- Email: [email protected]
Interventions to improve hand hygiene compliance in patient care
Journal of Hospital Infection (2008) 68, 193e202
Available online at www.sciencedirect.com
www.elsevierhealth.com/journals/jhin
REVIEW
Interventions to improve hand hygiene compliance in patient care D.J. Gould a,*, N.S. Drey a, D. Moralejo b, J. Grimshaw c, J. Chudleigh a a
Health Care Research Unit, City University, London, UK Memorial University, St John’s N.L., Canada c University of Ottowa, Ontario, Canada b
Available online 15 January 2008
KEYWORDS Hand hygiene; Healthcare-associated infection; Research design
Summary Healthcare-associated infection is a major cause of morbidity and mortality. Hand hygiene is regarded as the most effective method of prevention but is poorly performed by health workers. We report a systematic review identifying studies which investigated the effectiveness of interventions to increase hand hygiene compliance short and longer term and to determine their success in terms of hand hygiene compliance and subsequent effect on rates of healthcare-associated infection. We employed the inclusion criteria employed by the Cochrane Effective Practice and Organisation of Care Group. Forty-eight studies and one thesis were identified. Only two met the stringent inclusion criteria. Overall studies remain small scale, poorly controlled and follow-up data collection is abandoned too soon to establish impact longer term. Furthermore, designs are insufficiently robust to attribute any observed changes to the intervention. Studies lack theoretical focus and seldom describe the intervention in sufficient detail, the change management process or contextual information about the organisation in the depth necessary to explain success or lack of it. The review concludes that interrupted time-series studies may offer the most rigorous approach to assessing the impact of interventions to increase hand hygiene compliance. In such study designs the number of new cases of healthcareassociated infection should be taken as an outcome measure, with data collection points at least 12 months before intervention and afterwards to allow for seasonal trends. Contextual factors at national and at local level should be carefully documented to take into consideration the influence of secular trends. ª 2007 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Address: Health Care Research Unit, City University, 24 Chiswell Street, London EC1 4TY, UK. Tel.: þ44 (0) 2070 405449; fax: þ44 (0) 207 5866. E-mail address: [email protected] 0195-6701/$ - see front matter ª 2007 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2007.11.013
194
Introduction Healthcare-associated infection (HAI) is a major problem globally. Ten percent of inpatients in the UK develop HAI, causing 5000 deaths and costing £930 million annually.1 Figures are similar in other developed countries.2 Although hand hygiene has been intensively promoted as the most important means of preventing HAI, numerous studies have demonstrated that compliance with hand hygiene recommendations is poor and interventions to improve it lack sustainability.3 In 2001 Naikoba and Hayward systematically reviewed 21 studies published before 2000.4 They classified 17 as uncontrolled trials. Fifteen took place in critical care units (CCUs). Numerous different interventions and combinations of interventions to improve hand hygiene were described. The reviewers concluded that multifaceted approaches promoted hand hygiene compliance more effectively than single interventions, in line with accepted opinion at that time.5 They suggested that education combining written material with reminders and continuous feedback on performance was more useful than interventions involving novel equipment (e.g. automated sinks, moisturised soaps). Multiple limitations were noted with these studies, including small sample sizes, short duration of follow-up; lack of or inappropriate control groups, and lack of generalisability from the CCU to other clinical settings. In most studies frequency of hand hygiene was taken as the outcome measure rather than microbiological data. However, this review included studies employing research designs that are too weak to justify causal inferences about the effects of interventions. The number of studies exploring the effectiveness of interventions to increase hand hygiene compliance has doubled since the publication of the above review.4 Interest has increased in response to a Swiss initiative which employed an uncontrolled before-and-after design.6 The findings suggested that a hospital-wide poster campaign combining performance feedback with alcohol hand rub placed at every bedside could result in sustained improvement in hand hygiene for some although not all groups of health workers (nurses but not doctors), reduction in HAI and reduced rates of meticillin-resistant Staphylococcus aureus (MRSA).6 Longer-term follow-up revealed continuing success.7 In this paper we present the findings of a newer, more rigorous systematic review with greater discussion of the methodological problems associated with study design and the content of
D.J. Gould et al. interventions intended to increase hand hygiene compliance. This paper is based on a Cochrane review first published in the Cochrane Database of Systematic Reviews, Issue 2, April 2007. Cochrane reviews are regularly updated as new evidence emerges and in response to comments and criticisms. The Cochrane Library should be consulted for the most recent version of the review.
Methods The aim of the review was to identify all studies investigating the effectiveness of interventions intended to increase hand hygiene compliance and/or use of alcohol hand rubs short term (less than six months) and longer term (six months or more) and to determine their success in terms of hand hygiene compliance and subsequent effect on rates of HAI.
Search strategy We employed the rigorous inclusion criteria employed by the Cochrane Effective Practice and Organisation of Care (EPOC) Group which is available on their website. We searched the Cochrane Central Register of Controlled Trials (CENTRAL) and the EPOC specialised register of trials, MEDLINE, PubMed, EMBASE, CINAHL and the BNI. All databases were searched to July 2006; MEDLINE was searched from 1980, CINAHL from its inception, and the remainder from 1990 until July 2006. The following search terms were used: ‘hand hygiene’ or ‘hand hygiene’ alone and combined with the following terms: ‘education’; ‘knowledge’; ‘health promotion’; ‘audit’; ‘compliance’; ‘product availability’; ‘cross-infection’; nosocomial infection’; hospital-acquired infection’ and ‘health care-associated infection’. Additional search strategies included: hand searching key journals from 1985 onwards (British Medical Journal, Journal of Hospital Infection, American Journal of Infection Control, Infection Control and Hospital Epidemiology). Conference proceedings from the UK Hospital Infection Society and the Infection Control Nurses’ Association were hand searched. Contact was established with the Hand Hygiene Liaison Group in the UK to discuss progress with their current study funded by the Department of Health. Conference proceedings from the Society of Hospital Epidemiologists of America, and the Community and Hospital Infection Control Association of Canada were hand searched. Colleagues from these organisations were contacted
Hand hygiene compliance in patient care for information about relevant unpublished work. Information was sought from pharmaceutical companies manufacturing hand hygiene products.
195 to highlight key issues for investigators, editors and referees.
Included studies Inclusion and exclusion criteria Two authors reviewed each paper independently. We considered randomised controlled trials, controlled clinical trials, controlled before-and-after studies and interrupted time-series analyses meeting explicit inclusion and quality criteria used by the EPOC Group. Study designs were analysed by annotating the points at which data collection took place in relation to intervention. Appendix 1 shows how annotation and analysis were performed.8 To be eligible for review, interrupted time-series studies had to include at least three data collection points before and after the intervention to take into account the influence of secular trends and the autocorrelation among measurements repeatedly taken over time.9 Seasonal variations (e.g. changes in humidity) might influence outcome measures in studies which examine HAI rates. Hand hygiene compliance is likely to be influenced by changes in staffing levels and replacement of usual personnel by temporary employees during national holidays and sickness. Studies reporting proxy indicators of hand hygiene compliance (e.g. increased use of soap or alcohol hand rub) were considered. We reviewed studies where the participants or target groups were nurses, doctors and other allied health professionals (except operating theatre staff because specific hand hygiene techniques are used in this setting). We also considered studies in any hospital or community setting, in any country that involved any type of intervention intended to improve hand hygiene compliance using aqueous solutions and/ or alcohol products (e.g. education; audit with performance feedback; health promotion; and variations in availability and types of hand hygiene products). Studies to promote compliance with universal precautions were included providing data relating specifically to hand hygiene were presented separately.
Results We identified 49 papers and one thesis published from 1980 onwards through electronic searching. No additional papers were identified by hand searching or communication with any organisations contacted. Only two studies met the minimum inclusion criteria for a Cochrane EPOC review.10,11 This paper describes included and excluded studies
Although meeting the criteria for review, both included studies were poorly designed (see Tables I and II). One had been included in the earlier systematic review.4 The other has been published since.10,11 Gould and Chamberlain reported a controlled before-and-after study conducted in four matched surgical wards in the same hospital.10 Two wards were randomly selected to serve as intervention wards. Two further matched wards were controls. Sixteen nurses in the intervention wards and 15 nurses in the controls were recruited randomly into the study. Each nurse was observed continuously for 2 h by the same observer, blinded to group allocation. The outcome measure was number of hand washes performed after activities judged to offer significant risk of cross-infection (‘essential’ hand hygiene episodes). The unit of analysis was the individual nurse. Baseline data were similar in control and intervention wards. Three months post-intervention the number of essential hand hygiene episodes was similar for both groups. In their randomised controlled trial Huang et al. recruited 100 nurses randomised into experimental and control groups.11 Data were collected from 98 nurses by direct observation by three researchers for 30 min each before the intervention and four months afterwards. Data were presented on the proportion of nurses who washed hands, but there was no information to explain how the proportion was calculated. The unit of analysis was the individual nurse. Four months post-intervention, hand hygiene compliance was significantly improved (P < 0.001) for the nurses in the experimental group but not the control. In both studies the outcome measure was the number of times hand hygiene was performed before and after specific types of patient contact collected by direct observation during day shifts. Microbiologically defined outcome measures were not used. The criteria used to determine when hand hygiene should be performed were explicitly stated, with reference to official guidelines from the USA or derived from them.12 Details of observers’ training were disclosed by Huang et al. but details of inter-rater reliability testing for the three data collectors were not supplied.11 The ‘Hawthorne effect’ (increased productivity, i.e. more hand hygiene episodes resulting from the presence of observers) was not discussed by Huang et al.
196 Table I
D.J. Gould et al. Characteristics of included studies
Gould and Chamberlain (1997)10 Methods Design Baseline measurements Appropriate choice of control Objective measure of performance Outcome measures Protection against contamination Validated audit tool used Inter-rater reliability tested Possibility of Hawthorne effect Participants Interventions Outcomes Notes Allocation concealment Huang et al. (2002)11 Methods Design Baseline measurements Appropriate choice of control Objective measure of performance Outcome measures Protection against contamination Validated audit tool used Inter-rate reliability tested Possibility of Hawthorne effect Participants Interventions Outcomes Notes
Controlled before and after study Done Done Done Reliable Not done Done Not appropriate Discussed UK nurses on general surgical wards Single teaching session: hand hygiene, universal precautions % frequency of hand washes after high-risk activities Intervention not successful at three months Not done
Randomised controlled clinical trial Done Done Done Not reliable Done Not stated Not done Not discussed People’s Republic of China Nurses throughout a hospital Education, mainly universal precautions % of nurses washing hands before and after patient contact Intervention successful after four months
and discounted by Gould and Chamberlain.10,11,13 They argued that hand hygiene is such an ingrained activity that it would be impossible for health workers to maintain changes in usual practice throughout the period of observation. This may have been true when their study was conducted as hand hygiene was not attracting its current level of interest, but would no longer be the case. Both studies featured a single intervention involving education relating to universal precautions as well as hand hygiene. In the study by Gould and Chamberlain teaching was provided by a nurse teacher with particular infection control expertise.10 The educational package was designed to consist of five different sessions each 30 min long, covering a comprehensive range of topics relating to infection control generally and specifically to universal precautions and hand hygiene. Practical demonstrations were included. Half the teaching sessions were cancelled because the wards were too busy. Thus some nurses failed to receive all the intended input. The teaching was,
however, well-evaluated. The intervention employed by Huang et al. involved 2 h of formal teaching about blood-borne pathogens and universal precautions delivered by specially trained nurses, 1 h of practical demonstration, 30 min of discussion, and written information.11 Details of evaluation and attendance were not disclosed. Neither study employed a theoretical framework to inform the intervention, but Gould and Chamberlain discussed the rationale for delivering work-based teaching with reference to the nursing education literature.10
Excluded studies (see Appendix 1) The thesis and the paper subsequently written from it were discounted because the data related to universal precautions with information pertaining to hand hygiene were not presented separately.14,15 A further paper was excluded because it did not contain data.16 Of the remaining 45 studies, three were excluded because no baseline data
Hand hygiene compliance in patient care Table II
197
Detailed comparison of included studies
Study
Results
Additional notes
Huang et al. (2002)11
Percentage of 49 nurses who used appropriate hand hygiene: e Before patient contact Experimental group Pre-intervention 51% Post-intervention 85.7% Control Pre-intervention 53.1% Post-intervention 53.1% e After patient contact Experimental group Pre-intervention 75.5% Post-intervention 91.8% Control Pre-intervention 75.5% Post-intervention 71.4%
Significant increase in experimental group at post-test for both before patient contact (PL.001) and after patient contact (PL.05) compared to control and baseline No confidence intervals reported
Gould and Chamberlain (1997)10
Percentage of essential hand decontamination: Experimental group Pre-intervention 54.5% Post-intervention 58.6% Control Pre-intervention 54.4% Post-intervention 64.1%
No significant difference between experimental group and control No confidence intervals reported
were collected or were collected on only a few of the participating wards.17e19 A further 21 studies were excluded because they reported uncontrolled before-and-after study designs.20e40 Three interrupted time-series studies were reported, each with less than three pre- and post-intervention data collection points.41e43 Of the remaining 20 studies, an additional 12 reported complicated before-and-after designs in which two or more sequential interventions took place, but with only one or two episodes of data collection after each new intervention.6,7,44e53 This group included an influential study which is widely quoted as evidence of the ability of hand hygiene campaigns to increase compliance and decrease rates of HAI and a longer follow-up building on the original work.6,7 In this group of studies, a single episode of baseline data collection took place with further data collected over extended periods. These long periods of data collection became interventional, because feedback performance was provided to health workers during each as part of a deliberately engineered Hawthorne effect. Six controlled before-and-after studies were identified. Each employed one intervention and
one control unit.54e59 These were excluded because the intervention was completely confounded by the study site. Thus it was difficult to attribute any observed changes to the intervention rather than to other site-specific variables. Additional weaknesses of these studies were the dissimilarities of the control and intervention sites, and, in some studies, imbalances between baseline hand hygiene rates. In the controlled before-and-after study by Mayer et al., the intervention and control wards were dissimilar.54 One was a medical ICU while the other was a surgical ICU. Wound care places greater demands on health workers to perform hand hygiene. In this study percentage number of hand hygiene episodes was the same (63%) on experimental and control wards, but on the experimental (surgical) unit demand for hand hygiene was more than double the demand on the control ward. Other differences between the wards likely to have influenced the outcome measures included differing admission rates and ratios of (less well qualified) nursing assistants to qualified nurses. The controlled before-and-after study reported by Larson et al. in 1991 was in two completely dissimilar clinical settings (neonatal intensive care
198 and post-anaesthesia recovery).55 Baseline measures of hand hygiene were not quoted separately. In the controlled before-and-after study reported by Larson et al. in 1997 a seven-bed neurosurgical ICU served as the experimental unit against a reasonably similar control (a seven-bed surgical ICU).56 However, provision for hand hygiene was not the same. In the experimental unit, five patients were nursed in a bay area with one sink. Two more beds were in single rooms, each with its own sink. The control unit was better equipped, with individual rooms for all patients, each with its own sink. Baseline observation of hand washing (not use of alcohol) was 151 episodes in the experimental unit compared to 310 episodes in the better-equipped control, although this difference was not statistically significant. The controlled before-and-after study published by Larson et al. in 2000 was sited in two hospitals.57 The authors state that the experimental and control hospitals were ‘similar’ in terms of nurse:patient ratios, staffing patterns and patient populations, but without supporting evidence. However, hand hygiene frequency was documented in only two dissimilar sites in these hospitals, adult medical critical care and neonatal intensive care, with no data to describe the characteristics of the two units. Baseline frequency of hand hygiene was greater in the experimental hospital [42.6%; confidence interval (CI): 1.3] than the control (30.3; CI: 1.52). The controlled before-and-after study by Bittner et al. was excluded because the experimental unit was a surgical CCU and the control was a medical CCU.58 The baseline rates of hand hygiene were similar. The controlled before-and-after study reported by Colombo et al. was excluded because of lack of matching between control and intervention sites.59 The intervention was conducted on surgical, medical and intensive care units. The control specifically excluded intensive care. This was not considered a valid comparison because of the differences between critical care and general ward patients. Nine of the excluded studies involved a single intervention featuring education or training related to hand hygiene, usually formal teaching with practical demonstrations.15,18,20,21,25,26,35,37,41 Hand hygiene was often discussed in conjunction with other topics (e.g. universal precautions, epidemiology). Dubbert et al. combined education with audit and feedback, while six studies looked at audit and feedback alone.27,36,38,39,43,52,59 Seven studies involved single interventions related to the introduction of a new hand hygiene product (e.g. emollient soap,55 alcohol hand rub17,22,28,29,34,59). Marena et al.
D.J. Gould et al. compared plain soap and antimicrobial solution, combined with education.32 Other single interventions studied were visual feedback of organisms from hand cultures,16 use of gowns,42 labelled teddy bears,30 labels on ventilators,47 reminders from patients,33,48 posters,19 voice prompts,50 automated sinks,55 and relocation to a new hospital.53 The remaining studies involved multidimensional campaigns featuring different combinations of an educational or training programme, a new product, audit and performance feedback, written information and written reminders such as posters or labels. Theoretical frameworks were clearly articulated in two studies only.24,57 Neither of the studies reviewed or those excluded considered economic outcomes. Apart from Pittet et al. none of the authors mentioned the cost of the resources required to increase hand hygiene compliance.6 Similarly there was no mention of health service utilisation outcomes such as: readmission rates, changes in levels of health care; length of patient stay; or the effects of any of the interventions on patients’ health.
Discussion Study designs There is a dearth of methodologically robust studies to explore the effectiveness of interventions to improve hand hygiene compliance. Studies remain small scale, poorly controlled, abandon follow-up too soon to establish longer-term impact and seldom include microbiological data as outcome measures. We rejected most studies included in the older review because they were uncontrolled before-and-after designs, interrupted time-series studies lacking sufficient data collection points or because they reported controlled before-and-after designs with poor controls and only one control and intervention site, making it impossible to attribute any observed changes to the intervention.4,9 Ethically conducted studies (those where data collection was not covert) were inherently weakened through the possibility of a Hawthorne effect.13 This was also a problem when, as is inevitable with hand hygiene, the intervention was not blind.60 Cluster-randomised controlled trials may offer a solution. They are recommended when it is not feasible to enroll individuals from the same venue (ward, hospital) because of cross-contamination (e.g. health workers discussing a training programme).61 For intervention studies to enhance hand hygiene they would be feasible because the intervention is usually applied to groups rather
Hand hygiene compliance in patient care than individuals.62 A future study design might involve random sampling wards of the same type. However, this approach has two drawbacks. Cluster-randomised controlled trials require large samples and are thus expensive to undertake as well as challenging in terms of project management.63 Difficulties matching intervention and control sites mean that for hand hygiene compliance, they suffer from the same disadvantages as controlled before-and-after studies; the many contextual factors have the potential to influence HAI rates and weaken findings.64 Well-designed interrupted time-series studies meeting strict criteria might offer a better solution.9 When applied to studies involving HAI as an outcome measure, data collection would need to take place for 12 months before intervention and again afterwards to allow for seasonal trends. Overcoming secular trends in studies lacking controls would never be entirely possible although qualitative data and routinely collected audit data for the extensive range of variables influencing hand hygiene compliance and HAI rates (e.g. policy initiatives, media storms, long-term trends in the epidemiology of HAI and the introduction of new drugs and therapies) could be taken into consideration. Arguably such studies would be better attempted outside the UK, since there have been excessively rapid changes in health policy directed towards reducing HAI, accompanied by increasing media attention.
Interventions Our review raises questions about the adequacy of interventions intended to achieve improved hand hygiene compliance. This is supported by one of the studies we reviewed,11 while in the other study failure to improve hand hygiene compliance can be attributed to inability to implement an otherwise well-designed, educationally sound programme.10 In future studies the nature and content of the intervention requires consideration in addition to its length. Older change management literature suggested that multifaceted campaigns are most effective,5 but this view has now been challenged.65 According to more recent evidence, interventions previously thought to be ineffective such as education are modestly successful.65,66 Audit with performance feedback is less successful than educational intervention.65 Few existing hand hygiene intervention studies described the intervention adequately. Authors used the terms ‘education’ and ‘training’ interchangeably or were confused about their meanings. The purpose of education is to promote intellectual curiosity, development (personally,
199 professionally and academically) and encourage the aspiration which must underpin service transformation.67 Training is a narrower endeavour, promoting discipline, but encouraging rigidity and inhibiting development.67 The interventions described by Gould and Chamberlain and Huang et al. fulfil the criteria of education; however, the Geneva initiative, and others that it has stimulated such as the CleanYourHandsCampaign in England and Wales, represent training because of their heavy reliance on audit with performance feedback as the drivers for behaviour change.6,10,11,68 Educational initiatives are resource intensive and expensive, but if well-designed and well-implemented, have the potential to effect sustainable change. Training is less expensive but its effects are likely to be shortlived and influenced by staff turnover and shortage, because health workers are coached to undertake a repetitive set of activities rather than problemsolve. Future ventures to increase hand hygiene compliance should state explicitly whether they are employing education or training, with underlying rationale. This would pave the way towards the development of strong theoretical frameworks, which are essential for the external validity, and generalisability of study findings. Descriptions of educational interventions should include: rationale for choice of educational approach and venue; who delivered the education and their preparation; programme content; numbers of health workers attending; evaluation; changes necessary to the planned programme; and their impact. Studies geared towards training that involve audit with performance feedback should indicate the nature of the audit tool and how feedback was provided. In conclusion, initiatives to enhance hand hygiene compliance lack rigour. Existing study designs are inherently weak. Possible alternatives are insufficiently robust because the many contextual variables likely to influence hand hygiene compliance and HAI rates are not amenable to control. Future studies must combine the expertise of clinical scientists with that of behavioural scientists able to include qualitative analysis of contextual factors into the analysis of findings. Interventions intended to improve compliance are poorly articulated and not grounded in theoretical frameworks necessary for external validity. Conflict of interest statement D.J.G. co-authored one of the studies included in this review. Funding sources None.
200
Appendix 1: Analysis of the research design of excluded studies The studies were summarised by annotating the data collection points as recommended by Cook and Campbell.8 O, data collection point; X, the intervention.
Studies with few/no baseline data Maury et al.17: X O1 Panhotra et al.18: X O1 O2 O3 On Thomas et al.19: O1 X1 X2 O2 (baseline data were collected in one of the participating units only)
Uncontrolled before-and-after study designs Baker20: O1 X O2, O3 Berg et al.21: O1 X O2 Brown et al.22: O1 X O2 Coignard et al.23: O1 X O2 Creedon24: O1 X O2 Diekema et al.25: O1 X O2 O3 O4 Dorsey et al.26: O1 X O2 Dubbert et al.27: O1 X O2 Earl et al.28: O1 X O2 O3 Graham29: O1 X O2 Hughes et al.30: O1 X O2 Lam et al.31: O1 X O2 Marena et al.32: O1 X O2 McGuckin et al.33: O1 X O2 Muto et al.34: O1 X O2 Prieto and Macleod-Clark et al.35: O1 X O2 O3 Salemi et al.36: O1 X O2 O3 O4 O5 Shaw and Tanner et al.37: O1 X O2 Van de Mortel and Heyman38: O1 X O2 Van de Mortel et al.39: O1 X O2 O3 Won et al.40: O1 X O2 O3
Interrupted time-series designs (less than three data entry points before/after each intervention Conly et al.41: O1 O2 X O3 O4 Donowitz42: O1 O2 X O3 O4 Raju and Kobler43: O1 O2 X O3 O4 O5
Uncontrolled before-and-after study designs with sequential data collection Pittet et al.6 with Hugonnet et al.7 comprising a unit: O1 X O2 O3 X2 X3 O4 Xn On Avila-Aguero et al.44: O1 X1 O2 X2 O3 O4
D.J. Gould et al. Bischoff et al.45: O1 X2 O2 X2 X3 04 Harbarth et al.46: O1 X1 O2 X2 O3 X3 O4 Khatib et al.47: O1 X1 O2a O2b O2c X2 O3a O3b O3c O3d McGuckin et al.48: O1 X1 O2 X2 O3 O4 Rosenthal et al.49: O1 X1 O2 X2 O3 Swaboda et al.50: O1 X1 O2 X2 O3 Simmons et al.51: O1 X1 O2 Xn Tibbals et al.52: O1 X1 O2 X2 O3 Whitby and McLaws53: O1 X1 O2 X2 O3
Controlled before-and-after study designs with one intervention and one control site Mayer et al.54 Larson et al.55 Larson et al.56 Larson et al.57 Bittner et al.58 Colombo et al.59
References 1. National Audit Office. The management and control of hospital-acquired infection in acute NHS trusts in England. London: National Audit Office; 1998. 2. Zoutman DE, Ford DB, Bryce E, et al. The state of infection surveillance and control in Canadian acute care hospitals. Am J Infect Control 2003;31:266e273. 3. Teare EL, Cookson B, French G, et al. Leader. UK handwashing initiative. J Hosp Infect 1999;43:1e3. 4. Naikoba S, Hayward A. The effectiveness of interventions aimed at increasing handwashing in healthcare workers e a systematic review. J Hosp Infect 2001;47:173e180. 5. National Health Service Centre for Research and Development. Getting evidence into practice 1999;5:1e15. 6. Pittet D, Hugonnet S, Mourouga P, et al. Effectiveness of a hospital-wide programme to improve compliance with hand hygiene. Lancet 2000;356:1307e1312. 7. Hugonnet S, Pernager TV, Pittet D. Alcohol-based handrub improves compliance with hand hygiene in intensive care units. Arch Intern Med 2002;162:1037e1043. 8. Cook CT, Campbell DT. Quasi experimentation. Boston: Houghton Mifflin; 1979. 9. Ramsay CR, Matowe L, Grilli R, et al. Interrupted time series designs in health technology assessment: lessons from two systematic reviews of behavior change strategies. Int J Technol Assess Health Care 2003;18:613e 623. 10. Gould DJ, Chamberlain A. The use of a ward-based educational package to enhance nurses’ compliance with infection control procedures. J Clin Nurs 1997;6:55e67. 11. Huang J, Jiang D, Wang X. Changing knowledge, behavior and practice related to universal precautions among hospital nurses in China. J Contin Educ Nurs 2002;35:217e224. 12. Larson E. APIC Guidelines for hand hygiene and hand antisepsis in health care settings. Am J Infect Control 1995; 23:251e269. 13. Roethlisberger FJ, Dickson WJ. Management and the worker. Cambridge, MA: Harvard University Press; 1939. 14. Moongtui W. Compliance with components of universal precautions guidelines: using peer feedback program as
Hand hygiene compliance in patient care
15.
16. 17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30. 31.
32.
33.
34.
a cue to action among health care workers in Thailand. PhD Dissertation, University of Alabama, Birmingham, 1999. Moongtui W, Gauthier D, Turner J. Using peer feedback to improve handwashing and glove usage among Thai health care workers. Am J Infect Control 2000;28:365e369. Moore M, Abbott NR. Can handwashing practice be changed? Am J Nurs 1980;80:80. Maury E, Alzieu M, Baudel JL, et al. Availability of an alcohol solution can improve hand disinfection compliance in an intensive care unit. Am J Respir Crit Care Med 2000;162: 324e332. Panhotra BR, Saxena AK, Al-Arabi AM. The effect of a continuous educational program on handwashing compliance among healthcare workers in an intensive care unit. Br J Infect Control 2004;5:15e18. Thomas M, Gilespie W, Krauss J, et al. Focus group data as a tool in assessing effectiveness of a hand hygiene campaign. Am J Infect Control 2005;33:368e373. Baker DJ. Effects of a video based staff training with manager led exercises in residential support. Ment Retard 1998; 36:198e204. Berg DE, Hershaw RC, Claudia AR, Ramirez CA, Weinstein RA. Control of nosocomial infections in an ICU in Guatemala City. Clin Infect Dis 1995;21:588e593. Brown SM, Lubimova AV, Khrustalyeva NM, et al. Use of an alcohol-based hand rub and quality improvement interventions to improve hand hygiene in a Russian intensive care unit. Infect Control Hosp Epidemiol 2003;24:172e179. Coignard B, Grandbastien B, Berrouane Y, et al. Handwashing quality impact of a special program. Infect Control Hosp Epidemiol 1998;19:510e513. Creedon SA. Health care workers’ hand decontamination practices: compliance with recommended guidelines. J Adv Nurs 2005;51:208e216. Diekema DJ, Schult S, Albanese MA, Doebbeling BN. Universal precautions training of preclinical students: impact on knowledge, attitudes and compliance. Prev Med 1995;24: 580e585. Dorsey ST, Cydulka RK, Emerman CL. Is hand washing teachable? Failure to improve hand washing behaviour in an urban emergency department. J Academic Emerg Care 1996;3:360e365. Dubbert PM, Dolce J, Richter W, Miller M, Chapman SW. Increasing ICU staff hand washing: effects of education and group feedback. Infect Control Hosp Epidemiol 1990; 11:191e193. Earl ML, Jackson M, Rickman LS. Improved rates of compliance with hand antisepsis guidelines: a three-phase observational study. Am J Nur 2001;101:26e33. Graham M. Frequency and duration of hand washing in an intensive care unit. Am J Infect Control 1990;18: 77e81. Hughes WT, Williams B, Pearson T. The nosocomial colonisation of T-Bear. Infect Control 1986;7:495e5000. Lam BC, Lee J, Lau YL. Hand hygiene in a neonatal intensive care unit: a multimodal intervention and impact on nosocomial infection. Pediatrics 2004;114:565e571. Marena C, Lodola L, Zecca M, et al. Assessment of handwashing practices with chemical and microbiologic methods: preliminary results from a prospective crossover study. Am J Infect Control 2002;30:334e340. McGuckin M, Waterman R, Porten L, et al. Patient education model for increasing hand washing compliance. Am J Infect Control 1999;27:309e314. Muto CA, Sistrom MG, Farr BM. Hand hygiene rates unchanged by installation of dispensers of a rapidly acting hand antiseptic. Am J Infect Control 2000;28:273e276.
201 35. Prieto J, Macleod-Clark J. Contact precautions for Clostridium difficile and methicillin-resistant Staphylococcus aureus (MRSA). J Res Nurs 2005;10:511e526. 36. Salemi C, Canola T, Eck E. Hand washing and physicians: how to get them to together. Infect Control Hosp Epidemiol 2002;23:32e35. 37. Shaw S, Tanner J. Hand washing practices on a neonatal unit: the influence of teaching sessions. J Neonatal Nurs 2003;9:162e166. 38. Van de Mortel T, Heyman L. Performance feedback increases the incidence of hand washing by staff following patient contact in an ICU. Aust Crit Care 1995;8:8e13. 39. Van de Mortel T, Bourke R, Phillips I, et al. Maximising handwashing rates in the critical care unit through yearly performance feedback. Aust Crit Care 2000;13:91e95. 40. Won SP, Chou HC, Hsiah WS, et al. Hand washing program for the prevention of nosocomial infections in a neonatal intensive care unit. Infect Control Hosp Epidemiol 2004; 25:742e746. 41. Conly JM, Hills S, Ross J, Lertzman J, Thomas LJ. Hand washing practices in an ICU: the effects of an educational program and its relationship to infection rates. Am J Infect Control 1989;17:330e339. 42. Donowitz LG. Failure of the overgown to prevent nosocomial infection in paediatric ICU. Pediatrics 1986;77: 35e38. 43. Raju NK, Kobler C. Improving hand hygiene habits in the newborn nurseries. Am J Med Sci 1991;302:355e358. 44. Avila-Aguero M, Umalha MA, Jiminez AL. Handwashing practices in tertiary-care, pediatric hospital and the effect on an educational program. Clin Perform Qual Health Care 1998;6:70e72. 45. Bischoff WE, Reynolds TM, Curtis N, Sessler CN, Edward MB, Wenzel RP. Handwashing compliance by health care workers: the impact of introducing accessible alcohol-based hand antiseptic. Arch Int Med 2000;160:1017e1021. 46. Harbarth S, Pittet D, Grady L, et al. Interventional study to evaluate the impact of an alcohol based hand gel in improving hand hygiene compliance. Pediatr Infect Dis J 2002;21: 496e497. 47. Khatib M, Jamaladdine G, Abdullah A, Ibrahim Y. Handwashing and use of gloves while managing patients receiving mechanical ventilation in the ICU. Chest 1999;116: 172e175. 48. McGuckin M, Taylor A, Martin V, et al. Evaluation of a patient education model for increasing hand hygiene compliance in an inpatient rehabilitation unit. Am J Infect Control 2004;32:235e238. 49. Rosenthal VD, McCormick RD, Guzman S, Villamayor C, Orellano PW. Effect of education and performance feedback on handwashing: the benefit of administrative support in Argentinean hospitals. Am J Infect Control 2003;31: 85e92. 50. Swaboda SM, Earsing K, Strauss K, Lane S, Lipsett PA. Electronic monitoring and voice prompts to improve hand hygiene and decrease nosocomial infections in critical care. Crit Care Med 2004;32:358e363. 51. Simmons B, Bryant J, Neiman R, Spencer L, Arheart K. The role of hand washing in preventing endemic ICU infections. Infect Control Hosp Epidemiol 1990;11:589e594. 52. Tibbals J. Teaching hospital medical staff to hand wash. Med J Aust 1996;164:395e398. 53. Whitby M, McLaws ML. Handwashing in healthcare workers: accessibility of sink location does not improve compliance. J Hosp Infect 2004;58:247e253. 54. Mayer JA, Dubbert PM, Miller M, Burkett P, Chapman S. Increasing handwashing in an ICU. Infect Control 1986;7:259e262.
202 55. Larson E, McGeer A, Quaraishi ZAE, et al. Effect of an automated sink on hand washing practices and attitudes in high risk units. Infect Control Hosp Epidemiol 1991;12: 422e428. 56. Larson E, Bryan JL, Adler LM, Lee M, Blane C. Multifaceted approach to changing hand washing behavior. Am J Infect Control 1997;25:3e10. 57. Larson EL, Early E, Cloonan P, Surgue S, Parides M. An organizational climate intervention associated with handwashing and decreased nosocomial infections. Behav Med 2000;26:14e22. 58. Bittner MJ, Rich EC, Turner PD, Arnold WH. Limited impact of sustained simple feedback based on soap and paper towel consumption on the frequency of hand washing in an adult intensive care unit. Infect Control Hosp Epidemiol 2002;23:120e126. 59. Colombo C, Giger H, Grote C, et al. Impact of teaching interventions on nurse compliance with hand disinfection. J Hosp Infect 2002;51:69e72. 60. Gould DJ, Chudleigh J, Drey NS, et al. Measuring handwashing performance in health service audits and research studies. J Hosp Infect 2007;66:109e115.
D.J. Gould et al. 61. Medical Research Council. Cluster randomised trials: methodological and ethical considerations. In: London: MRC Clinical Trials Series 2002. 62. Donner A, Klar N. Pitfalls in cluster randomization trials. Am J Public Health 2004;94:416e422. 63. Kerry SM, Bland JM. Statistics notes: sample size in cluster randomisation. Br Med J 1998;316:549e550. 64. Togerson DJ. Contamination in trials: is cluster randomisation the answer? Br Med J 2001;322:355e357. 65. Jamtvedt G, Young JM, Kristofferson DT, et al. Audit and feedback: effects on professional practice and health outcomes [Review]. Cochrane Library 2007;(1):1e57. 66. Grimshaw J, Eccles M, Tetroe J. Implementing clinical guidelines: current evidence and future implications. J Contin Educ Health Prof 2004;24:S31eS37. 67. Tosey P, McNair S. Work-related learning. In: Jarvis P, editor. The age of learning: education and the learning society. London: Kogan Page; 2001. p. 95e108. 68. Storr J. The effectiveness of the national cleanyourhandscampaign. Nurs Times 2005;101:50e51.
Available online at www.sciencedirect.com
www.elsevierhealth.com/journals/jhin
REVIEW
Interventions to improve hand hygiene compliance in patient care D.J. Gould a,*, N.S. Drey a, D. Moralejo b, J. Grimshaw c, J. Chudleigh a a
Health Care Research Unit, City University, London, UK Memorial University, St John’s N.L., Canada c University of Ottowa, Ontario, Canada b
Available online 15 January 2008
KEYWORDS Hand hygiene; Healthcare-associated infection; Research design
Summary Healthcare-associated infection is a major cause of morbidity and mortality. Hand hygiene is regarded as the most effective method of prevention but is poorly performed by health workers. We report a systematic review identifying studies which investigated the effectiveness of interventions to increase hand hygiene compliance short and longer term and to determine their success in terms of hand hygiene compliance and subsequent effect on rates of healthcare-associated infection. We employed the inclusion criteria employed by the Cochrane Effective Practice and Organisation of Care Group. Forty-eight studies and one thesis were identified. Only two met the stringent inclusion criteria. Overall studies remain small scale, poorly controlled and follow-up data collection is abandoned too soon to establish impact longer term. Furthermore, designs are insufficiently robust to attribute any observed changes to the intervention. Studies lack theoretical focus and seldom describe the intervention in sufficient detail, the change management process or contextual information about the organisation in the depth necessary to explain success or lack of it. The review concludes that interrupted time-series studies may offer the most rigorous approach to assessing the impact of interventions to increase hand hygiene compliance. In such study designs the number of new cases of healthcareassociated infection should be taken as an outcome measure, with data collection points at least 12 months before intervention and afterwards to allow for seasonal trends. Contextual factors at national and at local level should be carefully documented to take into consideration the influence of secular trends. ª 2007 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Address: Health Care Research Unit, City University, 24 Chiswell Street, London EC1 4TY, UK. Tel.: þ44 (0) 2070 405449; fax: þ44 (0) 207 5866. E-mail address: [email protected] 0195-6701/$ - see front matter ª 2007 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2007.11.013
194
Introduction Healthcare-associated infection (HAI) is a major problem globally. Ten percent of inpatients in the UK develop HAI, causing 5000 deaths and costing £930 million annually.1 Figures are similar in other developed countries.2 Although hand hygiene has been intensively promoted as the most important means of preventing HAI, numerous studies have demonstrated that compliance with hand hygiene recommendations is poor and interventions to improve it lack sustainability.3 In 2001 Naikoba and Hayward systematically reviewed 21 studies published before 2000.4 They classified 17 as uncontrolled trials. Fifteen took place in critical care units (CCUs). Numerous different interventions and combinations of interventions to improve hand hygiene were described. The reviewers concluded that multifaceted approaches promoted hand hygiene compliance more effectively than single interventions, in line with accepted opinion at that time.5 They suggested that education combining written material with reminders and continuous feedback on performance was more useful than interventions involving novel equipment (e.g. automated sinks, moisturised soaps). Multiple limitations were noted with these studies, including small sample sizes, short duration of follow-up; lack of or inappropriate control groups, and lack of generalisability from the CCU to other clinical settings. In most studies frequency of hand hygiene was taken as the outcome measure rather than microbiological data. However, this review included studies employing research designs that are too weak to justify causal inferences about the effects of interventions. The number of studies exploring the effectiveness of interventions to increase hand hygiene compliance has doubled since the publication of the above review.4 Interest has increased in response to a Swiss initiative which employed an uncontrolled before-and-after design.6 The findings suggested that a hospital-wide poster campaign combining performance feedback with alcohol hand rub placed at every bedside could result in sustained improvement in hand hygiene for some although not all groups of health workers (nurses but not doctors), reduction in HAI and reduced rates of meticillin-resistant Staphylococcus aureus (MRSA).6 Longer-term follow-up revealed continuing success.7 In this paper we present the findings of a newer, more rigorous systematic review with greater discussion of the methodological problems associated with study design and the content of
D.J. Gould et al. interventions intended to increase hand hygiene compliance. This paper is based on a Cochrane review first published in the Cochrane Database of Systematic Reviews, Issue 2, April 2007. Cochrane reviews are regularly updated as new evidence emerges and in response to comments and criticisms. The Cochrane Library should be consulted for the most recent version of the review.
Methods The aim of the review was to identify all studies investigating the effectiveness of interventions intended to increase hand hygiene compliance and/or use of alcohol hand rubs short term (less than six months) and longer term (six months or more) and to determine their success in terms of hand hygiene compliance and subsequent effect on rates of HAI.
Search strategy We employed the rigorous inclusion criteria employed by the Cochrane Effective Practice and Organisation of Care (EPOC) Group which is available on their website. We searched the Cochrane Central Register of Controlled Trials (CENTRAL) and the EPOC specialised register of trials, MEDLINE, PubMed, EMBASE, CINAHL and the BNI. All databases were searched to July 2006; MEDLINE was searched from 1980, CINAHL from its inception, and the remainder from 1990 until July 2006. The following search terms were used: ‘hand hygiene’ or ‘hand hygiene’ alone and combined with the following terms: ‘education’; ‘knowledge’; ‘health promotion’; ‘audit’; ‘compliance’; ‘product availability’; ‘cross-infection’; nosocomial infection’; hospital-acquired infection’ and ‘health care-associated infection’. Additional search strategies included: hand searching key journals from 1985 onwards (British Medical Journal, Journal of Hospital Infection, American Journal of Infection Control, Infection Control and Hospital Epidemiology). Conference proceedings from the UK Hospital Infection Society and the Infection Control Nurses’ Association were hand searched. Contact was established with the Hand Hygiene Liaison Group in the UK to discuss progress with their current study funded by the Department of Health. Conference proceedings from the Society of Hospital Epidemiologists of America, and the Community and Hospital Infection Control Association of Canada were hand searched. Colleagues from these organisations were contacted
Hand hygiene compliance in patient care for information about relevant unpublished work. Information was sought from pharmaceutical companies manufacturing hand hygiene products.
195 to highlight key issues for investigators, editors and referees.
Included studies Inclusion and exclusion criteria Two authors reviewed each paper independently. We considered randomised controlled trials, controlled clinical trials, controlled before-and-after studies and interrupted time-series analyses meeting explicit inclusion and quality criteria used by the EPOC Group. Study designs were analysed by annotating the points at which data collection took place in relation to intervention. Appendix 1 shows how annotation and analysis were performed.8 To be eligible for review, interrupted time-series studies had to include at least three data collection points before and after the intervention to take into account the influence of secular trends and the autocorrelation among measurements repeatedly taken over time.9 Seasonal variations (e.g. changes in humidity) might influence outcome measures in studies which examine HAI rates. Hand hygiene compliance is likely to be influenced by changes in staffing levels and replacement of usual personnel by temporary employees during national holidays and sickness. Studies reporting proxy indicators of hand hygiene compliance (e.g. increased use of soap or alcohol hand rub) were considered. We reviewed studies where the participants or target groups were nurses, doctors and other allied health professionals (except operating theatre staff because specific hand hygiene techniques are used in this setting). We also considered studies in any hospital or community setting, in any country that involved any type of intervention intended to improve hand hygiene compliance using aqueous solutions and/ or alcohol products (e.g. education; audit with performance feedback; health promotion; and variations in availability and types of hand hygiene products). Studies to promote compliance with universal precautions were included providing data relating specifically to hand hygiene were presented separately.
Results We identified 49 papers and one thesis published from 1980 onwards through electronic searching. No additional papers were identified by hand searching or communication with any organisations contacted. Only two studies met the minimum inclusion criteria for a Cochrane EPOC review.10,11 This paper describes included and excluded studies
Although meeting the criteria for review, both included studies were poorly designed (see Tables I and II). One had been included in the earlier systematic review.4 The other has been published since.10,11 Gould and Chamberlain reported a controlled before-and-after study conducted in four matched surgical wards in the same hospital.10 Two wards were randomly selected to serve as intervention wards. Two further matched wards were controls. Sixteen nurses in the intervention wards and 15 nurses in the controls were recruited randomly into the study. Each nurse was observed continuously for 2 h by the same observer, blinded to group allocation. The outcome measure was number of hand washes performed after activities judged to offer significant risk of cross-infection (‘essential’ hand hygiene episodes). The unit of analysis was the individual nurse. Baseline data were similar in control and intervention wards. Three months post-intervention the number of essential hand hygiene episodes was similar for both groups. In their randomised controlled trial Huang et al. recruited 100 nurses randomised into experimental and control groups.11 Data were collected from 98 nurses by direct observation by three researchers for 30 min each before the intervention and four months afterwards. Data were presented on the proportion of nurses who washed hands, but there was no information to explain how the proportion was calculated. The unit of analysis was the individual nurse. Four months post-intervention, hand hygiene compliance was significantly improved (P < 0.001) for the nurses in the experimental group but not the control. In both studies the outcome measure was the number of times hand hygiene was performed before and after specific types of patient contact collected by direct observation during day shifts. Microbiologically defined outcome measures were not used. The criteria used to determine when hand hygiene should be performed were explicitly stated, with reference to official guidelines from the USA or derived from them.12 Details of observers’ training were disclosed by Huang et al. but details of inter-rater reliability testing for the three data collectors were not supplied.11 The ‘Hawthorne effect’ (increased productivity, i.e. more hand hygiene episodes resulting from the presence of observers) was not discussed by Huang et al.
196 Table I
D.J. Gould et al. Characteristics of included studies
Gould and Chamberlain (1997)10 Methods Design Baseline measurements Appropriate choice of control Objective measure of performance Outcome measures Protection against contamination Validated audit tool used Inter-rater reliability tested Possibility of Hawthorne effect Participants Interventions Outcomes Notes Allocation concealment Huang et al. (2002)11 Methods Design Baseline measurements Appropriate choice of control Objective measure of performance Outcome measures Protection against contamination Validated audit tool used Inter-rate reliability tested Possibility of Hawthorne effect Participants Interventions Outcomes Notes
Controlled before and after study Done Done Done Reliable Not done Done Not appropriate Discussed UK nurses on general surgical wards Single teaching session: hand hygiene, universal precautions % frequency of hand washes after high-risk activities Intervention not successful at three months Not done
Randomised controlled clinical trial Done Done Done Not reliable Done Not stated Not done Not discussed People’s Republic of China Nurses throughout a hospital Education, mainly universal precautions % of nurses washing hands before and after patient contact Intervention successful after four months
and discounted by Gould and Chamberlain.10,11,13 They argued that hand hygiene is such an ingrained activity that it would be impossible for health workers to maintain changes in usual practice throughout the period of observation. This may have been true when their study was conducted as hand hygiene was not attracting its current level of interest, but would no longer be the case. Both studies featured a single intervention involving education relating to universal precautions as well as hand hygiene. In the study by Gould and Chamberlain teaching was provided by a nurse teacher with particular infection control expertise.10 The educational package was designed to consist of five different sessions each 30 min long, covering a comprehensive range of topics relating to infection control generally and specifically to universal precautions and hand hygiene. Practical demonstrations were included. Half the teaching sessions were cancelled because the wards were too busy. Thus some nurses failed to receive all the intended input. The teaching was,
however, well-evaluated. The intervention employed by Huang et al. involved 2 h of formal teaching about blood-borne pathogens and universal precautions delivered by specially trained nurses, 1 h of practical demonstration, 30 min of discussion, and written information.11 Details of evaluation and attendance were not disclosed. Neither study employed a theoretical framework to inform the intervention, but Gould and Chamberlain discussed the rationale for delivering work-based teaching with reference to the nursing education literature.10
Excluded studies (see Appendix 1) The thesis and the paper subsequently written from it were discounted because the data related to universal precautions with information pertaining to hand hygiene were not presented separately.14,15 A further paper was excluded because it did not contain data.16 Of the remaining 45 studies, three were excluded because no baseline data
Hand hygiene compliance in patient care Table II
197
Detailed comparison of included studies
Study
Results
Additional notes
Huang et al. (2002)11
Percentage of 49 nurses who used appropriate hand hygiene: e Before patient contact Experimental group Pre-intervention 51% Post-intervention 85.7% Control Pre-intervention 53.1% Post-intervention 53.1% e After patient contact Experimental group Pre-intervention 75.5% Post-intervention 91.8% Control Pre-intervention 75.5% Post-intervention 71.4%
Significant increase in experimental group at post-test for both before patient contact (PL.001) and after patient contact (PL.05) compared to control and baseline No confidence intervals reported
Gould and Chamberlain (1997)10
Percentage of essential hand decontamination: Experimental group Pre-intervention 54.5% Post-intervention 58.6% Control Pre-intervention 54.4% Post-intervention 64.1%
No significant difference between experimental group and control No confidence intervals reported
were collected or were collected on only a few of the participating wards.17e19 A further 21 studies were excluded because they reported uncontrolled before-and-after study designs.20e40 Three interrupted time-series studies were reported, each with less than three pre- and post-intervention data collection points.41e43 Of the remaining 20 studies, an additional 12 reported complicated before-and-after designs in which two or more sequential interventions took place, but with only one or two episodes of data collection after each new intervention.6,7,44e53 This group included an influential study which is widely quoted as evidence of the ability of hand hygiene campaigns to increase compliance and decrease rates of HAI and a longer follow-up building on the original work.6,7 In this group of studies, a single episode of baseline data collection took place with further data collected over extended periods. These long periods of data collection became interventional, because feedback performance was provided to health workers during each as part of a deliberately engineered Hawthorne effect. Six controlled before-and-after studies were identified. Each employed one intervention and
one control unit.54e59 These were excluded because the intervention was completely confounded by the study site. Thus it was difficult to attribute any observed changes to the intervention rather than to other site-specific variables. Additional weaknesses of these studies were the dissimilarities of the control and intervention sites, and, in some studies, imbalances between baseline hand hygiene rates. In the controlled before-and-after study by Mayer et al., the intervention and control wards were dissimilar.54 One was a medical ICU while the other was a surgical ICU. Wound care places greater demands on health workers to perform hand hygiene. In this study percentage number of hand hygiene episodes was the same (63%) on experimental and control wards, but on the experimental (surgical) unit demand for hand hygiene was more than double the demand on the control ward. Other differences between the wards likely to have influenced the outcome measures included differing admission rates and ratios of (less well qualified) nursing assistants to qualified nurses. The controlled before-and-after study reported by Larson et al. in 1991 was in two completely dissimilar clinical settings (neonatal intensive care
198 and post-anaesthesia recovery).55 Baseline measures of hand hygiene were not quoted separately. In the controlled before-and-after study reported by Larson et al. in 1997 a seven-bed neurosurgical ICU served as the experimental unit against a reasonably similar control (a seven-bed surgical ICU).56 However, provision for hand hygiene was not the same. In the experimental unit, five patients were nursed in a bay area with one sink. Two more beds were in single rooms, each with its own sink. The control unit was better equipped, with individual rooms for all patients, each with its own sink. Baseline observation of hand washing (not use of alcohol) was 151 episodes in the experimental unit compared to 310 episodes in the better-equipped control, although this difference was not statistically significant. The controlled before-and-after study published by Larson et al. in 2000 was sited in two hospitals.57 The authors state that the experimental and control hospitals were ‘similar’ in terms of nurse:patient ratios, staffing patterns and patient populations, but without supporting evidence. However, hand hygiene frequency was documented in only two dissimilar sites in these hospitals, adult medical critical care and neonatal intensive care, with no data to describe the characteristics of the two units. Baseline frequency of hand hygiene was greater in the experimental hospital [42.6%; confidence interval (CI): 1.3] than the control (30.3; CI: 1.52). The controlled before-and-after study by Bittner et al. was excluded because the experimental unit was a surgical CCU and the control was a medical CCU.58 The baseline rates of hand hygiene were similar. The controlled before-and-after study reported by Colombo et al. was excluded because of lack of matching between control and intervention sites.59 The intervention was conducted on surgical, medical and intensive care units. The control specifically excluded intensive care. This was not considered a valid comparison because of the differences between critical care and general ward patients. Nine of the excluded studies involved a single intervention featuring education or training related to hand hygiene, usually formal teaching with practical demonstrations.15,18,20,21,25,26,35,37,41 Hand hygiene was often discussed in conjunction with other topics (e.g. universal precautions, epidemiology). Dubbert et al. combined education with audit and feedback, while six studies looked at audit and feedback alone.27,36,38,39,43,52,59 Seven studies involved single interventions related to the introduction of a new hand hygiene product (e.g. emollient soap,55 alcohol hand rub17,22,28,29,34,59). Marena et al.
D.J. Gould et al. compared plain soap and antimicrobial solution, combined with education.32 Other single interventions studied were visual feedback of organisms from hand cultures,16 use of gowns,42 labelled teddy bears,30 labels on ventilators,47 reminders from patients,33,48 posters,19 voice prompts,50 automated sinks,55 and relocation to a new hospital.53 The remaining studies involved multidimensional campaigns featuring different combinations of an educational or training programme, a new product, audit and performance feedback, written information and written reminders such as posters or labels. Theoretical frameworks were clearly articulated in two studies only.24,57 Neither of the studies reviewed or those excluded considered economic outcomes. Apart from Pittet et al. none of the authors mentioned the cost of the resources required to increase hand hygiene compliance.6 Similarly there was no mention of health service utilisation outcomes such as: readmission rates, changes in levels of health care; length of patient stay; or the effects of any of the interventions on patients’ health.
Discussion Study designs There is a dearth of methodologically robust studies to explore the effectiveness of interventions to improve hand hygiene compliance. Studies remain small scale, poorly controlled, abandon follow-up too soon to establish longer-term impact and seldom include microbiological data as outcome measures. We rejected most studies included in the older review because they were uncontrolled before-and-after designs, interrupted time-series studies lacking sufficient data collection points or because they reported controlled before-and-after designs with poor controls and only one control and intervention site, making it impossible to attribute any observed changes to the intervention.4,9 Ethically conducted studies (those where data collection was not covert) were inherently weakened through the possibility of a Hawthorne effect.13 This was also a problem when, as is inevitable with hand hygiene, the intervention was not blind.60 Cluster-randomised controlled trials may offer a solution. They are recommended when it is not feasible to enroll individuals from the same venue (ward, hospital) because of cross-contamination (e.g. health workers discussing a training programme).61 For intervention studies to enhance hand hygiene they would be feasible because the intervention is usually applied to groups rather
Hand hygiene compliance in patient care than individuals.62 A future study design might involve random sampling wards of the same type. However, this approach has two drawbacks. Cluster-randomised controlled trials require large samples and are thus expensive to undertake as well as challenging in terms of project management.63 Difficulties matching intervention and control sites mean that for hand hygiene compliance, they suffer from the same disadvantages as controlled before-and-after studies; the many contextual factors have the potential to influence HAI rates and weaken findings.64 Well-designed interrupted time-series studies meeting strict criteria might offer a better solution.9 When applied to studies involving HAI as an outcome measure, data collection would need to take place for 12 months before intervention and again afterwards to allow for seasonal trends. Overcoming secular trends in studies lacking controls would never be entirely possible although qualitative data and routinely collected audit data for the extensive range of variables influencing hand hygiene compliance and HAI rates (e.g. policy initiatives, media storms, long-term trends in the epidemiology of HAI and the introduction of new drugs and therapies) could be taken into consideration. Arguably such studies would be better attempted outside the UK, since there have been excessively rapid changes in health policy directed towards reducing HAI, accompanied by increasing media attention.
Interventions Our review raises questions about the adequacy of interventions intended to achieve improved hand hygiene compliance. This is supported by one of the studies we reviewed,11 while in the other study failure to improve hand hygiene compliance can be attributed to inability to implement an otherwise well-designed, educationally sound programme.10 In future studies the nature and content of the intervention requires consideration in addition to its length. Older change management literature suggested that multifaceted campaigns are most effective,5 but this view has now been challenged.65 According to more recent evidence, interventions previously thought to be ineffective such as education are modestly successful.65,66 Audit with performance feedback is less successful than educational intervention.65 Few existing hand hygiene intervention studies described the intervention adequately. Authors used the terms ‘education’ and ‘training’ interchangeably or were confused about their meanings. The purpose of education is to promote intellectual curiosity, development (personally,
199 professionally and academically) and encourage the aspiration which must underpin service transformation.67 Training is a narrower endeavour, promoting discipline, but encouraging rigidity and inhibiting development.67 The interventions described by Gould and Chamberlain and Huang et al. fulfil the criteria of education; however, the Geneva initiative, and others that it has stimulated such as the CleanYourHandsCampaign in England and Wales, represent training because of their heavy reliance on audit with performance feedback as the drivers for behaviour change.6,10,11,68 Educational initiatives are resource intensive and expensive, but if well-designed and well-implemented, have the potential to effect sustainable change. Training is less expensive but its effects are likely to be shortlived and influenced by staff turnover and shortage, because health workers are coached to undertake a repetitive set of activities rather than problemsolve. Future ventures to increase hand hygiene compliance should state explicitly whether they are employing education or training, with underlying rationale. This would pave the way towards the development of strong theoretical frameworks, which are essential for the external validity, and generalisability of study findings. Descriptions of educational interventions should include: rationale for choice of educational approach and venue; who delivered the education and their preparation; programme content; numbers of health workers attending; evaluation; changes necessary to the planned programme; and their impact. Studies geared towards training that involve audit with performance feedback should indicate the nature of the audit tool and how feedback was provided. In conclusion, initiatives to enhance hand hygiene compliance lack rigour. Existing study designs are inherently weak. Possible alternatives are insufficiently robust because the many contextual variables likely to influence hand hygiene compliance and HAI rates are not amenable to control. Future studies must combine the expertise of clinical scientists with that of behavioural scientists able to include qualitative analysis of contextual factors into the analysis of findings. Interventions intended to improve compliance are poorly articulated and not grounded in theoretical frameworks necessary for external validity. Conflict of interest statement D.J.G. co-authored one of the studies included in this review. Funding sources None.
200
Appendix 1: Analysis of the research design of excluded studies The studies were summarised by annotating the data collection points as recommended by Cook and Campbell.8 O, data collection point; X, the intervention.
Studies with few/no baseline data Maury et al.17: X O1 Panhotra et al.18: X O1 O2 O3 On Thomas et al.19: O1 X1 X2 O2 (baseline data were collected in one of the participating units only)
Uncontrolled before-and-after study designs Baker20: O1 X O2, O3 Berg et al.21: O1 X O2 Brown et al.22: O1 X O2 Coignard et al.23: O1 X O2 Creedon24: O1 X O2 Diekema et al.25: O1 X O2 O3 O4 Dorsey et al.26: O1 X O2 Dubbert et al.27: O1 X O2 Earl et al.28: O1 X O2 O3 Graham29: O1 X O2 Hughes et al.30: O1 X O2 Lam et al.31: O1 X O2 Marena et al.32: O1 X O2 McGuckin et al.33: O1 X O2 Muto et al.34: O1 X O2 Prieto and Macleod-Clark et al.35: O1 X O2 O3 Salemi et al.36: O1 X O2 O3 O4 O5 Shaw and Tanner et al.37: O1 X O2 Van de Mortel and Heyman38: O1 X O2 Van de Mortel et al.39: O1 X O2 O3 Won et al.40: O1 X O2 O3
Interrupted time-series designs (less than three data entry points before/after each intervention Conly et al.41: O1 O2 X O3 O4 Donowitz42: O1 O2 X O3 O4 Raju and Kobler43: O1 O2 X O3 O4 O5
Uncontrolled before-and-after study designs with sequential data collection Pittet et al.6 with Hugonnet et al.7 comprising a unit: O1 X O2 O3 X2 X3 O4 Xn On Avila-Aguero et al.44: O1 X1 O2 X2 O3 O4
D.J. Gould et al. Bischoff et al.45: O1 X2 O2 X2 X3 04 Harbarth et al.46: O1 X1 O2 X2 O3 X3 O4 Khatib et al.47: O1 X1 O2a O2b O2c X2 O3a O3b O3c O3d McGuckin et al.48: O1 X1 O2 X2 O3 O4 Rosenthal et al.49: O1 X1 O2 X2 O3 Swaboda et al.50: O1 X1 O2 X2 O3 Simmons et al.51: O1 X1 O2 Xn Tibbals et al.52: O1 X1 O2 X2 O3 Whitby and McLaws53: O1 X1 O2 X2 O3
Controlled before-and-after study designs with one intervention and one control site Mayer et al.54 Larson et al.55 Larson et al.56 Larson et al.57 Bittner et al.58 Colombo et al.59
References 1. National Audit Office. The management and control of hospital-acquired infection in acute NHS trusts in England. London: National Audit Office; 1998. 2. Zoutman DE, Ford DB, Bryce E, et al. The state of infection surveillance and control in Canadian acute care hospitals. Am J Infect Control 2003;31:266e273. 3. Teare EL, Cookson B, French G, et al. Leader. UK handwashing initiative. J Hosp Infect 1999;43:1e3. 4. Naikoba S, Hayward A. The effectiveness of interventions aimed at increasing handwashing in healthcare workers e a systematic review. J Hosp Infect 2001;47:173e180. 5. National Health Service Centre for Research and Development. Getting evidence into practice 1999;5:1e15. 6. Pittet D, Hugonnet S, Mourouga P, et al. Effectiveness of a hospital-wide programme to improve compliance with hand hygiene. Lancet 2000;356:1307e1312. 7. Hugonnet S, Pernager TV, Pittet D. Alcohol-based handrub improves compliance with hand hygiene in intensive care units. Arch Intern Med 2002;162:1037e1043. 8. Cook CT, Campbell DT. Quasi experimentation. Boston: Houghton Mifflin; 1979. 9. Ramsay CR, Matowe L, Grilli R, et al. Interrupted time series designs in health technology assessment: lessons from two systematic reviews of behavior change strategies. Int J Technol Assess Health Care 2003;18:613e 623. 10. Gould DJ, Chamberlain A. The use of a ward-based educational package to enhance nurses’ compliance with infection control procedures. J Clin Nurs 1997;6:55e67. 11. Huang J, Jiang D, Wang X. Changing knowledge, behavior and practice related to universal precautions among hospital nurses in China. J Contin Educ Nurs 2002;35:217e224. 12. Larson E. APIC Guidelines for hand hygiene and hand antisepsis in health care settings. Am J Infect Control 1995; 23:251e269. 13. Roethlisberger FJ, Dickson WJ. Management and the worker. Cambridge, MA: Harvard University Press; 1939. 14. Moongtui W. Compliance with components of universal precautions guidelines: using peer feedback program as
Hand hygiene compliance in patient care
15.
16. 17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30. 31.
32.
33.
34.
a cue to action among health care workers in Thailand. PhD Dissertation, University of Alabama, Birmingham, 1999. Moongtui W, Gauthier D, Turner J. Using peer feedback to improve handwashing and glove usage among Thai health care workers. Am J Infect Control 2000;28:365e369. Moore M, Abbott NR. Can handwashing practice be changed? Am J Nurs 1980;80:80. Maury E, Alzieu M, Baudel JL, et al. Availability of an alcohol solution can improve hand disinfection compliance in an intensive care unit. Am J Respir Crit Care Med 2000;162: 324e332. Panhotra BR, Saxena AK, Al-Arabi AM. The effect of a continuous educational program on handwashing compliance among healthcare workers in an intensive care unit. Br J Infect Control 2004;5:15e18. Thomas M, Gilespie W, Krauss J, et al. Focus group data as a tool in assessing effectiveness of a hand hygiene campaign. Am J Infect Control 2005;33:368e373. Baker DJ. Effects of a video based staff training with manager led exercises in residential support. Ment Retard 1998; 36:198e204. Berg DE, Hershaw RC, Claudia AR, Ramirez CA, Weinstein RA. Control of nosocomial infections in an ICU in Guatemala City. Clin Infect Dis 1995;21:588e593. Brown SM, Lubimova AV, Khrustalyeva NM, et al. Use of an alcohol-based hand rub and quality improvement interventions to improve hand hygiene in a Russian intensive care unit. Infect Control Hosp Epidemiol 2003;24:172e179. Coignard B, Grandbastien B, Berrouane Y, et al. Handwashing quality impact of a special program. Infect Control Hosp Epidemiol 1998;19:510e513. Creedon SA. Health care workers’ hand decontamination practices: compliance with recommended guidelines. J Adv Nurs 2005;51:208e216. Diekema DJ, Schult S, Albanese MA, Doebbeling BN. Universal precautions training of preclinical students: impact on knowledge, attitudes and compliance. Prev Med 1995;24: 580e585. Dorsey ST, Cydulka RK, Emerman CL. Is hand washing teachable? Failure to improve hand washing behaviour in an urban emergency department. J Academic Emerg Care 1996;3:360e365. Dubbert PM, Dolce J, Richter W, Miller M, Chapman SW. Increasing ICU staff hand washing: effects of education and group feedback. Infect Control Hosp Epidemiol 1990; 11:191e193. Earl ML, Jackson M, Rickman LS. Improved rates of compliance with hand antisepsis guidelines: a three-phase observational study. Am J Nur 2001;101:26e33. Graham M. Frequency and duration of hand washing in an intensive care unit. Am J Infect Control 1990;18: 77e81. Hughes WT, Williams B, Pearson T. The nosocomial colonisation of T-Bear. Infect Control 1986;7:495e5000. Lam BC, Lee J, Lau YL. Hand hygiene in a neonatal intensive care unit: a multimodal intervention and impact on nosocomial infection. Pediatrics 2004;114:565e571. Marena C, Lodola L, Zecca M, et al. Assessment of handwashing practices with chemical and microbiologic methods: preliminary results from a prospective crossover study. Am J Infect Control 2002;30:334e340. McGuckin M, Waterman R, Porten L, et al. Patient education model for increasing hand washing compliance. Am J Infect Control 1999;27:309e314. Muto CA, Sistrom MG, Farr BM. Hand hygiene rates unchanged by installation of dispensers of a rapidly acting hand antiseptic. Am J Infect Control 2000;28:273e276.
201 35. Prieto J, Macleod-Clark J. Contact precautions for Clostridium difficile and methicillin-resistant Staphylococcus aureus (MRSA). J Res Nurs 2005;10:511e526. 36. Salemi C, Canola T, Eck E. Hand washing and physicians: how to get them to together. Infect Control Hosp Epidemiol 2002;23:32e35. 37. Shaw S, Tanner J. Hand washing practices on a neonatal unit: the influence of teaching sessions. J Neonatal Nurs 2003;9:162e166. 38. Van de Mortel T, Heyman L. Performance feedback increases the incidence of hand washing by staff following patient contact in an ICU. Aust Crit Care 1995;8:8e13. 39. Van de Mortel T, Bourke R, Phillips I, et al. Maximising handwashing rates in the critical care unit through yearly performance feedback. Aust Crit Care 2000;13:91e95. 40. Won SP, Chou HC, Hsiah WS, et al. Hand washing program for the prevention of nosocomial infections in a neonatal intensive care unit. Infect Control Hosp Epidemiol 2004; 25:742e746. 41. Conly JM, Hills S, Ross J, Lertzman J, Thomas LJ. Hand washing practices in an ICU: the effects of an educational program and its relationship to infection rates. Am J Infect Control 1989;17:330e339. 42. Donowitz LG. Failure of the overgown to prevent nosocomial infection in paediatric ICU. Pediatrics 1986;77: 35e38. 43. Raju NK, Kobler C. Improving hand hygiene habits in the newborn nurseries. Am J Med Sci 1991;302:355e358. 44. Avila-Aguero M, Umalha MA, Jiminez AL. Handwashing practices in tertiary-care, pediatric hospital and the effect on an educational program. Clin Perform Qual Health Care 1998;6:70e72. 45. Bischoff WE, Reynolds TM, Curtis N, Sessler CN, Edward MB, Wenzel RP. Handwashing compliance by health care workers: the impact of introducing accessible alcohol-based hand antiseptic. Arch Int Med 2000;160:1017e1021. 46. Harbarth S, Pittet D, Grady L, et al. Interventional study to evaluate the impact of an alcohol based hand gel in improving hand hygiene compliance. Pediatr Infect Dis J 2002;21: 496e497. 47. Khatib M, Jamaladdine G, Abdullah A, Ibrahim Y. Handwashing and use of gloves while managing patients receiving mechanical ventilation in the ICU. Chest 1999;116: 172e175. 48. McGuckin M, Taylor A, Martin V, et al. Evaluation of a patient education model for increasing hand hygiene compliance in an inpatient rehabilitation unit. Am J Infect Control 2004;32:235e238. 49. Rosenthal VD, McCormick RD, Guzman S, Villamayor C, Orellano PW. Effect of education and performance feedback on handwashing: the benefit of administrative support in Argentinean hospitals. Am J Infect Control 2003;31: 85e92. 50. Swaboda SM, Earsing K, Strauss K, Lane S, Lipsett PA. Electronic monitoring and voice prompts to improve hand hygiene and decrease nosocomial infections in critical care. Crit Care Med 2004;32:358e363. 51. Simmons B, Bryant J, Neiman R, Spencer L, Arheart K. The role of hand washing in preventing endemic ICU infections. Infect Control Hosp Epidemiol 1990;11:589e594. 52. Tibbals J. Teaching hospital medical staff to hand wash. Med J Aust 1996;164:395e398. 53. Whitby M, McLaws ML. Handwashing in healthcare workers: accessibility of sink location does not improve compliance. J Hosp Infect 2004;58:247e253. 54. Mayer JA, Dubbert PM, Miller M, Burkett P, Chapman S. Increasing handwashing in an ICU. Infect Control 1986;7:259e262.
202 55. Larson E, McGeer A, Quaraishi ZAE, et al. Effect of an automated sink on hand washing practices and attitudes in high risk units. Infect Control Hosp Epidemiol 1991;12: 422e428. 56. Larson E, Bryan JL, Adler LM, Lee M, Blane C. Multifaceted approach to changing hand washing behavior. Am J Infect Control 1997;25:3e10. 57. Larson EL, Early E, Cloonan P, Surgue S, Parides M. An organizational climate intervention associated with handwashing and decreased nosocomial infections. Behav Med 2000;26:14e22. 58. Bittner MJ, Rich EC, Turner PD, Arnold WH. Limited impact of sustained simple feedback based on soap and paper towel consumption on the frequency of hand washing in an adult intensive care unit. Infect Control Hosp Epidemiol 2002;23:120e126. 59. Colombo C, Giger H, Grote C, et al. Impact of teaching interventions on nurse compliance with hand disinfection. J Hosp Infect 2002;51:69e72. 60. Gould DJ, Chudleigh J, Drey NS, et al. Measuring handwashing performance in health service audits and research studies. J Hosp Infect 2007;66:109e115.
D.J. Gould et al. 61. Medical Research Council. Cluster randomised trials: methodological and ethical considerations. In: London: MRC Clinical Trials Series 2002. 62. Donner A, Klar N. Pitfalls in cluster randomization trials. Am J Public Health 2004;94:416e422. 63. Kerry SM, Bland JM. Statistics notes: sample size in cluster randomisation. Br Med J 1998;316:549e550. 64. Togerson DJ. Contamination in trials: is cluster randomisation the answer? Br Med J 2001;322:355e357. 65. Jamtvedt G, Young JM, Kristofferson DT, et al. Audit and feedback: effects on professional practice and health outcomes [Review]. Cochrane Library 2007;(1):1e57. 66. Grimshaw J, Eccles M, Tetroe J. Implementing clinical guidelines: current evidence and future implications. J Contin Educ Health Prof 2004;24:S31eS37. 67. Tosey P, McNair S. Work-related learning. In: Jarvis P, editor. The age of learning: education and the learning society. London: Kogan Page; 2001. p. 95e108. 68. Storr J. The effectiveness of the national cleanyourhandscampaign. Nurs Times 2005;101:50e51.