare that they have no competing interests.
References
1. European Centre for Disease Prevention and Control. Annual epide-
miological report on communicable diseases in Europe 2008. Stockholm:
European Centre for Disease Prevention and Control, 2008.
2. McGill L. Patient safety: a European Union priority. Clin Med 2009; 9:
136–139.
3. The Council of the European Union. Council recommendation on
patient safety, including the prevention and control of healthcare
associated infections. Off J Eur Union 2009; C151: 1–6.
Clinical Microbiology and Infection ª2012 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 941–945
Bed occupancy rates and hospital-acquired infections 945
4. McGowan JE. Economic impact of antimicrobial resistance. Emerg
Infect Dis 2001; 7: 286–292.
5. Kaier K, Meyer E, Dettenkofer M, Frank U. Epidemiology meets
econometrics: using time-series analysis to observe the impact of bed
occupancy rates on the spread of multidrug-resistant bacteria. J Hosp
Infect 2010; 76: 108–113.
6. Hugonnet S, Harbarth S, Sax H, Duncan RA, Pittet D. Nursing
resources: a major determinant of nosocomial infection? Curr Opin
Infect Dis 2004; 17: 329–333.
7. Clements A, Halton K, Graves N et al. Overcrowding and understaff-
ing in modern health-care systems: key determinants in meticillin-
resistant Staphylococcus aureus transmission. Lancet Infect Dis 2008; 8:
427–434.
8. Orendi J. Health-care organisation, hospital-bed occupancy, and
MRSA. Lancet 2008; 371: 1401–1402.
9. Cunningham JB, Kernohan WG, Rush T. Bed occupancy, turnover
intervals and MRSA rates in English hospitals. Br J Nurs 2006; 15:
656–660.
10. Borg MA. Bed occupancy and overcrowding as determinant factors in
the incidence of MRSA infections within general ward settings. J Hosp
Infect 2003; 54: 316–318.
11. Borg MA, Suda D, Scicluna E. Time-series analysis of the impact of
bed occupancy rates on the incidence of methicillin-resistant Staphylo-
coccus aureus infection in overcrowded general wards. Infect Control
Hosp Epidemiol 2008; 29: 496–502.
12. Conrad A, Kaier K, Frank U, Dettenkofer M. Are short training ses-
sions on hand hygiene effective in preventing hospital-acquired
MRSA? A time-series analysis. Am J Infect Control 2010; 38: 559–561.
13. Cunningham JB, Kernohan WG, Sowney R. Bed occupancy and turn-
over interval as determinant factors in MRSA infections in acute set-
tings in Northern Ireland: 1 April 2001–31 March 2003. J Hosp Infect
2005; 61: 189–193.
14. Kong F, Cook D, Paterson DL, Whitby M, Clements A. Do staffing
and workload levels influence the risk of new meticillin-resistant
Staphylococcus aureus acquisitions in a well-resourced intensive care
unit? J Hosp Infect 2012; 80: 331–339.
15. Sakamoto F, Yamada H, Suzuki C, Sugiura H, Tokuda Y. Increased
use of alcohol-based hand sanitizers and successful eradication of
methicillin-resistant Staphylococcus aureus from a neonatal intensive
care unit: a multivariate time series analysis. Am J Infect Control 2010;
38: 529–534.
16. Vicca AF. Nursing staff workload as a determinant of methicillin-resis-
tant Staphylococcus aureus spread in an adult intensive therapy unit.
J Hosp Infect 1999; 43: 109–113.
17. Schwab F, Meyer E, Geffers C, Gastmeier P. Understaffing, over-
crowding, inappropriate nurse:ventilated patient ratio and nosocomial
infections: which parameter is the best reflection of deficits? J Hosp
Infect 2012; 80: 133–139.
18. Kaier K, Luft D, Dettenkofer M, Kist M, Frank U. Correlations
between bed occupancy rates and Clostridium difficile infections—a
time-series analysis. Epidemiol Infect 2011; 139: 482–485.
19. Harbarth S, Sudre P, Dharan S, Cadenas M, Pittet D. Outbreak of
Enterobacter cloacae related to understaffing, overcrowding, and poor
hygiene practices. Infect Control Hosp Epidemiol 1999; 20: 598–603.
20. Keegan AD. Hospital bed occupancy: more than queuing for a bed.
Med J Aust 2010; 193: 291–293.
21. Jones R. Hospital bed occupancy demystified. Br J Healthcare Manage
2011; 17: 242–248.
22. Kaier K. Economic implications of the dynamic relationship between
antibiotic use and hospital-acquired infections. Value Health 2012; 15:
87–93.
ª2012 The Authors