Asssigment

Effectiveness and Limitations of Hand Hygiene Promotion on Decreasing Healthcare–Associated Infections Yee-Chun Chen1,2,3*, Wang-Huei Sheng1,2, Jann-Tay Wang1,2, Shan-Chwen Chang2,3, Hui-Chi Lin2, Kuei-

Lien Tien2, Le-Yin Hsu2, Keh-Sung Tsai1,3

1 Center for Infection Control, National Taiwan University Hospital, Taipei, Taiwan, 2 Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan,

3 Department of Medicine, National Taiwan University College of Medicine, Taipei, Taiwan

Abstract

Background: Limited data describe the sustained impact of hand hygiene programs (HHPs) implemented in teaching hospitals, where the burden of healthcare-associated infections (HAIs) is high. We use a quasi-experimental, before and after, study design with prospective hospital-wide surveillance of HAIs to assess the cost effectiveness of HHPs.

Methods and Findings: A 4-year hospital-wide HHP, with particular emphasis on using an alcohol-based hand rub, was implemented in April 2004 at a 2,200-bed teaching hospital in Taiwan. Compliance was measured by direct observation and the use of hand rub products. Poisson regression analyses were employed to evaluate the densities and trends of HAIs during the preintervention (January 1999 to March 2004) and intervention (April 2004 to December 2007) periods. The economic impact was estimated based on a case-control study in Taiwan. We observed 8,420 opportunities for hand hygiene during the study period. Compliance improved from 43.3% in April 2004 to 95.6% in 2007 (p,.001), and was closely correlated with increased consumption of the alcohol-based hand rub (r = 0.9399). The disease severity score (Charlson comorbidity index) increased (p = .002) during the intervention period. Nevertheless, we observed an 8.9% decrease in HAIs and a decline in the occurrence of bloodstream, methicillin-resistant Staphylococcus aureus, extensively drug-resistant Acinetobacter baumannii, and intensive care unit infections. The intervention had no discernable impact on HAI rates in the hematology/oncology wards. The net benefit of the HHP was US$5,289,364, and the benefit-cost ratio was 23.7 with a 3% discount rate.

Conclusions: Implementation of a HHP reduces preventable HAIs and is cost effective.

Citation: Chen Y-C, Sheng W-H, Wang J-T, Chang S-C, Lin H-C, et al. (2011) Effectiveness and Limitations of Hand Hygiene Promotion on Decreasing Healthcare– Associated Infections. PLoS ONE 6(11): e27163. doi:10.1371/journal.pone.0027163

Editor: Adam J. Ratner, Columbia University, United States of America

Received February 9, 2011; Accepted October 11, 2011; Published November 16, 2011

Copyright: � 2011 Chen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: Dr. Chen received grants (DOH96-DC-1010, DOH97-DC-1005) from the Center for Disease Control, Department of Health, and a grand (DOH99-TD-B- 111-001) from the Department of Health, Taiwan. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing Interests: None of the authors declared a conflict of interest.

* E-mail: [email protected]

Introduction

Healthcare-associated infections (HAIs) cause significant con-

cern regarding the safety and quality of healthcare quality

worldwide [1,2]. The World Health Organization launched the

World Alliance for Patient Safety in October 2004 [2]. HAIs have

been identified as a fundamental priority, and were selected as the

topic of the first Global Patient Safety Challenge. Hand hygiene

was identified as the core component of this strategy because it is a

simple, standardized, low-cost measure based on solid scientific

evidence. The major barrier is poor compliance by healthcare

providers, regardless of available resources [1–3].

Following the SARS epidemic in 2003, we reviewed the

strengths and weaknesses of infection control strategies at the

National Taiwan University Hospital in anticipation of other

infectious diseases emerging [4,5]. We decided to focus first on

compliance with hand hygiene. An unannounced hand hygiene

audit by infection control nurses was conducted of hospital wards

in December 2003. We found that of the 226 opportunities

presented, hospital staff washed their hands for only 16.6%. This

finding led us to introduce alcohol-based hand rubs, and

implement a hospital-wide program promoting hand hygiene

through using alcohol-based hand rubs.

Facing an increase in overall HAIs and infections caused by

multidrug-resistant organisms [6], the goal of this initiative was to

decrease HAIs by blocking the transmission of microorganisms via

the hands of healthcare workers (HCW). We used a quasi-

experimental, before and after, study design [7] with prospective

hospital-wide surveillance of HAIs to assess the effectiveness of

hand hygiene. We found that the sustained improvement of hand

hygiene compliance reduces HAIs and is cost effective.

Methods

Hospital and study population National Taiwan University Hospital (NTUH) is a 2,200-bed

major teaching hospital in Taiwan that provides both primary and

PLoS ONE | www.plosone.org 1 November 2011 | Volume 6 | Issue 11 | e27163

tertiary medical care. Approximately one third of the hospital’s

house staff on the floor are replaced each year. The distribution

and time trends of HAIs and infection control programs during

1981 to 2007 have been described previously [6]. An antimicrobial

stewardship program was not conducted during the study period.

Baseline evaluation We conducted a baseline evaluation in December 2003 [8]. At

that time, hands-free washing facilities with unmedicated liquid

soap (not refilled) and paper towels were located in every room of

the wards and by every intensive care unit (ICU) bed. However,

alcohol-based hand rubs were unavailable. The baseline evalua-

tion included direct observation of hand hygiene compliance by

infection control nurses, a survey of knowledge and the reasons or

factors affecting hand hygiene adherence using a structured, self-

administered questionnaire, and suggestions to improve hand

hygiene performance. Accordingly, we introduced accessible

alcohol-based hand rubs to improve compliance.

Hospital-wide hand hygiene program With substantial support from the hospital superintendent, the

hospital-wide hand hygiene promotion program began in April

2004. A multidisciplinary approach involving cognition, equip-

ment, and behavior was designed based on literature, the baseline

evaluation, and the concept of total quality improvement [9].

Disposable alcohol-based hand rub sanitizers (75% isopropyl

alcohol in plastic hand-compressing dispensers) (So Easy liquid,

PBF, Taiwan) were wall mounted between every two beds in

general wards, by each bed in special units (such as intensive care

units), and affixed to trolleys (including treatment trolleys and

resuscitation trolleys) to ensure accessibility near or at the point of

care [2].

Promotion consisted of lectures and/or web-based self-learning

with a post test, reminders located near points of care, use of hand

hygiene compliance as a quality indicator, observation and verbal

reminders by infection control nurses, periodic audits and

performance feedback provided to units and departments, and

incentives of US$160.00 for an outstanding performance (unit and

department level). A fine of US$3.00 for compliance failures

(individual level) was implemented in 2007 in highly specific

situations, that is, individuals not modify their behavior even after

face-to-face communication. The hand hygiene program was

announced and promoted hospital-wide for one month every year

while preparing for pandemic influenza (typically during Novem-

ber). The program was reviewed, revised, and promoted annually

according to the plan-do-check-act cycle [8]. The target hand

hygiene compliance rate was determined based on the perfor-

mance in the preceding year.

Determination of hand hygiene compliance Hand hygiene compliance was assessed by direct observation

during day shifts (8 a.m. to 5 p.m.) on weekdays according to the

U.S.A. CDC criteria [10], using a standardized case report form.

Emergency procedures were excluded. Direct observation was

performed by infection control nurses (ICN) after training and

consensus development [11]. The site audit period was an-

nounced, though healthcare workers were not informed of the

specific observation time. The hand hygiene compliance rates

recorded for comparison between hospital services or between

years were limited to the opportunities before and after patient

contact to maintain consistency during the study period. Six

surveys were conducted during the implementation period. To

avoid the Hawthorne effect [12], the duration and number of

opportunities for hand hygiene monitoring were increased

gradually for the six surveys (Fig. 1A). The annual consumption

of liquid soap, antiseptics, and alcohol hand rub (product volume

use per 1,000 patient-days) was used as a surrogate marker of hand

hygiene over time.

Outcome assessment and data collection We measured the overall HAI rates, HAI rates by site of

infection, and HAI rates by selected pathogen to assess the

effectiveness of HHP. We chose methicillin-resistant Staphylococcus

aureus (MRSA), Acinetobacter, and extensively drug-resistant A.

baumannii (XDRAB) as marker organisms because our previous

study showed a significant increase in HAI incidences, and

outbreaks of these infections occurred during the study period [6].

Conversely, the incidences of HAIs by methicillin-susceptible S.

aureus (MSSA) and Escherichia coli, which were chosen as control

organisms, were relatively stable or decreased.

Prospective, hospital-wide on-site surveillance of HAIs at

NTUH began in 1981, and were conducted through weekly visits

from full-time ICN to all patient units [6]. Data were collected on

standardized data collection forms and inputted into the computer

Figure 1. Trends in compliance with the hand hygiene during 6 consecutive hospital-wide surveys conducted from May 2004 to December 2007. Panel A shows significant increases in percent of adherence to hand hygiene before and after patient contact for all healthcare workers in in-patient service units (p,0.001), by doctors (p,0.001), nurses (p,0.001), and other healthcare workers (p,0.001). Panel B shows significant increases in the annual consumption of alcohol-based hand rub (p = 0.001), antiseptics (p = 0.04), and liquid soap (p = 0.03). doi:10.1371/journal.pone.0027163.g001

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 2 November 2011 | Volume 6 | Issue 11 | e27163

database manually. The severity of underlying diseases, compris-

ing a maximum of six diseases, was scored using the Charlson

comorbidity index [13]. The patient population, bed occupancy

rate, age, gender, severity of underlying diseases, length of hospital

stay, cumulative incidence of HAI, and outcome at discharge were

determined during the preintervention (January 1999 to March

2004) and intervention (April 2004 to December 2007) periods.

Definitions HAIs were classified according to definitions provided by the

Centers for Disease Control and Prevention, U.S.A. [6,14]. The

cumulative incidence of HAIs was defined as episodes per 1,000

patient-days. Extensively drug-resistant A. baumannii (XDRAB) was

defined as isolates that were resistant to five or more classes of

antibacterial agents [6]. Compliance with hand hygiene is the ratio

of the number of performed actions using correct technique to the

number of opportunities [2,11].

Statistical evaluation To investigate the impact of the intervention on the levels and

trends of cumulative HAI incidence over time, we adopted the

generalized autoregressive Poisson regression analyses [15,16].

This model autocorrelates the HAI cases in successive periods by

incorporating the autoregressive order j term and number of HAI

cases at an earlier time t-j [17]. The factors considered by the

model (see Text S1) include the levels and trends in the

preintervention period, the changes in levels and trends during

the intervention period, and autoregressive terms. Because the

SARS epidemic occurred during the later part of the preinterven-

tion period (April 2003 to July 2003), an additional variable was

added to the model to assess the effect of SARS. To account for

possible seasonal and other event variations, such as the lengthy

Chinese New Year holiday during January and February, the ‘‘old

guy’’ effect during April and May, and the new staff effect during

June to July, the model defined spring as March to May, summer

as June to August, autumn as September to November, and winter

as December to February. The analysis was further stratified by

the site of infection, such as bloodstream, urinary tract, and

surgical site infections, by the pathogens, such as MSSA, MRSA,

Acinetobacter, XDRAB, and E. coli, and by ward units, such as ICUs,

oncology wards, and hematology wards. Heterogeneity factors

were calculated by dividing the deviance with the degrees of

freedom, and used as an indicator to assess whether an extra-

Poisson variation (overdispersion) was present.

The difference in patient population during the two periods was

examined using Student’s t-test (continuous variables) and a chi-

Table 1. Characteristics of the Patient Population.

Parameter Preintervention period (Jan 1999–Mar 2004)

Intervention period (April 2004–Dec 2007)

Number of acute care beds 2,0276180 2,20268.6

Occupancy rate, % 79.766.7 84.160.9a

Total patient-days 648,305632,109 720,56469,543b

Length of hospital stay (days) 9.8060.62 9.7960.65

Number of patients 290,056 262,090

Age (in years) 46.4623.1 48.3623.3a

Gender, % male 50.461.2 49.261.3

Charlson comorbidity index 2.0760.27 2.8160.22c

Underlying diseases, %

Cardiac vascular diseases 16.663.2 22.960.3a

Congestive heart failure 1.860.3 2.460.1a

Cerebrovascular diseases 4.460.8 4.860.3a

Chronic pulmonary diseases 3.560.8 4.160.3a

Moderately-severe liver diseases 5.960.8 3.860.1a

Moderately-severe renal diseases 4.860.8 3.660.5a

Diabetes mellitus without end organ damage 7.461.5 9.860.2a

Diabetes mellitus with end organ damage 1.560.2 1.660.1

Any tumor 22.061.7 31.763.3a

Leukemia 0.3760.06 0.4760.03a

Lymphoma 1.1060.25 1.6360.16a

Metastatic solid tumor 7.5361.16 12.8062.01a

Acquired immunodeficiency syndrome 0.3760.07 0.4060.02

Solid organ transplant (kidney, liver, heart, or pancreas) 0.5960.06 0.9360.14a

Hematopoietic stem cell transplant 0.0560.01 0.0560.02

In-hospital mortality, % 3.6760.19 3.7760.09d

aP,0.001. bP = 0.004. cP = 0.002. dP = 0.04. doi:10.1371/journal.pone.0027163.t001

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 3 November 2011 | Volume 6 | Issue 11 | e27163

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 4 November 2011 | Volume 6 | Issue 11 | e27163

squared test (categorical variables). Time trends in the consump-

tion of alcohol-based hand rubs, soap, and antiseptics were

examined using simple linear regression. To investigate improve-

ments to hand hygiene compliance during the six consecutive

hospital-wide surveys, Pearson’s chi-squared test for trend was

performed. Correlation between hand hygiene compliance and the

consumption of alcohol-based hand rubs, soap, and antiseptics,

was examined using the Spearman method. All statistical tests

were considered two-tailed and were significant at p,.05.

Economic evaluation The relevant parameters used in the economic evaluation are

shown in Table S2. The cost-effectiveness of the program was

evaluated by calculating the extra cost required to prevent one

episode of HAI from a hospital perspective. The number of

expected episodes of HAI averted by the program was derived

using the generalized autoregressive Poisson regression model.

The costs considered in the analysis included alcohol-based hand

hygiene products and promotional efforts (posters, wall displays,

rewards, and other expenses) over the 4-year intervention period.

The cost of personnel involved in the program (including

planning, training, and auditing) was not considered in the base-

case analysis (Table S2), because the program did not incur any

increases in staff or manpower costs. However, the opportunity

costs of personnel were considered in the sensitivity analysis by

converting the number of working hours that infection control

nurses spent on the program to their salary (Table S3). Cost-

benefit analyses were also conducted to examine the net benefit

and benefit-cost ratio of the program. The benefit was measured

by subtracting the cost savings from the extra costs caused by

HAIs, which was determined by a case-control study conducted in

our hospital [18]. All future costs, number of HAI episodes, and

benefits were discounted to the present value at an annual rate of

3% [19,20]. One-way sensitivity analyses were performed to

explore the influence of the uncertainty of several parameters, such

as discount rates, cost of alcohol hand rub, campaign expenses,

extra cost per HAI episode, and the number of averted HAIs. The

ranges used for the sensitivity analysis were a 50% increase or

decrease to the base-case estimates of alcohol hand rub costs and

campaign expenses, the twenty-fifth and seventy-fifth percentile of

the estimated additional costs of each HAI episode, and a 95%

confidence interval for the model-predicted number of averted

HAIs.

Ethics statement We followed the principles expressed in the Declaration of

Helsinki. This study was approved by the Institutional Ethics

Review Board of the National Taiwan University Hospital

(No. NTUH-200805033R).

Results

Patient population The patient population characteristics during the preinterven-

tion and intervention periods are summarized in Table 1. During

the intervention period, the total patient days (p = .004), mean

Charlson comorbidity index (p = .002), and frequency of underly-

ing illnesses, such as hematological malignancies and solid tumors

(p,.001), increased significantly.

Hand hygiene compliance During the six surveys conducted during the intervention

period, 8,420 opportunities to observe hand hygiene were

presented, as shown in Fig. 1A. Overall compliance improved

from 43.3% in 2004 to 95.6% in 2007 (p,.001). Compliance

improved significantly for all professional categories of HCW

(p,.001), in both general wards and intensive care units (p,.001).

We also observed an increased use of alcohol-based hand rubs

(p = .001), liquid soap (p = .03), and antiseptics (p = 0.04), as shown

in Fig. 1B. The overall improvement in hand hygiene compliance

was significantly correlated with the increased consumption of

alcohol-based hand rub (correlation coefficient r = 0.9399,

p = .005), but less correlated with the consumption of antiseptics

(r = 0.7930, p = .06) and soap (r = 0.7686, p = .07).

Outcome assessment The predicted monthly cumulative incidence of HAI using the

full and most parsimonious segmented regression models are

shown in Table S1. Before intervention, HAI increased gradually,

and the impact of SARS and seasons were significant. After

implementation of the hand hygiene program, the levels and

trends of HAIs changed significantly (p = .02, p = .04, respectively),

as shown in Fig. 2A. Bloodstream (p,.001), urinary tract (trend,

p = .03), and skin and soft tissue infections (trend, p,.001)

decreased significantly. Additionally, though surgical site infections

showed a downward trend, the incidence rate did not differ

significantly. No significant changes in the respiratory tract and

gastrointestinal tract infection rates were observed during the

intervention. The time trends of HAI significantly decreased in

ICUs (p,.001) during the intervention period, as shown in Fig. 2B;

however, no significant changes in the HAI rates of oncology and

hematological wards was noted, as shown in Fig. 2C.

The cumulative incidence of HAIs caused by MRSA,

Acinetobacter, and XDRAB decreased substantially during the

intervention period, as shown in Figs. 3A and 3B. This was

associated with a decrease in the annual consumption of

glycopeptides, anti-Pseudomonas fluoroquinolones, and carbape-

nems (data not shown). Finally, no significant change in the trends

or levels of infections caused by MSSA and E. coli during the

intervention period was observed, as shown in Fig. 3C.

Economic evaluation The hand hygiene program was associated with an estimated

reduction of 1,504 (95% confidence interval: 526 to 2,544) HAI

episodes (8.9%) during the intervention period without discounting

(Table 2). Results of the economic analysis with a 3% discount rate

are shown in Table 3. From a hospital perspective, the discounted

additional cost of preventing one HAI episode was US$163.6.

However, the net benefit of the hand hygiene program is

US$5,289,364, which indicates that the benefits from savings to

the additional costs of HAI could outweigh the program costs. The

benefit-cost ratio shows that every US$1 spent on the program

could result in a US$23.7 benefit. The results of one-way

sensitivity analysis (Table S3) show that the cost of alcohol hand

Figure 2. Time trends of monthly cumulative incidences of overall healthcare-associated infection before (January 1999 to March 2004) and during the hand hygiene program (April 2004 through December 2007). (A) hospital-wide (change in levels, p = 0.02; change in trends, p = 0.04); (B) intensive care units (change in levels, p = 0.26; change in trends, p,0.001); (C) hematology ward (p = 0.21, p = 0.38, respectively). Observed incidences, black solid line, —; mean, red dash line, ---; green and blue shadow, 95% confidence interval of observed incidences; yellow shadow, 95% confidence interval (CI) of predicted incidences. The vertical dashed lines (- - -) separate the preintervention and intervention periods. doi:10.1371/journal.pone.0027163.g002

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 5 November 2011 | Volume 6 | Issue 11 | e27163

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 6 November 2011 | Volume 6 | Issue 11 | e27163

rub and the number of averted HAI episodes are the two factors

influencing the additional cost of preventing one HAI episode;

extra cost per HAI episode and the number of averted HAI

episodes significantly impact the net benefit and benefit-cost ratio.

Discussion

This study demonstrates that excellent compliance with hand

hygiene by HCW was achieved and maintained over 4 years

through an intensive hospital-wide program. The program was

associated with significant decreasing trends for all HAIs and HAIs

caused by MRSA and XDRAB, and was cost effective. However,

we were unable to demonstrate a reduction in the HAIs of

hematology and oncology services. We were also unable to

demonstrate a reduction in HAIs caused by MSSA and E. coli.

This result was unsurprising because of the relatively greater

importance of endogenous infections in immune and structurally

compromised hosts.

To provide consistent care and protect patient safety throughout

the hospital, a hand hygiene program requires multidisciplinary

efforts and encompasses all hospital units. However, conducting a

randomized, controlled trial in such a large and complex situation

is difficult. According to a recent, comprehensive review of

literature published as a Cochrane Review [21], the quality of

intervention studies intended to increase hand hygiene compliance

remains disappointing. We chose to superimpose the hand hygiene

program on a well-established infection control program using

standardized surveillance methods. This enabled us to determine

the differences in HAI incidences over 4-year preintervention and

postintervention periods, and conduct numerous observations

while adjusting to changes in the frequency and severity of

underlying diseases.

We elected to use Poisson regression analysis with an

interrupted time series [22,23] to assess the efficacy of the

hospital-wide hand hygiene intervention over time, and determine

whether factors other than the intervention could explain the

change. In the interrupted time series, the level and trend of the

preintervention segment served as the controls for the postinter-

vention segment, providing a methodologically acceptable design

for measuring the intervention effect [22,23]. This method

requires data of the continuous or counted outcome measures,

summarized at regular, evenly spaced intervals. Thus, evaluation

of the longitudinal effect of a hospital-wide hand hygiene program

is only feasible in limited healthcare settings.

Although evidence exists that improved adherence to hand

hygiene is linked to reduced infection rates [24–26], other studies

failed to report these effects [27–30]. In addition, the impact of

hand hygiene improvement on HAI incidence varied [31,32].

Several potentially confounding factors are relevant to this study.

First, although direct observation is the criterion standard for

measuring HH compliance, the method is subject to observation

bias, selection bias, and the Hawthorne effect [2,12], which may

result in an overestimated HH rate. Second, the HH compliance

rate was limited to the opportunities before and after patient

Figure 3. Time trends of monthly cumulative incidences by pathogen. (A) methicillin-resistant S. aureus (MRSA) (change in level, p = 0.03; change in trend, p = 0.04); (B) extensively drug-resistant Acinetobacter (XDRAB) (p = 0.78; p,0.001, respectively); (C) Escherichia coli (p = 0.89; p = 0.33, respectively). Infection control measures for XDRAB were intensified during June 2001 to June 2002. These efforts resulted in only a transient reduction in the rates of infection for XDRAB and MRSA. Observed incidences, black solid line, —; mean, red dash line, ---; green and blue shadow, 95% confidence interval of observed incidences; yellow shadow, 95% confidence interval (CI) of predicted incidences. The vertical dashed lines (- - -) separate the preintervention and intervention periods. doi:10.1371/journal.pone.0027163.g003

Table 2. Decrease in Healthcare-associated Infections and Cost Savings Attributed to the Hand Hygiene Program.

Parameter Number

Episodes of healthcare-associated infections

Observed 15,301

Predicteda 16,805

Total reduction (%) 1,504 (8.9%)

Costs of the hand hygiene program in US dollars

Total alcohol handrub expense 221,517b

Campaign costs (posters, salaries,c etc.) 22,953

Total costd 244,470

Average cost per 1000 patient-day 90.60

Average cost to prevent one episode of HAI 162.50

Extra costs per episode of HAIe 5,335613,872

Mean cost reduction for HAIs 8,023,840

Mean net cost savings from the hand hygiene program 7,779,370

aPredicted by Poisson regression model described in Methods and Text S1. bAll monetary values are expressed in US dollars. The average exchange rate in

2007: 1 US dollar = 32.842 New Taiwan dollars, 1 Euro = 44.952 New Taiwan dollars.

cThe salaries were limited to the proportion of working hours of infection control nurses spent for the hand hygiene audits, not include those of infection control nurses in planning and education and not include those of infectious disease specialists.

dVariable costs including soap, water and materials used for drying hands (e.g., towels) were not included.8

eBased on a case-control study conducted at this hospital and two local hospitals in Taiwan.18

doi:10.1371/journal.pone.0027163.t002

Table 3. Results of Cost-effectiveness Analysis and Cost- benefit Analysis of the Hand Hygiene Programa.

Parameter Value

Episodes of healthcare-associated infectionsb

With hand hygiene program 14,608

Without hand hygiene program 16,032

Number of averted episodes of HAI 1,424

Saving from extra costs arose from episodes of HAI 5,522,408c

Extra costs of the hand hygiene program 233,044

Extra cost to prevent one episode of HAI 163.6

Net benefitd 5,289,364

Benefit cost ratioe 23.7

aAll future costs, number of episodes of HAI, and benefits were discounted to the present value at an annual rate of 3%.

bPredicted by Poisson regression model described in Methods and Text S1. cAll monetary values are expressed in US dollars. The average exchange rate in 2007: 1 US dollar = 32.842 New Taiwan dollars, 1 Euro = 44.952 New Taiwan dollars.

dNet benefit = Benefit – Cost. eBenefit cost ratio = benefit/cost. doi:10.1371/journal.pone.0027163.t003

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 7 November 2011 | Volume 6 | Issue 11 | e27163

contact to maintain consistency during the study period. Our

recent survey demonstrated that the HH rate for 5 moments was

only approximately 60%; whereas the HH rate for two moments

exceeded 90% (unpublished data). Third, the hand hygiene

programs in this study focused only on HCWs. Patients, patients’

family, and other caregivers may contaminate the environment

and/or transfer organisms. Fourth, not all HAIs of exogenous

origin can be prevented by hand hygiene [33,34]. Augmenting

other infection control measures, such as multimodal implemen-

tation strategies (bundle care), environment cleanliness, appropri-

ate use of antimicrobial agents, and active microbial surveillance of

multidrug-resistant organisms is also necessary [33–38]. Finally,

not all HAI were preventable. Our data failed to support the

effects of HHP on decreasing HAIs among hemato-oncology

patients and patients with infections caused by MSSA and E. coli.

Endogenous infection may be reduced through improvements in

host immunity.

The economic impact of effective hand hygiene programs on

decreasing HAIs was first evaluated by Pittet et al. [2,24]. Pittet

and colleagues [24] estimated the program costs to be less than

US$57,000 per year for a 2,600-bed hospital, an average of

US$1.42 per admitted patient. Supplementary costs associated

with the increased use of alcohol-based hand rub averaged

US$6.07 per 100 patient-days. Based on conservative estimates,

US$100 was saved per averted infection. The economic evaluation

in this study produced similar results. In addition, this study

showed that the cost of alcohol hand rub and the number of

averted HAI episodes are the two factors influencing the additional

costs of preventing one HAI episode, which subsequently

influences the net benefit and benefit-cost ratio. However, in the

model used in this study, the cost of manpower was underesti-

mated, and the indirect costs related to HAIs were not included.

Several studies have demonstrated a temporal relationship

between improved hand hygiene practices and a reduction in

HAI incidence and multidrug resistant microorganisms [2,25];

however, few have been able to sustain a lasting impact [2,26].

Demonstrating to hospital administrators that these programs are

cost-effective, relatively inexpensive, improve patient safety, and

require long-term and stable investment is essential [2,24]. This is

particularly required by teaching hospitals where the house staff

and attending physicians change periodically [25]. Interest in pay-

for-performance and other funding schemes is growing, which

should further strengthen financial incentives to foster hand

hygiene [39–41].

In conclusion, this study demonstrates that implementing a

hospital-wide hand hygiene program is feasible. The program was

associated with a reduction in the HAIs of most hospital units and

HAIs caused by MRSA and XDRAB. This effect was achieved

and confirmed using a before and after study design combined

with a prospective HAI surveillance program and hospital-wide

annual promotion, which resulted in sustained effects and high-

quality HH observation. The costs of implementing the program

were low compared to the costs saved by reducing the HAI

incidence. The impact may be even greater with full adherence to

the five moments for hand hygiene. However, hand hygiene

programs are only one component of hospital infection control.

They must be supplemented with measures directed at device-

associated and endogenous infections in compromised hosts, and

augmented by antimicrobial stewardship.

Supporting Information

Text S1 The generalized autoregressive Poisson model.

(DOC)

Table S1 Parameter estimates, standard errors and p values from the full and most parsimonious segmented regression models predicting monthly incidence density (episodes per 1000 patient-days) over time.

(DOC)

Table S2 Base-case estimates and ranges used in sensitivity analyses of parameters.

(DOC)

Table S3 Results of one-way sensitivity analysis.

(DOC)

Acknowledgments

We are grateful to members of the NTUH Center for Infection Control

and Center for Quality Control for their support of the hand hygiene

program and the hospital staff for their commitment to improve patient

safety and reduce healthcare-associated infections. We appreciate members

of the Biostatistics Laboratory of the College of Public Health, National

Taiwan University, and Dr. Grace Hui-Min Wu for their assistance and

suggestions regarding the statistical analysis; they received no compensa-

tion for their contribution. Additionally, the authors are grateful to Prof.

Calvin Kunin and Prof. Wei-Chuan Hsieh for their suggestions and critical

review of the manuscript.

Author Contributions

Conceived and designed the experiments: YCC WHS JTW HCL.

Performed the experiments: KLT. Analyzed the data: YCC WHS LYH.

Contributed reagents/materials/analysis tools: LYH. Wrote the paper:

YCC WHS JTW. Review and comments of the manuscript: SCC KST.

References

1. Burke JP (2003) Infection Control - A problem for patient safety. N Engl J Med

348: 651–656.

2. World Health Organization (2009) WHO guidelines on hand hygiene in health

care. First global patient safety challenge: clean care is safe care. http://

whqlibdoc.who.int/publications/2009/9789241597906_eng.pdf. Accessed May

5, 2010.

3. Saint S, Howell JD, Krein SL (2010) Implementation Science: How to Jump-

Start Infection Prevention. Infect Control Hosp Epidemiol 31: S14–S17.

4. Chen YC, Chen PJ, Chang SC, Kao CL, Wang SH, et al. (2004) Infection

control and SARS transmission among healthcare workers, Taiwan. Emerg

Infect Dis 10: 895–898.

5. Chen YC, Huang LM, Chan CC, Su CP, Chang SC, et al. (2004) SARS in

hospital emergency room. Emerg Infect Dis 10: 782–788.

6. Chuang YC, Chen YC, Chang SC, Sun CC, Chang YY, et al. (2010) Secular

trends of healthcare-associated infections at a teaching hospital in Taiwan,

1981–2007. J Hosp Infect 76: 143–149.

7. Shardell M, Harris AD, El Kamary SS, Furuno JP, Miller RR, et al. (2007)

Statistical analysis and application of quasi experiments to antimicrobial

resistance intervention studies. Clin Infect Dis 45: 901–907.

8. Lin HC, Tien KL, Sun CC, Wang SH, Chen YC, et al. (2010) Promotion

and achievement of a hospital-wide hand hygiene program implemented

during 2004–2007 at a teaching hospital in Taiwan. J Infect Control 20:

146–162.

9. Shortell SM, O’Brien JL, Carman JM, Foster RW, Hughes EFX, et al. (1995)

Assessing the impact of continuous quality improvement/total quality manage-

ment: concept versus implementation. Health Services Research 30: 377–401.

10. Boyce JM, Pittet D (2002) Guideline for hand hygiene in health-care settings.

Recommendations of the Healthcare Infection Control Practices Advisory

Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force.

Society for Healthcare Epidemiology of America/Association for Professionals in

Infection Control/Infectious Diseases Society of America. MMWR Recomm

Rep 51(RR-16): 1–45.

11. Sax H, Allegranzi B, Uçkay I, Larson E, Boyce J, et al. (2007) ‘My five moments

for hand hygiene’: a user-centred design approach to understand, train, monitor

and report hand hygiene. J Hosp Infect 67: 9–21.

12. Roethlisberger FJ, Dickson WJ (1939) Management and the worker: an account

of a research program conducted by the Western Electric Company, Hawthorne

Works, Chicago. Cambridge, MA: Harvard University Press.

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 8 November 2011 | Volume 6 | Issue 11 | e27163

13. Charlson ME, Pompei P, Ales KL, Mackenzie CR (1987) A new method of

classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40: 373–383.

14. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM (1988) CDC

definitions for nosocomial infections, 1988. Am J Infect Control 16: 128–140. 15. Zeger SL (1988) A regression model for time series of counts. Biometrika 75:

621–629. 16. Hsu SM, Yen AMF, Chen THH (2008) The impact of climate on Japanese

encephalitis. Epidemiol infect 136: 980–987.

17. Chatfield C (1989) The analysis of time series. An introduction. 4th ed. London, England: Chapman and Hall.

18. Sheng WH, Wang JT, Lu DCT, Chie WC, Chen YC, et al. (2005) Comparative impact of hospital-acquired infections on medical costs, length of hospital stay

and outcome between community hospitals and medical centers. J Hosp Infect 59: 205–214.

19. Drummond MF, Sculpher MJ, Torrance GW, O’Brien BJ, Stoddart GL (2005)

Methods for the Economic Evaluation of Health Care Programmes. 3rd ed. Oxford, UK: Oxford University Press.

20. Petrou S, Gray A (2011) Economic evaluation alongside randomised controlled trials: design, conduct, analysis, and reporting. BMJ 342: d1548. doi: 10.1136/

bmj.d1548.

21. Gould DJ, Moralejo D, Drey N, Chudleigh JH (2010) Interventions to improve hand hygiene compliance in patient care. Cochrane Database of Systematic

Reviews Issue 9. Art. No.: CD005186. DOI: 10.1002/14651858.CD005186.pub3. 22. Gillings D, Makuc D, Siegel E (1981) Analysis of interrupted time series

mortality trends: an example to evaluate regionalized perinatal care. Am J Public Health 71: 38–46.

23. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D (2002) Segmented

regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther 27: 299–309.

24. Pittet D, Sax H, Hugonnet S, Harbarth S (2004) Cost implications of successful hand hygiene promotion. Infect Control Hosp Epidemiol 25: 264–266.

25. Pittet D, Hugonnet S, Harbarth S, Mourouga P, Sauvan V, et al. (2000)

Effectiveness of a hospital-wide programme to improve compliance with hand hygiene. Lancet 356: 1307–1312.

26. Trick WE, Vernon MO, Welbel SF, Demarais P, Hayden MK, et al. (2007) Chicago Antimicrobial Resistance Project. Multicenter intervention program to

increase adherence to hand hygiene recommendations and glove use and to reduce the incidence of antimicrobial resistance. Infect Control Hosp Epidemiol

28: 42–49.

27. Simmons B, Bryand J, Neiman K, Spencer L, Arheart K (1990) The role of handwashing in prevention of endemic intensive care unit infections. Infect

Control Hosp Epidemiol 11: 589–594. 28. Rupp ME, Fitzgerald T, Puumala S, Anderson JR, Craig R, et al. (2008)

Prospective, controlled, cross-over trial of alcohol-based hand gel in critical care

units. Infect Control Hosp Epidemiol 29: 8–15.

29. Capretti MG, Sandri F, Tridapalli E, Galletti S, Petracci E, et al. (2008) Impact

of a standardized hand hygiene program on the incidence of nosocomial

infection in very low birth weight infants. Am J Infect Control 36: 430–435.

30. Mertz D, Dafoe N, Walter SD, Brazil K, Loeb M (2010) Effect of a multifaceted

intervention on adherence to hand hygiene among healthcare workers: a cluster-

randomized trial. Infect Control Hosp Epidemiol 31: 1170–1176.

31. Pittet D, Allegranzi B, Sax H, Dharan S, Pessoa-Silva CL, et al. (2006) Evidence-

based model for hand transmission during patient care and the role of improved

practices. Lancet Infect Dis 6: 641–652.

32. Silvestri L, Petros AJ, Sarginson RE, de la Cal MA, Murray AE, et al. (2005)

Handwashing in the intensive care unit: a big measure with modest effects.

J Hosp Infect 59: 172–179.

33. Dancer SJ (2010) Control of transmission of infection in hospitals requires more

than clean hands. Infect Control Hosp Epidemiol 31: 958–960.

34. Grundmann H, Barwolff S, Tami A, Behnke M, Schwab F, et al. (2005) How

many infections are caused by patient-to-patient transmission intensive care

units? Crit Care Med 33: 946–951.

35. Siegel JD, Rhinehart E, Jackson M, Chiarello L, the Healthcare Infection

Control Practices Advisory Committee (2007) Guideline for isolation precau-

tions: preventing transmission of infectious agents in healthcare settings.

Available: http://www.cdc.gov/hicpac/2007IP/2007isolationPrecautions.html

Accessed Jun 5, 2009.

36. Siegel JD, Rhinehart E, Jackson M, Chiarello L, the Healthcare Infection

Control Practices Advisory Committee (2006) Management of multidrug-

resistant organisms in healthcare settings, 2006. Avaiable: http://www.cdc.gov/

hicpac/mdro/mdro_0.html Accessed Jun 5, 2009.

37. Jarlier V, Trystram D, Brun-Buisson C, Fournier S, Carbonne A, et al. (2010)

Curbing Methicillin-Resistant Staphylococcus aureus in 38 French Hospitals

Through a 15-Year Institutional Control Program. Arch Intern Med 170:

552–559.

38. Chuang YC, Wang JT, Chen ML, Chen YC (2010) Comparison of an

automated repetitive-sequence-based PCR microbial typing system with pulsed-

field gel electrophoresis for molecular typing of vancomycin-resistant Enterococcus

faecium. J Clin Microbiol 48: 2897–2901.

39. Wenzel RP (1985) Nosocomial infections, diagnosis-related groups, and study on

the efficacy of nosocomial infection control. Economic implications for hospitals

under the prospective payment system. Am J Med 78: 3–7.

40. Lindenauer PK, Remus D, Roman S, Rothberg MB, Benjamin EM, et al. (2007)

Public reporting and pay for performance in hospital quality improvement. New

Engl J Med 356: 486–496.

41. Perencevich EN, Stone PW, Wright SB, Carmeli Y, Fisman DN, et al. (2007)

Raising standards while watching the bottom line: making a business case for

infection control. Infect Control Hosp Epidemiol 28: 1121–1133.

Hand Hygiene and Healthcare-Associated Infections

PLoS ONE | www.plosone.org 9 November 2011 | Volume 6 | Issue 11 | e27163