infection control and hospital epidemiology

february 2011, vol. 32, no. 2

original article

Estimating the Proportion of Healthcare-Associated Infections That

Are Reasonably Preventable and the Related Mortality and Costs
Craig A. Umscheid, MD, MSCE;1,2,3 Matthew D. Mitchell, PhD;1 Jalpa A. Doshi, PhD;1,3
Rajender Agarwal, MD, MPH;1 Kendal Williams, MD, MPH;1,3 Patrick J. Brennan, MD2,3,4

objective. To estimate the proportion of healthcare-associated infections (HAIs) in US hospitals that are reasonably preventable,
along with their related mortality and costs.
methods. To estimate preventability of catheter-associated bloodstream infections (CABSIs), catheter-associated urinary tract infections
(CAUTIs), surgical site infections (SSIs), and ventilator-associated pneumonia (VAP), we used a federally sponsored systematic review of
interventions to reduce HAIs. Ranges of preventability included the lowest and highest risk reductions reported by US studies of moderate
to good quality published in the last 10 years. We used the most recently published national data to determine the annual incidence of
HAIs and associated mortality. To estimate incremental cost of HAIs, we performed a systematic review, which included costs from studies
in general US patient populations. To calculate ranges for the annual number of preventable infections and deaths and annual costs, we
multiplied our infection, mortality, and cost figures with our ranges of preventability for each HAI.
results. As many as 65%70% of cases of CABSI and CAUTI and 55% of cases of VAP and SSI may be preventable with current
evidence-based strategies. CAUTI may be the most preventable HAI. CABSI has the highest number of preventable deaths, followed by
VAP. CABSI also has the highest cost impact; costs due to preventable cases of VAP, CAUTI, and SSI are likely less.
conclusions. Our findings suggest that 100% prevention of HAIs may not be attainable with current evidence-based prevention
strategies; however, comprehensive implementation of such strategies could prevent hundreds of thousands of HAIs and save tens of
thousands of lives and billions of dollars.
Infect Control Hosp Epidemiol 2011;32(2):101-114

In 1999, the Institute of Medicine released its groundbreaking

report To Err Is Human.1 The report acknowledged the
overwhelming incidence and cost of medical errors that occur
in our healthcare system and identified healthcare-associated
infections (HAIs) as an important patient safety challenge.
Subsequent studies have further documented the extent of
HAIs in the United States.2
Numerous studies have examined interventions to reduce
the most common HAIs: catheter-associated bloodstream infection (CABSI), ventilator-associated pneumonia (VAP),
catheter-associated urinary tract infection (CAUTI), and surgical site infection (SSI).3,4 These strategies have demonstrated
some success, and payers have responded by using financial
incentives to encourage hospitals to adopt them. In October
2008, Medicare stopped providing reimbursement for treatment of 8 largely preventable conditions, 3 of which it deemed
never events and 5 reasonably preventable.5 Three of the
5 reasonably preventable conditions are HAIs; namely,

CABSI, CAUTI, and SSI. VAP is being considered for inclusion in an expanded list scheduled for release in 2011.6
Although nonpayment for treatment of HAIs may be an
effective incentive for hospitals and physicians to reduce the
incidence of HAIs, some have asserted that not all HAIs are
preventable and that this new incentive may be a challenge
for hospitals that care for patients at high risk for HAIs.5,6 To
inform discussions regarding the preventability of HAIs, we
estimated the proportion of HAIs that are reasonably preventable in US hospitals, as well as their associated mortality
rates and costs.

methods
HAI Incidence, Associated Mortality, and Risk Reduction
An accurate estimation of the annual number of preventable
HAIs requires accurate estimates of 2 underlying values: the
current total annual number of HAIs and the proportion of

Affiliations: 1. Center for Evidence-Based Practice, University of Pennsylvania, Philadelphia, Pennsylvania; 2. Center for Clinical Epidemiology and
Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania; 3. Department of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania;
4. Office of the Chief Medical Officer, University of Pennsylvania, Philadelphia, Pennsylvania.
Received May 12, 2010; accepted August 10, 2010; electronically published January 14, 2011.
2011 by the Society for Healthcare Epidemiology of America. All rights reserved. 0899-823X/2011/3202-0001$15.00. DOI: 10.1086/657912

102

infection control and hospital epidemiology

february 2011, vol. 32, no. 2

these that are reasonably preventable. Likewise, an accurate

estimation of the number of preventable HAI-associated
deaths requires accurate estimates of the current total annual
number of deaths from HAIs and the proportion of these
deaths that are reasonably preventable. We obtained data
for CABSI, VAP, CAUTI, and SSI.
To estimate the number of HAIs and their associated mortality rates, we used figures assembled by Klevens et al2 from
the National Nosocomial Infections Surveillance (NNIS) system, the National Hospital Discharge Survey, and the American Hospital Association. These estimates include HAIs in
infants, children, and adults in and outside of the intensive
care unit (ICU) setting. HAIs were defined as infections that
occurred during a hospitalization and that were not present
prior to hospital admission. Infections had to meet body site
specific criteria. Death was considered associated with an HAI
if an infection preventionist determined that the HAI directly
contributed to or caused the death.
To estimate the proportion of HAIs that could be prevented, we used the estimates of HAI risk reductions resulting
from quality improvement strategies reported in an Agency
for Healthcare Research and Quality (AHRQ) Evidence-Based
Practice Center (EPC) report,3 a systematic review of the
published literature on HAI prevention. We used only HAI
risk reductions reported in US studies published in the past
10 years that had a controlled or time-series design or were
graded as good quality by the AHRQ EPC report.3 When
there were fewer than 3 studies that met these criteria, we
included studies graded as moderate quality. Quality ratings
were based on answers to questions that pertained to the
internal and external validity of the studies.3
We obtained the full text of all studies cited in our report
for further detailed analysis. For studies with a simple beforeafter design, we calculated risk reductions from the reported
infection rates before and after the study intervention. For
studies with a controlled before-after design, we estimated
the risk reduction of the intervention by calculating the difference between the risk reductions in the control arm and
the intervention arm. For studies using an interrupted timeseries design, we used the risk reductions estimated by modeling, if that was available; if modeling was not available, we
estimated risk reductions using data from the first and last
time periods.
Because the patient populations and interventions tested
in the published studies of HAI reduction varied, it was not
appropriate to combine the risk reductions into a single summary estimate using meta-analysis. Thus, we used the highest
and lowest infection reduction values from US studies for
each type of HAI to generate a range of possible risk reductions for each type of HAI. The ranges were multiplied by
the estimated number of infections, as reported by Klevens
et al,2 to yield an estimated range of the number of reasonably
preventable infections for each type of HAI. The ranges were
also multiplied by the estimated number of deaths as reported
by Klevens et al,2 to yield an estimated range of the number

of reasonably preventable deaths for each type of HAI.2 To

estimate the costs associated with each type of HAI, we multiplied the number of potentially preventable infections by
the estimated incremental cost for that type of infection.
HAI Cost
Estimating the cost of preventable HAIs requires knowledge
of the total incidence of HAIs, the proportion of HAIs that
are reasonably preventable, and the cost of each type of
HAI. The costs of CABSI, VAP, CAUTI, and SSI were obtained
from studies identified by a MEDLINE search. Search strategies for HAIs were based on those used in the AHRQ EPC
report.3 Search strategies for economic impact were based on
those used by the Canadian Agency for Drugs and Technologies in Health.7 The 2 search elements were combined, and
results were limited to English-language studies published in
the 10 years prior to the search; searches were completed in
May 2008.
The titles and abstracts of all articles found by the searches
were reviewed by a research analyst (M.D.M.), and potentially
relevant articles were retrieved. Studies were included in the
analysis if they reported original calculations of costs for one
of the selected HAIs, were conducted at US hospitals, and
included 10 or more patients with infection. Studies from
outside the United States were excluded because their cost
estimates may not be reflective of those in the United States.
Studies of highly specialized or narrow patient populations
were also excluded, since our aim was to estimate costs for
the general population.
Cost results from each study were reported as the incremental cost to the hospital per case patient with an HAI. We
calculated confidence intervals for the mean cost per case if
these intervals were not reported by study authors and if
sufficient information was available to do so. All results were
converted to 2009 dollars using the Consumer Price Index
(CPI) for Hospital Services (US Bureau of Labor Statistics),
except for one study of CABSI,8 which included data from
before 1997; for this study, the CPI for Medical Services was
used as the inflator, because the CPI for Hospital Services
was not published until 1997. Where possible, summary estimates of cost were based only on studies that used regression
models to isolate the costs of the infection from costs that
may have been coincident with the infection. Where multiple
studies for a particular infection measured costs the same
way, we took their range of estimates.

results
Number of HAIs and Associated Mortality
A comprehensive estimate of annual incidence of the mortality rates associated with HAIs in US hospitals was reported
in 2007 (Appendix Table A1).2 These data suggest that CABSI
and VAP cause more than two-thirds of the deaths resulting
from HAIs and that they are 5 times as deadly as the other
HAIs.

reasonably preventable hai s

Proportion of HAIs and HAI-Associated Deaths That Are

Reasonably Preventable
From a total of 4,847 potentially relevant articles identified
in the AHRQ EPC report,3 434 articles were retrieved, and
64 ultimately met the inclusion criteria (Appendix Table A2).3
However, the quality of the studies was generally poor. Few
were controlled or time-series analyses, and most of the simple before-after studies were categorized by the AHRQ EPC
report3 as moderate or poor quality. They did not grade
the quality of controlled and interrupted time-series trials,
assuming they were of higher quality than the simple beforeafter studies. Because there was little consistency among patient groups studied or interventions tested, the AHRQ investigators could not perform any quantitative synthesis of
the data, and they did not attempt to make a summary estimate of the proportion of infections or deaths that could
be considered preventable.3 Appendix Table A2 shows the
number of studies excluded on quality grounds and for other
reasons.
The characteristics of the included studies are summarized
in Table 1. All of the good-quality CABSI prevention studies
reported in the AHRQ EPC report3 focused on ICU patients.
The interventions tested varied from study to study, as did
absolute infection rates both before and after the interventions, suggesting that either patient populations varied across
studies or some centers already had effective infection prevention measures in place before their studies commenced.
With respect to VAP prevention studies, there were only 2
good-quality before-after studies and no studies utilizing
other designs, so we broadened our inclusion criteria to include 3 additional moderate-quality studies. As with CABSI
rates, VAP rates also varied from study to study, both before
and after implementation of infection prevention programs.
In the case of CAUTI prevention studies, there were no available good-quality studies and only 2 studies of moderate
quality. The SSI studies tested a variety of interventions, and
their results showed more heterogeneity than those seen in
studies of any other type of HAI.
Costs of HAIs
The HAI-related costs to hospitals in the included studies are
summarized for each type of HAI in Tables 2, 3, 4, and 5.
Four studies of the cost of CABSI in ICU patients met our
inclusion criteria (Table 2). One study10 identified costs associated with CABSIs by reviewing charts and each line item
in patient bills to ascertain whether the cost was attributable
to infection. Costs determined by this method were consistent
with those from the other studies. Two studies11,12 used regression models to isolate incremental costs of CABSI; however, the mean incremental cost was $21,400 in 2009 dollars
in one11 and $110,800 in the other.12 The first of these 2 studies
was hampered by an extremely wide confidence interval.
For VAP, of the 4 included cost studies (Table 3), 3 reported
similar results in unadjusted mean incremental costs per case,

103

but the other study13 reported a much higher figure. Only

the study of Warren and colleagues14 used regression to calculate a mean adjusted incremental cost per infection
($23,000).
All 3 studies of CAUTI (Table 4) estimated costs by summing the costs of specific line items associated with CAUTI.
One study15 considered only the cost of laboratory tests used
to diagnose CAUTI and medications used to treat it. The
other 2 studies16,17 counted those costs plus an assumed 0.5day increase and a 1-day increase in hospital length of stay,
respectively. Our summary cost calculation ($1,200$4,700)
is based on the range of estimates in the included studies.
The 4 studies of SSI costs (Table 5) differed considerably
in methods. Unadjusted mean or median costs per infection
fell in the range of $5,600$12,900. One study18 also reported
adjusted mean incremental costs, which were much lower, at
$2,200 per SSI.
To arrive at a national total cost of HAIs, we selected a
summary estimate of incremental costs per infection case for
each type of infection. To do this, we first used adjusted
estimates on the basis of regression models to account for
confounding variables. Regression models were available for
all infections except CAUTI, for which cost estimates were
necessarily based on studies that simply summed line-item
costs presumed to be associated with the infection, so we
took the range of those estimates as our summary estimate.
There were 2 estimates of CABSI costs based on regression
models, but they differed widely from each other; therefore,
we used their range as the basis for calculating our range of
the estimated total cost of CABSI. Only 1 adjusted incremental cost estimate was available for VAP and for SSI to be
used in our total cost calculations.
Estimates of the Total Impact of HAIs
The literature suggests that as many as 65%70% of cases of
CABSI and CAUTI and 55% of cases of VAP and SSI are
preventable with current evidence-based strategies. CAUTI
may be the most preventable HAI; the number of avoidable
infections ranges from 95,483 to 387,550 per year. This is
followed by CABSI, with 44,762164,127 preventable infections; VAP, with 95,078137,613 preventable infections; and
SSI, with 75,526156,862 preventable infections.
Our calculations demonstrate that CABSI is associated with
the highest number of preventable deaths, followed by VAP.
If best practices in infection control were applied at all US
hospitals, the reduction in the number of cases of CABSI
could save as many as 5,52020,239 lives, and for VAP 13,667
19,782 lives could be saved. The potential to save lives by
reducing the number of cases of CAUTI and SSI is smaller:
2,2259,031 lives annually for CAUTI and 2,1334,431 lives
annually for SSI.
Of the HAIs we examined, preventable cases of CABSI are
likely to have the highest associated costs, ranging anywhere
from $960 million to $18.2 billion annually. The hospital costs

Sherertz et al32 (2000)

Coopersmith et al31 (2002)

Warren et al30 (2003)

Warren et al29 (2004)

Coopersmith et al28 (2004)

Berenholtz et al27 (2004)

CABSI (good-quality studies)

Pronovost et al26 (2006)

Type of HAI, study

Intervention

Interrupted time-series; Preventive: Hand hygiene; maximum sterile

ICU patients
barrier precautions; insertion site selection; chlorhexidine disinfection; removal
of unnecessary catheters
QI: Clinician education, audit, and feedback; clinician reminder; organizational
change
Controlled before-after Intervention period: Preventive: Hand hystudy; ICU patients
giene; maximum sterile barrier precautions; insertion site selection; chlorhexidine disinfection; removal of
unnecessary catheters
QI: Clinician education, audit, and
feedback
Control period: Clinician education only
Before-after study;
Preventive: Hand hygiene; maximum sterile
ICU patients
barrier precautions; insertion site
selection
QI: Clinician education
Before-after study;
Preventive: Hand hygiene; maximum sterile
ICU patients
barrier precautions; insertion site
selection
QI: Clinician education, audit, and
feedback
Before-after study;
Preventive: Maximum sterile barrier preICU patients
cautions; insertion site selection
QI: Clinician education, audit, and
feedback
Before-after study;
Preventive: Hand hygiene
ICU patients
QI: Clinician education, audit, and
feedback
Before-after study;
Preventive: Hand hygiene; maximum sterile
ICU patients
barrier precautions
QI: Clinician education

Study design and

type of patients

2.1 cases per 1,000 CDs

5.5 cases per 1,000 CDs

2.8 cases per 1,000 CDs

Intervention period: 0
cases per 1,000 CDs;
control period: 1.6
cases per 1,000 CDs

1.4 cases per 1,000 CDs

4.51 cases per 1,000 CDs 2.92 cases per 1,000 CDs

10.8 cases per 1,000 CDs 3.7 cases per 1,000 CDs

4.9 cases per 1,000 CDs

9.4 cases per 1,000 CDs

3.4 cases per 1,000 CDs

Intervention period:
11.3 cases per 1,000
CDs; control period:
5.7 cases per 1,000
CDs

7.7 cases per 1,000 CDs

After intervention

Incidence or risk of infection

Before intervention

table 1. Summaries of Studies of Prevention of Healthcare-Associated Infections Included in the Present Analysis

35%

66%

57%

42%

18%

28%b

66%a

Reduction

Before-after study;
ICU patients

Before-after study;
ICU patients

Before-after study;
ICU patients

Before-after study;
type of patients
not reported

Rao et al40 (2004)

Preventive: Appropriate use of perioperative antibiotics; decreased use of preoperative shaving; improvement in perioperative glucose control
QI: Audit and feedback; clinician education; clinician reminder
Preventive: Improvement in perioperative
glucose control
QI: Audit and feedback; clinician education; patient education
Preventive: Appropriate use of perioperative antibiotics; decreased use of preoperative shaving; improvement in perioperative glucose control
QI: Clinician education, clinician reminder

Preventive: Reduction in placement of

catheters; removal of unnecessary
catheters
QI: Clinician education; clinician reminder; organizational change
Preventive: Aseptic insertion and catheter
care; removal of unnecessary catheters
QI: Clinician education; organizational
change

Preventive: Head of bed angle 130

QI: Clinician education, audit, and
feedback

2.1%

7.58%

2.3%

SICU: 10.3 cases per

1,000 CDs; MICU:
15.8 cases per 1,000
CDs; CICU: 15.1
cases per 1,000 CDs

36 cases per 1,000 CDs

SICU: 45.1 cases per

1,000 VDs; MICU:
22.4 cases per 1,000
VDs

1.5%

3.47%

1.7%

SICU: 8.6 cases per

1,000 CDs; MICU:
11.2 cases per 1,000
CDs; CICU: 8.3 cases
per 1,000 CDs

11 cases per 1,000 CDs

SICU: 27.9 cases per

1,000 VDs; MICU:
11.6 cases per 1,000
VDs

Preventive: Hand hygiene; head of bed an- 8.75 cases per 1,000 VDs 4.74 cases per 1,000 VDs
gle 130; daily interruption of sedation
QI: Clinician education
Preventive: Head of bed angle 130
12.6 cases per 1,000 VDs 5.7 cases per 1,000 VDs
QI: Clinician education

29%

54%

26%

SICU: 17%;
MICU: 29%;
CICU: 45%

69%

SICU: 38%;
MICU: 48%

55%

46%

note. CABSI, catheter-associated bloodstream infection; CAUTI, catheter-associated urinary tract infection; CD, catheter-day; ICU, intensive care unit; CICU, cardiology ICU;
MICU, medical ICU; QI, quality initiative; SICU, surgical ICU; SSI, surgical site infection; VAP, ventilator-associated pneumonia; VD, ventilator-day.
a
Reported risk reduction resulting from interrupted time-series modeling.
b
Risk reduction calculated by taking the difference between the risk reductions of the intervention arm and the control arm.

Before-after study;
type of patients
not reported

Before-after study;
type of patients
not reported

Before-after study;
ICU patients

Lutarewych et al39 (2004)

SSI (moderate quality studies)

Dellinger et al38 (2005)

Dumigan et al37 (1998)

CAUTI (moderate-quality studies)

Topal et al36 (2005)
Before-after study;
ward patients

VAP (moderate-quality studies)

Lai et al35 (2003)

Zack et al34 (2002)

VAP (good-quality studies)

Babcock et al33 (2004)

St. Louis, MO
Surgical and medical ICU
41
Primary
Cost identification
By infection control team using CDC
criteria

Cost identification
Not reported

Warren et al11

Pittsburgh, PA
Medical and coronary ICU
54
Primary

Shannon et al10
Baltimore, MD
Surgical ICU
86
Primary

Dimick et al12
Ann Arbor, MI
Medical ICU
68
Secondary

DiGiovine et al8

note. Costs were converted to 2009 dollars using the Consumer Price Index (CPI) for Hospital Services (US Bureau of Labor Statistics), except for DiGiovine et al,8 for which costs were converted
using the CPI for Medical Services, since the hospital index was not calculated before 1997. APACHE II, Acute Physiology and Chronic Health Evaluation II; CD, catheter-day; CDC, Centers for Disease
Control and Prevention; CFU, colony-forming unit; CI, confidence interval; ICU, intensive care unit; LOS, length of stay; SD, standard deviation; VD, ventilator-day.
a
No. of patients with infection; excludes matched control subjects.

Cost identification
Cost identification
Colonization of catheter (115 CFU) By infection control team using
with organism found in periphCDC criteria
eral blood specimen
Control group
None
ICU patients without infection
ICU patients without infection
Matched control subjects; matching
based on predicted mortality,
sex, age, race, admitting diagnosis, and chronic health
Method of determining cost
Clinicians retrospective review of charts,
Multiple linear regression model
Multiple linear regression model
By LOS and by total direct costs;
bills, and payments
which costs were reported is not
stated
Source of cost data (baseline year) Hospital cost reports, 20022005 dollars
Hospital cost accounting database, 2000 Hospital charges converted to costs, Hospital cost accounting database,
dollars
1998 dollars
19941996 dollars
Costs measured
Line item costs of care attributable to CABSI All costs in hospital accounting database, All costs in hospital billing database; Cost analysis methods ambiguous;
(eg, additional hospital days, antibiotics,
including overhead costs
overhead costs not reported
it appears that only direct costs
and tests) and/or its complications (eg,
for ICU care were considered
exploratory laparotomy, hemodialysis)
Perspective
Hospital
Hospital
Hospital
Hospital
Time horizon
Inpatient stay
Inpatient stay
Inpatient stay
ICU component of stay
Main economic outcome
Mean incremental direct cost per hospitali- Adjusted mean incremental total cost
Adjusted mean incremental total
Mean incremental ICU cost attribzation attributable to the CABSI
per hospitalization attributable to the
cost per hospitalization attributautable to the CABSI
CABSI
ble to the CABSI
Multivariate adjustment made to No
Yes; regression model controlled for
Yes; regression model controlled for No
cost estimates
APACHE II score, heart failure, heAPACHE III score and age
modialysis, ventilator-days, and corticosteroid use
Unadjusted results (as published) Mean, $40,179 (SD not reported)
Median, $63,572 (75th95th quartile
Median, $62,652 (75th95th quartile All patients: mean, $23,751;
range, $39,314$84,871)
range, $17,439$170,799)
survivors: mean, $34,508 (SDs
not reported)
Adjusted results (as published)
No multivariate analysis
Mean, $11,971 (95% CI, $6,732
Mean, $56,167 (95% CI, $11,523
No multivariate analysis
$18,352)
$165,735)
Unadjusted results (2009 dollars) Mean, $56,000 (SD not reported)
Median, $113,700 (75th95th quartile
Median, $123,600 (75th95th quar- All patients: mean, $41,900;
range, $70,300$151,700)
tile range, $34,400$337,000)
survivors: mean, $60,900
(SDs not reported)
Adjusted results (2009 dollars)
No multivariate analysis
Mean, $21,400 (95% CI, $12,000
Mean, $110,800 (95% CI, $22,700 No multivariate analysis
$32,800)
$327,000)
Comments
Three clinician reviewers had to agree costs
Separate analyses done for all pawere attributable to the CABSI or its
tients and for patients who surcomplications for costs to be included;
vived to discharge
average loss to hospital, $26,885

City
Type of patients, by hospital site
No. of patientsa
Cost analysis primary or
secondary aim?
Purpose of economic analysis
Method of defining infection

Variable

table 2. Summary of 4 Studies of the Costs Associated with Catheter-Associated Bloodstream Infection (CABSI)

Houston, TX
Trauma ICU
70
Primary
Cost identification
By infection control team using
NNIS criteria
Matched control subjects: matching
based on age and Injury Severity
Score

Cost identification
By infection control team using NNIS
criteria
Patients in same ICU without
infection

Cocanour et al41

Kansas City, MO
Trauma ICU
13
Secondary

Lansford et al13

Cost identification
By infection control team using NNIS
criteria
Patients in same ICU without infection

St. Louis, MO
Surgical and medical ICUs
127
Primary

Warren et al14

Cost identification
Not reported

Nationwide
ICU
816
Secondary

Rello et al42

note. Costs were converted to 2009 dollars using the Consumer Price Index for Hospital Services (US Bureau of Labor Statistics). CDC, Centers for Disease Control and Prevention; CFU,
colony-forming unit; CI, confidence interval; ICU, intensive care unit; NNIS, National Nosocomial Infections Surveillance.
a
No. of patients with infection; excludes matched control subjects.

Matched control subjects:

matching based on type of
admission, predicted mortality, duration of ventilation,
and age
Method of determining cost
Average total costs for patients with
Average total costs for patients with Average total costs for patients with
Average total charges for patients
VAP vs patients without VAP
VAP vs patients without VAP
VAP vs patients without VAP
with VAP vs patients without
VAP
Source of cost data (baseline year) Not reported, 20032004 dollars
Hospital cost accounting database,
Hospital cost accounting database,
Hospital billed charges database,
20022003 dollars
19981999 dollars
19981999 dollars
Costs measured
Total hospital costs and charges; details Total ICU costs; details and overhead All costs in database, including overhead All charges in database, overhead
and overhead costs not reported
costs not reported
costs not reported
Perspective
Hospital
Hospital
Hospital
Hospital
Time horizon
Not reported
ICU stay
Inpatient stay
Not reported
Main economic outcome
Mean incremental charges per hospiMean incremental ICU costs per stay Adjusted mean incremental costs per
Mean incremental charges per
talization attributable to VAP
attributable to VAP
hospitalization attributable to VAP
hospitalization attributable to
VAP
Multivariate adjustment made to No
No
Yes; regression model controlled for
No
cost estimates
APACHE II score, heart failure,
CABSI, hemodialysis, tracheostomy,
number of CVCs, H2 blocker use,
corticosteroid use
Unadjusted results (as published) Mean, $233,099 (95% CI, $106,200
Mean, $57,158 (95% CI, $39,300
Mean, $48,948 (95% CI, $38,617
Mean, $41,294 (95% CI,
$360,000)
$75,000)
$59,278)
$34,900$47,700)
Adjusted results (as published)
No multivariate analysis
No multivariate analysis
Mean, $11,897 (95% CI, $5,265
No multivariate analysis
$26,214)
Unadjusted results (2009 dollars) Mean, $324,000 (95% CI, $148,000
Mean, $84,700 (95% CI, $58,200
Mean, $94,600 (95% CI, $75,600
Mean, $80,200 (95% CI,
$500,000)
$111,000)
$114,000)
$67,500$92,200)
Adjusted results (2009 dollars)
No multivariate analysis
No multivariate analysis
Mean, $23,000 (95% CI, $10,100
No multivariate analysis
50,700)
Comments
Article method reports costs, results
report charges

Control group

City
Type of patients, by hospital site
No. of patientsa
Cost analysis primary or
secondary aim?
Purpose of economic analysis
Method of defining infection

Variable

table 3. Summary of 4 Studies of the Costs Associated with Ventilator-Associated Pneumonia (VAP)

Seattle, WA
Inpatients
No cases
Secondary
Cost estimation
No cases

Cost identification
Bacteria or fungi at concentration of
11,000 CFU/mL
No control
Clinicians retrospective review of
charts and bills

Saint et al16

Madison, WI
Inpatients
123
Secondary

Tambyah et al15

Estimate, $2,471
No multivariate analysis
Estimate, $4,700
No multivariate analysis
Part of a multicenter study; cost analysis for just
1 hospital; ignored treatment, other costs; selected low estimate so as not to overestimate
impact of prevention measures

No

Hospital
Fixed as 1 additional inpatient day
Cost of ICU stay and diagnostic workup used in
management of CAUTI

1-day ICU stay, fever evaluation

Not reported, baseline year not reported

No control
Investigators estimate of additional length of
stay and testing needed

Cost estimation
No cases

Philadelphia, PA
ICU
No cases
Secondary

Bologna et al17

note. Costs were converted to 2009 dollars using the Consumer Price Index for Hospital Services (US Bureau of Labor Statistics). CFU, colony-forming unit; ICU,
intensive care unit.
a
No. of patients with infection; excludes matched control subjects.

No control
Investigators estimate of additional
length of stay, testing, and treatment
needed
Source of cost data (baseline year) Not reported; 1998 dollars
Standard hospital charge multiplied by
cost/charge ratio; 1998 dollars
Costs measured
Lab test costs and medication costs
0.5-day inpatient stay, urine analysis,
urine culture, and sensitivity testing;
antimicrobial therapy
Perspective
Hospital
Hospital
Time horizon
Diagnostic and treatment period
Fixed as 0.5 additional inpatient day
Main economic outcome
Cost of lab tests and medications used Cost of hospital stay, laboratory tests, and
in management of CAUTI
medications used in management of
CAUTI
Multivariate adjustment made to No
No
cost estimates
Unadjusted results (as published) Mean, $589
Estimate, $2,041
Adjusted results (as published)
No multivariate analysis
No multivariate analysis
Unadjusted results (2009 dollars) Estimate, $1,200
Estimate, $4,000
Adjusted results (2009 dollars)
No multivariate analysis
No multivariate analysis
Comments
Ignored additional physician and
Ignored all other costs, such as
nursing costs and cost of bloodnursing and physician costs; selected
stream infections
low estimates of costs

Control group
Method of determining cost

City
Type of patients, by hospital site
No. of patientsa
Cost analysis primary or
secondary aim?
Purpose of economic analysis
Method of defining infection

Variable

table 4. Summary of 3 Studies of the Costs Associated with Catheter-Associated Urinary Tract Infection (CAUTI)

255
Primary

Durham, NC
All surgery

Kirkland et al44

Patients in same hospital without

infection

Median total costs for patients with SSI

vs patients without SSI
Cost accounting database; 20012002
dollars
Labor and supply costs, details not reported, overhead not reported
Hospital
Not reported
Mean incremental cost per hospitalization associated with infectious
complication
Yes; regression model controlled for
procedure complexity, patient characteristics (details not reported), and
other complications
Median, $8,039

Mean, $1,398 (95% CI, $377$2,418)

Median, $12,900

Matched control subjects: matching

based on type of procedure, age,
and duration of procedure
Mean total costs for patients with SSI
vs patients without SSI
Hospital charges multiplied by cost/
charge ratio; 1998 dollars
All hospital and outpatient charges,
overhead not reported
Health system
8 weeks after discharge
Mean incremental cost per case associated with the SSI

Mean, $3,382 (95% CI, $1,314

$5,450)

No multivariate analysis
Mean, $6,700 (95% CI, $2,600
$10,800)

No multivariate analysis
Mean, $2,200 (95% CI, $600$3,900)
Most cases of SSI are diagnosed after Infectious complications included 41
discharge
wound infections, 10 cases of sepsis,
24 cases of wound dehiscence

No

Cost identification
Using NSQIP criteria

75
Primary

Ann Arbor, MI
General or vascular surgery

Dimick et al46

Cost identification
By investigator, using NNIS criteria

267
Primary

Boston, MA
All surgery

Perencevich et al45

note. Costs were converted to 2009 dollars using the Consumer Price Index (CPI) for Hospital Services (US Bureau of Labor Statistics), except for Kirkland et al,44 for which costs were converted
using the CPI for Healthcare Services, because the hospital services component was not calculated before 1997. CDC, Centers for Disease Control and Prevention; CFU, colony-forming unit; CI,
confidence interval; ICU, intensive care unit; LOS, length of stay; NNIS, National Nosocomial Infections Surveillance; NSQIP, National Surgical Quality Improvement Program.
a
No. of patients with infection; excludes matched control subjects.
b
Depending on the type of surgery and whether the infection was fatal.

Cost identification
By independent physicians, using CDC
criteria
Patients in same hospital without
infection

Iowa City, IA
General, cardiothoracic surgery, or
neurosurgery
316
Secondary

Herwaldt et al43

Cost identification
By infection control nurse, using
NNIS criteria
Control group
Matched control subjects: matching
based on NNIS risk index, type of
procedure, age, date of surgery,
and surgeon
Method of determining cost
Median total costs for patients with SSI Mean total costs for patients with SSI
vs patients without SSI
vs patients without SSI
Source of cost data (baseline year) Hospital financial department, details
Cost accounting database; 19911995
not reported; 19951998 dollars
dollars
Costs measured
All hospital costs excluding physicians Direct costs: overhead excluded, defees, overhead not reported
tails not reported
Perspective
Hospital
Hospital
Time horizon
30 days after operation
30 days after operation
Main economic outcome
Total postoperative costs attributable to Mean incremental direct cost per
the SSI
hospitalization associated with the
SSI
Multivariate adjustment made to Yes, but results reported only as perNo
cost estimates
centages; regression model controlled
for Karnofsky score, NNIS risk index, number of comorbidities, obesity, preoperative LOS, and age
Unadjusted results (as published) Median, $3,343
Initial stay: mean, $3,089 (95% CI,
$2,139$4,163); with readmission:
mean, $5,038 (95% CI, $4,020
$6,289)
Adjusted results (as published)
Increase of 25%106%b
No multivariate analysis
Unadjusted results (2009 dollars) Median, $5,600
Initial stay: mean, $6,000 (95% CI,
$4,200$8,100); with readmission:
mean, $9,800 (95% CI, $7,900
$12,300)
Adjusted results (2009 dollars)
Increase of 25%106%b
No multivariate analysis
Comments
Cost methods given in separate
report47

No. of patientsa
Cost analysis primary or
secondary aim?
Purpose of economic analysis
Method of defining infection

City
Type of patients, by hospital site

Variable

table 5. Summary of 4 Studies of the Costs Associated with Surgical Site Infection (SSI)

110

infection control and hospital epidemiology

february 2011, vol. 32, no. 2

of preventable VAP are estimated to be $2.19 billion to 3.17

billion dollars annually. Costs of preventable CAUTIs are estimated to be $115 million to $1.82 billion annually, and the
costs of preventable SSIs are estimated to be $166 million to
$345 million.

discussion
Past studies have estimated the number of infections prevented or lives saved if hospitals followed best practices in
infection prevention and control. The Centers for Disease
Control and Preventions Study on the Efficacy of Nosocomial
Infection Control (SENIC) project made such an estimate in
1975.19 Its estimate considered that 30%35% of most HAIs
were preventable with effective surveillance and control programs, including 22% of cases of pneumonia. In a 1985 follow-up survey,20 the SENIC project found that only a fraction
of those infections were actually being prevented, because
many hospitals still had not implemented recommended infection control measures. This was still the case in the present
decade.21 Our estimated ranges of potential reductions in
HAIs are in line with the most recent estimates by Kaye et al.22
The considerable uncertainty in our estimates of preventable HAIs and the associated mortality and costs stems from
both the component numbers and the calculations themselves. First, while our estimates of the annual numbers of
HAIs and associated deaths are based on broad national surveillance systems,2 those data are more than 5 years old and
do not capture the possibly lower infection and mortality
rates resulting from improved care practices implemented
since 2002. If care has improved since that time, the current
number of infections and deaths would be lower than those
observed in 2002. That would continue the trend observed
since 19751976, when the total number of HAIs estimated
by the SENIC project was 2.15 million.19 Second, there is no
definite way to attribute a death to an HAI, because patient
deaths frequently have multiple causes, and the role of infection may not always be clear. Klevens et al2 attempted to
address this by only including deaths for which an infection
preventionist determined that the HAI caused or directly contributed to the death, but this may overstate the number of
deaths of patients with HAI who may have actually died of
other causes. However, for some infectionsspecifically,
CABSIother investigators have provided higher estimates
of attributable mortality than Klevens et al.2 Pittet and colleagues23 estimated an attributable mortality of 35% in surgical ICU patients. For other HAIs, such as VAP, recent systematic reviews of the literature have highlighted difficulty of
quantifying the attributable mortality.24 Therefore, for most
HAIs additional studies are needed to determine the attributable mortality.
Certainty in the estimate of the proportion of HAIs that
are reasonably preventable is limited by the quality of the
HAI reduction studies. None of the studies was randomized,
and few were controlled, limiting the validity of reported risk

reductions. Most utilized a simple before-after design, comparing outcomes before and after an intervention to reduce
the incidence of HAIs, a design that cannot control for other
changes in patient care between the control period and the
intervention period and makes it difficult to attribute the
results to the intervention rather than to random variation,
patient selection, or other uncontrolled variables. To address
this limitation, we only included studies of good or moderate
quality in which causality could reasonably be attributed to
the intervention. In addition, some of the published studies
included in the AHRQ EPC report3 date back a decade or
more; infection prevention and control practices examined
in these older studies may be standard practice currently,
making large HAI reductions resulting from these interventions less likely in modern hospitals. To address this limitation, our analyses only included studies published in the past
decade.
Another source of uncertainty is generalizing from the results of studies in specialized populations, such as ICU patients, to patients on general hospital wards. In our review,
all but one of the CABSI, VAP, and CAUTI studies were
carried out in an ICU. The one study not performed in an
ICU examined CAUTI on a general medical ward.36 If that
study were discounted, the upper limit for the percentage of
HAIs that were reasonably preventable would fall from 69%
to 45%, which corresponds to 134,800 fewer preventable infections, 3,100 fewer preventable deaths, and $160 million to
$630 million less in costs.
The key uncertainty in estimating reasonably preventable
HAI deaths is the fact that the studies we reviewed did not
directly measure death as an outcome. Instead, we extrapolated reductions in death rates from the estimates of reductions in the number of HAIs, which have their own limitations. In addition, in multiplying the estimated fraction of
HAIs that are preventable by the estimated number of HAIrelated deaths, we assume that the proportion of deaths that
are preventable is the same as the proportion of infections
that are preventable. The true effect on deaths could be larger
or smaller, depending on the extent to which preventive measures affect the severity of HAIs and the extent to which
preventive measures work for the kinds of patients who are
more susceptible to fatal HAIs. In addition, this review focused on HAIs associated with invasive devices and surgical
procedures but did not capture data on morbidity and mortality associated with other infections, such as Clostridium
difficile infection.
Cost estimates are also limited, mostly by the poor design
of the available studies. In general, 2 types of cost analyses
were available in the published literature. The first was a raw
comparison of costs between patients with the HAI in question and patients without an infection (unadjusted results).
Some of these studies attempted to control for confounding
variables, such as patient age and disease severity, by selecting
uninfected matched control patients for each infected case
patient. Others simply compared mean or median costs for

reasonably preventable hai s

all infected patients and uninfected patients, an approach that

likely overestimates HAI costs because some of the variables
predicting increased risk of infection also predict increased
cost irrespective of infection. This limitation on the precision
of true cost estimates has been reviewed elsewhere.25 The
second type of cost analysis in the literature used regression
modeling to account for the effect of multiple variables, including HAIs, on hospital costs (adjusted results). In all instances, adjusted analyses yielded lower incremental costs per
infection than did the unadjusted analyses in the same study.
We used adjusted estimates if they were available, but in some
instances, such as for CAUTI, only a range of unadjusted
estimates was available, which increased the risk that we overestimated incremental costs.
Importantly, the cost objective of our study was to estimate
the incremental costs of HAIs to hospitals, not the cost-effectiveness of various interventions or bundles of practices
to prevent or decrease the incidence of HAIs. Hence, our cost
estimates do not factor in the costs of those interventions
required to prevent or reduce HAIs, such as those described
in the AHRQ EPC report (eg, use of chlorhexidine).3
Our study suggests that, in the patients and settings examined, HAIs have never been 100% preventable, even with
the implementation of comprehensive evidence-based infection control strategies. Instead, risk reductions may be limited
to 65%70% for CABSI and CAUTI and approximately 55%
for VAP and SSIreductions that may actually be overestimated given the limitations cited above. The magnitude of
risk reductions did not appear to be associated with study
design or study quality. However, for all HAIs the studies of
patients with the highest risk of infection prior to the intervention were often the studies that showed the greatest risk
reductions, and the studies of patients with the lowest risk
were the ones that demonstrated the smallest reductions; this
suggests regression to the mean.
Given the difficulty of preventing 100% of HAIs even with

111

comprehensive evidence-based interventions, it may be appropriate to consider reimbursement strategies that encourage hospitals to reduce the incidence of HAIs while also accounting for hospitals case mix indices. For example,
reimbursement based on a percentage reduction in the incidence of an HAI or a reduction of the number of cases of
an HAI below a threshold set according to the case mix.
In conclusion, our findings suggest that the goal of preventing 100% of HAIs may not be attainable even with use
of current evidence-based HAI prevention strategies; however, comprehensive implementation of such strategies could
prevent hundreds of thousands of HAIs and save tens of
thousands of lives and billions of dollars. Given their limitations, the figures in our study should not be used as a basis
for policy decisions but should prompt future studies with
robust designs to measure accurately the impact of HAI reduction strategies and the incremental cost of HAIs.

acknowledgments
We thank our colleague David Goldmann, MD, at the University of Pennsylvania, for reviewing the manuscript and for his many thoughtful
suggestions.
Potential conflicts of interest. P.J.B. reports that he is chair of the Healthcare Infection Control Practices Advisory Committee of the Centers for
Disease Control and Prevention and past president of SHEA. The authors
report no other potential conflicts of interest.
Address reprint requests to Craig A. Umscheid, MD, MSCE, Assistant
Professor of Medicine and Epidemiology, Director, Center for Evidence-Based
Practice, University of Pennsylvania, 3535 Market Street, Mezzanine, Suite
50, Philadelphia, PA 19104 (craig.umscheid@uphs.upenn.edu
This study was originally performed for the Society for Healthcare Epidemiology of America (SHEA) and was included in its written testimony to
the Committee on Oversight and Government Reform in its Hearing on
Healthcare-Associated Infections: A Preventable Epidemic, chaired by Henry
A. Waxman on April 16, 2008, in Washington, DC. Findings from the manuscript were subsequently presented at the 19th Annual Scientific Meeting
of SHEA in San Diego, California, in 2009.

112

infection control and hospital epidemiology

february 2011, vol. 32, no. 2

appendix a
table a1.

Reported Rates of Healthcare-Associated Infections in US Hospitals in 2002

Type of infection

No. of
infections

No. of deaths
from infection

Case fatality
rate, %

Catheter-associated bloodstream infection

Ventilator-associated pneumonia
Catheter-associated urinary tract infection
Surgical site infection
Other
Total

248,678
250,205
561,667
290,485
386,090
1,737,125

30,665
35,967
13,088
8,205
11,062
98,987

12.3
14.4
2.3
2.8
2.9
5.7

note.

Data are from Klevens et al.2

table a2. Evaluation of Studies of Prevention of Healthcare-Associated Infections

in US Hospitals Reported in the Agency for Healthcare Research and Quality (AHRQ)
Report and in the Present Report
No. of studies included,
by type of HAI
Variable
EPC report
Controlled studies
Time-series studies
Good pre-post studies
Moderate pre-post studies
Poor pre-post studies
Total
Present report
Excluded from analysis
Because quality was low
Because more than 10 years old
Because only reported process outcomes
Because from outside US
Included in analysis

CABSI

VAP

CAUTI

SSI

2
1
6
2
8
19

0
0
3
4
5
12

3
0
0
6
1
10

4
2
1
6
15
28

10
0
0
2
7

5
1
1
2
3

1
3
2
2
2

15
3
2
5
3

note. Data are from the AHRQ report by Ranji et al.3 CABSI, catheter-associated
bloodstream infection; CAUTI, catheter-associated urinary tract infection; EPC, evidence-based practice center; SSI, surgical site infection; VAP, ventilator-associated pneumonia; US, United States.

reasonably preventable hai s

references
1. Kohn LT, Corrigan JM, Donaldson MS, eds. To err is human:
building a safer health system. National Academy of Sciences,
2000.
2. Klevens RM, Edwards JR, Richards CL Jr, et al. Estimating health
care-associated infections and deaths in US hospitals, 2002. Public Health Rep 2007;122(2):160166.
3. Ranji SR, Shetty K, Posley KA, et al. Volume 6: prevention of
healthcare-associated infections. Rockville, MD: Agency for
Healthcare Research and Quality; 2007 January 2007. AHRQ
publication 04(07)-0051-6.
4. Yokoe DS, Mermel LA, Anderson DJ, et al. A compendium of
strategies to prevent healthcare-associated infections in acute
care hospitals. Infect Control Hosp Epidemiol 2008;29(suppl 1):
S12S21.
5. Wald HL, Kramer AM. Nonpayment for harms resulting from
medical care: catheter-associated urinary tract infections. JAMA
2007;298(23):27822784.
6. Pronovost PJ, Goeschel CA, Wachter RM. The wisdom and
justice of not paying for preventable complications. JAMA
2008;299(18):21972199.
7. Brown A, Wells P, Jaffey J, et al. Point-of-care monitoring devices
for long-term oral anticoagulation therapy: clinical and cost
effectiveness. Ottawa: Canadian Agency for Drugs and Technologies in Health, 2007. Technology report 72.
8. DiGiovine B, Chenoweth C, Watts C, Higgins M. The attributable mortality and costs of primary nosocomial bloodstream
infections in the intensive care unit. Am J Respir Crit Care Med
1999;160(3):976981.
9. Harbarth S, Sax H, Gastmeier P. The preventable proportion of
nosocomial infections: An overview of published reports. J Hosp
Infect 2003;54(4):258266.
10. Shannon RP, Patel B, Cummins D, Shannon AH, Ganguli G,
Lu Y. Economics of central lineassociated bloodstream infections. Am J Med Qual 2006;21(suppl 6):7S-16S.
11. Warren DK, Quadir WW, Hollenbeak CS, Elward AM, Cox MJ,
Fraser VJ. Attributable cost of catheter-associated bloodstream
infections among intensive care patients in a nonteaching hospital. Crit Care Med 2006;34(8):20842089.
12. Dimick JB, Pelz RK, Consunji R, Swoboda SM, Hendrix CW,
Lipsett PA. Increased resource use associated with catheter-related bloodstream infection in the surgical intensive care unit.
Arch Surg 2001;136(2):229234.
13. Lansford T, Moncure M, Carlton E, et al. Efficacy of a pneumonia prevention protocol in the reduction of ventilator-associated pneumonia in trauma patients. Surg Infect (Larchmt)
2007;8(5):505510.
14. Warren DK, Shukla SJ, Olsen MA, et al. Outcome and attributable cost of ventilator-associated pneumonia among intensive
care unit patients in a suburban medical center. Crit Care Med
2003;31(5):13121317.
15. Tambyah PA, Knasinski V, Maki DG. The direct costs of nosocomial catheter-associated urinary tract infection in the era of
managed care. Infect Control Hosp Epidemiol 2002;23(1):2731.
16. Saint S, Veenstra DL, Sullivan SD, Chenoweth C, Fendrick AM.
The potential clinical and economic benefits of silver alloy urinary catheters in preventing urinary tract infection. Arch Intern
Med 2000;160(17):26702675.
17. Bologna RA, Tu LM, Polansky M, Fraimow HD, Gordon DA,

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

113

Whitmore KE. Hydrogel/silver ion-coated urinary catheter reduces nosocomial urinary tract infection rates in intensive care
unit patients: a multicenter study. Urology 1999;54(6):982987.
Dimick JB, Chen SL, Taheri PA, Henderson WG, Khuri SF,
Campbell DA Jr. Hospital costs associated with surgical complications: a report from the private-sector national surgical
quality improvement program. J Am Coll Surg 2004;199(4):531
537.
Haley RW, Culver DH, White JW, Morgan WM, Emori TG. The
nationwide nosocomial infection rate: a new need for vital statistics. Am J Epidemiol 1985;121(2):159167.
Haley RW, Culver DH, White JW, et al. The efficacy of infection
surveillance and control programs in preventing nosocomial
infections in US hospitals. Am J Epidemiol 1985;121(2):182205.
Braun BI, Kritchevsky SB, Wong ES, et al. Preventing central
venous catheter-associated primary bloodstream infections:
characteristics of practices among hospitals participating in the
evaluation of processes and indicators in infection control
(EPIC) study. Infect Control Hosp Epidemiol 2003;24(12):926
935.
Kaye KS, Engemann JJ, Fulmer EM, Clark CC, Noga EM, Sexton
DJ. Favorable impact of an infection control network on nosocomial infection rates in community hospitals. Infect Control
Hosp Epidemiol 2006;27(3):228232.
Pittet D, Harbarth S. What techniques for diagnosis of ventilator-associated pneumonia? [see comment]. Lancet 1998;
352(9122):8384.
Melsen WG, Rovers MM, Bonten MJ. Ventilator-associated
pneumonia and mortality: a systematic review of observational
studies. Crit Care Med 2009;37(10):27092718.
Perencevich EN, Stone PW, Wright SB, et al. Raising standards
while watching the bottom line: making a business case for
infection control. Infect Control Hosp Epidemiol 2007;28(10):
11211133.
Pronovost P, Needham D, Berenholtz S, et al. An intervention
to decrease catheter-related bloodstream infections in the ICU.
N Engl J Med 2006;355(26):27252732.
Berenholtz SM, Pronovost PJ, Lipsett PA, et al. Eliminating catheter-related bloodstream infections in the intensive care unit.
Crit Care Med 2004;32(10):20142020.
Coopersmith CM, Zack JE, Ward MR, et al. The impact of
bedside behavior on catheter-related bacteremia in the intensive
care unit. Arch Surg 2004;139(2):131136.
Warren DK, Zack JE, Mayfield JL, et al. The effect of an education program on the incidence of central venous catheterassociated bloodstream infection in a medical ICU. Chest 2004;
126(5):16121618.
Warren DK, Zack JE, Cox MJ, Cohen MM, Fraser VJ. An educational intervention to prevent catheter-associated bloodstream infections in a nonteaching, community medical center.
Crit Care Med 2003;31(7):19591963.
Coopersmith CM, Rebmann TL, Zack JE, et al. Effect of an
education program on decreasing catheter-related bloodstream
infections in the surgical intensive care unit. Crit Care Med 2002;
30(1):5964.
Sherertz RJ, Ely EW, Westbrook DM, et al. Education of physicians-in-training can decrease the risk for vascular catheter
infection. Ann Intern Med 2000;132(8):641648.
Babcock HM, Zack JE, Garrison T, et al. An educational intervention to reduce ventilator-associated pneumonia in an inte-

114

34.

35.

36.

37.

38.

39.

40.

infection control and hospital epidemiology

february 2011, vol. 32, no. 2

grated health system: a comparison of effects. Chest 2004;125(6):

22242231.
Zack JE, Garrison T, Trovillion E, et al. Effect of an education
program aimed at reducing the occurrence of ventilator-associated pneumonia. Crit Care Med 2002;30(11):24072412.
Lai KK, Baker SP, Fontecchio SA. Impact of a program of intensive surveillance and interventions targeting ventilated patients in the reduction of ventilator-associated pneumonia and
its cost-effectiveness. Infect Control Hosp Epidemiol 2003;24(11):
859863.
Topal J, Conklin S, Camp K, Morris V, Balcezak T, Herbert P.
Prevention of nosocomial catheter-associated urinary tract infections through computerized feedback to physicians and a
nurse-directed protocol. Am J Med Qual 2005;20(3):121126.
Dumigan DG, Kohan CA, Reed CR, Jekel JF, Fikrig MK. Utilizing
national nosocomial infection surveillance system data to improve urinary tract infection rates in three intensive-care units.
Clin Perform Qual Health Care 1998;6(4):172178.
Dellinger EP, Hausmann SM, Bratzler DW, et al. Hospitals collaborate to decrease surgical site infections. Am J Surg 2005;
190(1):915.
Lutarewych M, Morgan SP, Hall MM. Improving outcomes of
coronary artery bypass graft infections with multiple interventions: putting science and data to the test. Infect Control Hosp
Epidemiol 2004;25(6):517519.
Rao N, Schilling D, Rice J, Ridenour M, Mook W, Santa E.

41.

42.

43.

44.

45.

46.

47.

Prevention of postoperative mediastinitis: a clinical process improvement model. J Healthc Qual 2004;26(1):2227.
Cocanour CS, Ostrosky-Zeichner L, Peninger M, et al. Cost of
a ventilator-associated pneumonia in a shock trauma intensive
care unit. Surg Infect (Larchmt) 2005;6(1):6572.
Rello J, Ollendorf DA, Oster G, et al. Epidemiology and outcomes of ventilator-associated pneumonia in a large US database. Chest 2002;122(6):21152121.
Herwaldt LA, Cullen JJ, Scholz D, et al. A prospective study of
outcomes, healthcare resource utilization, and costs associated
with postoperative nosocomial infections. Infect Control Hosp
Epidemiol 2006;27(12):12911298.
Kirkland KB, Briggs JP, Trivette SL, Wilkinson WE, Sexton DJ.
The impact of surgical-site infections in the 1990s: attributable
mortality, excess length of hospitalization, and extra costs. Infect
Control Hosp Epidemiol 1999;20(11):725730.
Perencevich EN, Sands KE, Cosgrove SE, Guadagnoli E, Meara
E, Platt R. Health and economic impact of surgical site infections
diagnosed after hospital discharge. Emerg Infect Dis 2003;9(2):
196203.
Dimick JB, Pronovost PJ, Cowan JA, Lipsett PA. Complications
and costs after high-risk surgery: where should we focus quality
improvement initiatives? J Am Coll Surg 2003;196(5):671678.
Herwaldt LA, Swartzendruber SK, Edmond MB, et al. The epidemiology of hemorrhage related to cardiothoracic operations.
Infect Control Hosp Epidemiol 1998;19(1):916.