Evidence Based Project

PSYCHOLOGY, HEALTH & MEDICINE, VOL. 7, NO. 3, 2002

Explaining hand hygiene practice: an extended application of the Theory of Planned Behaviour

E. A. JENNER,1 P. W. B. WATSON,1 L. MILLER,1 F. JONES1 & G. M. SCOTT2 1University of Hertfordshire, HatŽ eld & 2University College Hospitals, London, UK

Abstract Health care workers’ hand hygiene practice is universally sub-optimal. This contributes to the 8% prevalence rate of hospital-acquired infection which is currently costing the National Health Service in England nearly £1 billion per annum. It is estimated that about 30% of hospital-acquired infections could be prevented if health care workers adhered to hand hygiene guidelines. The aim of the study was to identify psychological constructs predictive of health care workers’ hand hygiene behaviour in order to determine ways to improve practice. We used a cross-sectional survey of 104 hospital-based health care workers. Data were analyzed through hierarchical logistic regression. The model correctly classiŽ ed 79% of cases in intention to perform appropriate hand hygiene and 87% of self-reported hand hygiene behaviour. Attitudes and personal responsibility were signiŽ cant predictors of intention, whilst perceived behavioural control and intention were signiŽ cant predictors of behaviour. The theoretical framework shows where future interventions to improve hand hygiene practice should be targeted.

Introduction

Hand hygiene is the single most important behaviour to prevent cross-infection (Black et al., 1981; Conly et al., 1989; Khan, 1982; Mortimer et al., 1966; Semmelweis, 1861; Shahid et al., 1996; Stanton & Clemens, 1987), yet many observational studies have shown that health care workers’ adherence to hand hygiene standards is universally sub-optimal. For example, Tibballs (1996) found that doctors washed their hands on only 8.6% (range 0–33%) of the occasions when it was deemed appropriate.

It is not only doctors, however, who have been found to be non-adherent but also nurses (Albert & Condie, 1981; Bartzokas et al., 1994; Gould, 1994; Gould et al., 1996), dentists and dental students (Porter et al., 1995), physiotherapists and radiographers (van de Mortel & Heyman, 1995), respiratory therapists, electrocardiographic and radiographic technicians (Wurtz et al., 1994) and occupational therapists (Marcil, 1993). As Goldmann and Larson (1992) point out, ‘Experts in infection control coax, cajole, threaten and plead, but still their colleagues neglect to wash their hands’ (p. 120). This failing contributes to approximately

Address for correspondence : Elizabeth A. Jenner, Department of Post-Registration Nursing, University of Hertfordshire, College Lane, HatŽ eld, Hertfordshire, AL10 9AB, UK. Tel: 1 44(01707) 284429; Fax: 1 44(01707) 284954; E-mail: [email protected]

ISSN 1354-8506 print/ISSN 1465-3966 online/02/030311-16 Ó Taylor & Francis Ltd DOI: 10.1080/13548500220139412

312 E. A. JENNER ET AL.

one in every 12 patients acquiring an infection as a direct result of their hospitalization (Emmerson et al., 1996). The cost to the National Health Service is nearly £1 billion pounds per annum (Plowman et al., 1999). However, it is not only the Ž nancial implications that are important but also the physical and psychological costs endured by patients and their families. Whilst the eradication of all hospital-acquired infections is an unrealistic goal, there is a commonly held view that perhaps 30% could be prevented (DOH/PHLS, 1995), leaving what Ayliffe (1986) termed ‘the irreducible minimum’. The Chief Medical OfŽ cer has recently issued guidance on how to tackle the growing problem of hospital care-associated infection (CMO, 2002). Given the relationship between hand washing and the prevention of cross-infection, increasing adherence is one obvious solution.

Reasons for non-adherence

Some reasons for health care workers’ non-adherence to hand hygiene guidelines have been identiŽ ed. Various studies have reported that environmental factors can effectively constitute barriers to performing the behaviour, such as harsh hand washing agents (Zimakoff et al., 1992) and paper towels (Heenan, 1992) and inaccessibility or insufŽ cient numbers of sinks (Kaplan & McGuckin, 1986). Another barrier is health care workers’ perceptions about a lack of time to wash their hands. The ideal duration for hand washing is not known and it is worth noting the difference in the lengths of time promulgated by various countries. For example, in the UK, the technique most commonly advocated takes 60 seconds and is based on a test procedure described by Ayliffe et al. (1978). In the USA, however, the guidelines produced by the Centers for Disease Control recommend that this behaviour can be performed in ten seconds (Garner & Favero, 1985). However, despite this quicker technique, health care workers still claim that they do not have time (Larson & Killien, 1982). Time constraints therefore would appear to be an important consideration, and Voss and Widmer (1997) ask whether we can afford 100% compliance. Using a mathematical model, they calculated that in a 14-bed intensive therapy unit, with 12 staff each working eight hours, it would take 16 hours, or two full-time nurse equivalents a day, to achieve 100% compliance with hand washing. Weeks (1999) estimates that by not washing his hands between each of the 60 ‘touch’ contacts he has every day with obstetric patients, he saves on average one or two hours, equivalent to 15% extra stafŽ ng that would be needed to cover the extra time.

In an attempt to address health care workers’ perceptions about a lack of time, several types of alcohol-based products have now been marketed, including wipes (Butz et al., 1990) and gel (Newman & Seitz, 1990), in addition to hand rubs (Mackintosh & Hoffman, 1984). These can be applied and rubbed to dryness in Ž ve seconds whilst moving between patients. However, such products should only be used for decontaminating hands that are not visibly soiled, hence hand washing is still indicated under certain circumstances. Voss and Widmer (1997) calculated that in an Intensive therapy unit, if hands were decontaminated with alcohol dispensed from containers on every bedside locker, it would take only four hours per day to achieve 100% compliance. It was to be hoped that the introduction of alcohol-based products would improve adherence, but unfortunately Gould (1994) found this not to be the case. It would appear therefore that ‘the primary problem with hand hygiene is not a paucity of good products, but rather the laxity of practice’ (Larson, 1995 p. 259).

Improving adherence

Numerous studies have been conducted in attempts to improve adherence to hand hygiene. Interventions aimed at addressing environmental barriers include providing an emollient soap

EXPLAINING HAND HYGIENE PRACTICE 313

(Kolari et al., 1989; Mayer et al., 1986) and making hand washing easier by, for example, increasing the ratio of sinks to beds (Kaplan & McGuckin, 1986) or introducing automated hand washing machines (Wurtz et al., 1994).

Studies also have been conducted to teach health care workers about the importance of hand washing. Interventions have included the use of educational and feedback strategies (Dubbert et al., 1990; Williams, 1987) and providing cues to action in the form of posters (Lohr et al., 1991). However, while numerous interventions to increase adherence have been designed and implemented, few have been successful in bringing about behavioural change. One possible reason for the failure of educational interventions may be explained by the tendency to assume a relationship between knowledge acquisition and subsequent behaviour change, when in fact this may not be the case.

This misconception has been well illustrated by several researchers. For example, Williams (1987) found that an increase in knowledge about hand hygiene did not translate into improved hand washing practice. Likewise, Alvaran et al. (1994) noted that neither level of knowledge nor positive opinion about the value of hand washing was associated with self-reported increases in hand washing practices. Similar Ž ndings have been reported by Gruber et al. (1989), who compared subjects’ knowledge scores with their scores for implementation of ‘universal precautions’ (a set of infection control practices designed to protect health care workers from contracting blood-borne virus infections through their work). They found that those with higher knowledge in fact had lower practice scores (r 5 2 0.12). Therefore, education alone is unlikely to impact upon hand washing behaviour. Pittet et al. (2000) have recently shown that a combination of strategic interventions is necessary to increase adherence to hand hygiene.

Co-workers can be one source of in uence over an individual’s practice. For example, Jenner et al. (2000) found that exposure to sub-optimal practice had an adverse effect on student nurses’ attitudes to hand hygiene, with attitudes declining signiŽ cantly throughout their three-year training programme. Similarly, Larson et al. (1986) showed that junior doctors may acquire poor hand washing practices from their peers; in contrast Larson and Larson (1983) reported that junior doctors’ hand washing practice improved when a consultant set an example. When attempting to improve adherence, it is clearly important therefore to target consultants who are considered by some to be the worst culprits (Araf et al., 1999) while at the same time are best positioned to act as role models.

Another strategy is to enhance an individual’s sense of ownership concerning health- related behaviours and outcomes. This can be thought of in terms of identifying the locus of responsibility for maintaining both one’s own health and the health of others. This has not yet been explored with regard to hand hygiene behaviour, but Rothman et al. (1993) explored the comparative usefulness of ‘internally-orientated’ messages versus ‘externally-ori- entated’ or ‘information-only’ messages to encourage women to attend mammography screening. Internally-oriented messages alone were successful in encouraging attendance. The relevance of this construct to hand hygiene practice would therefore seem worthy of exploration.

The need to explore constructs that are predictive of hand hygiene behaviour has been highlighted by Kretzer and Larson (1998). The primary aim of this study, therefore, was to develop a theoretical framework in order to identify perceived cognitive and physical factors that may explain health care workers’ hand hygiene behaviour. It was postulated that the Ž ndings would provide direction for the design of theoretically driven interventions to improve adherence to hand hygiene guidelines.

314 E. A. JENNER ET AL.

Theoretical framework

Theoretical models developed to explain the in uence of beliefs on individuals’ actions include the Health Belief Model (HBM; Rosenstock, 1966), the Theory of Reasoned Action (TRA; Fishbein, 1967) and the Theory of Planned Behaviour (TPB; Ajzen, 1985). The original aim of the HBM was to explain preventative health behaviours. The model consists of Ž ve core constructs, namely perceived susceptibility, perceived severity, costs, beneŽ ts and cues to action. The TRA was developed to explore the relationship between attitudes, subjective norms and intention to perform a behaviour (Fishbein, 1967). The TRA was later expanded by the addition of the construct of perceived behavioural control to form the TPB (Ajzen, 1985). The TPB consists of constructs believed to predict intention to perform behaviour. These are attitudes which can be deŽ ned as the extent to which individuals have a favourable or unfavourable evaluation of the behaviour; subjective norms which refer to the social pressure individuals perceive themselves to be under to perform a behaviour; and perceived behavioural control, which can be deŽ ned as the extent to which individuals believe that they have adequate resources, physical or cognitive, to perform the behaviour.

The models described above have been utilized in previous research to examine a wide variety of health behaviours such as exercise (Norman & Smith, 1995) and the uptake of screening for cervical cancer (Hill et al., 1985). They appear to have useful levels of predictive validity in terms of both explaining behaviour and providing information for the construction of theoretically-based interventions. However, to our knowledge, they have not previously been used to examine hand hygiene behaviour.

In this study, the TPB formed the foundation of the theoretical framework. The theory dictates that attitudes, subjective norms and perceived behavioural control are predictive of behavioural intention, while behaviour is predicted by intentions and perceived behavioural control. However, given that the construct of perceived behavioural control would not identify speciŽ c obstacles that may hinder health care workers’ hand hygiene behaviour, the model included the construct of barriers drawn from the HBM (Rosenstock, 1966). These were based on existing literature related to poor hand hygiene adherence: acceptability of hand hygiene agents (Zimakoff et al., 1992), time availability (Larson & Killien, 1982), satisfaction with paper towels (Heenan, 1992) and the number/location of sinks (Kaplan & McGuckin, 1986). It was postulated that these would impact upon intention and behaviour. The model was extended further by the addition of the construct of personal responsibility, which was postulated to be a predictor of intention. The proposed predictive model is shown in Figure 1.

Method

Design and sample

A cross-sectional survey was used. A total of 304 questionnaires were given to the infection control nurse and senior nurse managers in a central London teaching hospital who agreed to distribute them to health care workers who were ward based. Anonymity and conŽ dentiality were ensured in order to enhance the response rate (Oppenheim, 1992) and to overcome possible apprehensions which might affect the honesty of responses. Data collection took place during December 1999 through January 2000.

One hundred and four questionnaires were returned via internal post to the infection control nurse, representing a completed return rate of 34%. The sample included registered nurses, n 5 76 (73%); therapists, n 5 17 (16%); health care assistants, n 5 4 (4%); and

EXPLAINING HAND HYGIENE PRACTICE 315

FIG. 1. Proposed predictive model.

doctors, n 5 3 (3%). Four participants failed to disclose their occupation. Eighty-eight (85%) of the respondents were female. Five participants did not indicate their gender.

Measures

The target behaviour of hand washing was deŽ ned in accordance with national guidelines (UK Health Departments, 1998). These specify that hands should be washed before and after contact with each patient, and before putting on and removing gloves.

A self-report questionnaire measuring attitudes, subjective norms, perceived behavioural control, intention, behaviour, personal responsibility and barriers was constructed. With the exception of behaviour, the TPB constructs were measured using multiple items as recom- mended by Conner and Norman (1995), whereas personal responsibility and barriers were measured by single items. The response format for all items was a seven-point differential scale. Some items were reverse scored in order to reduce the probability of a positive response bias (Howell, 1992).

Attitudes were measured through four items and had an overall internal consistency value of 0.77 (Cronbach’s alpha), e.g. ‘Washing my hands before and after every patient contact is …’ (1 5 ‘not important’, 7 5 ‘very important’). Subjective norms were measured through two items (Cronbach’s alpha, 0.71; Spearman’s Rho, r 5 0.55), e.g. ‘I feel under social pressure from other health care workers to wash my hands’ (1 5 ‘strongly disagree’, 7 5 ‘strongly agree’). Perceived behavioural control was measured through two items (Cronbach’s alpha, 0.83; Spearman’s Rho, r 5 0.73), e.g. ‘If I wanted to, I could easily wash my hands before and after each patient contact’ (1 5 ‘unlikely’, 7 5 ‘likely’). The construct of intention was measured through four items (Cronbach’s alpha, 0.78), e.g. ‘I always intend to wash my hands before and after contact with each patient’ (1 5 ‘strongly disagree’, 7 5 ‘strongly agree’). For all multiple-item constructs, mean scores were calculated. The single item used

316 E. A. JENNER ET AL.

to measure behaviour was ‘I always wash my hands before and after contact with each patient’ (1 5 ‘strongly disagree’, 7 5 ‘strongly agree’). The single item used to measure the construct of personal responsibility was ‘I believe that I have a role to play in reducing the risk of cross-infection by washing my hands before and after every patient contact’, (1 5 ‘strongly disagree’ and 7 5 ‘strongly agree’).

The Ž ve barrier items were all measured separately on seven-point scales ranging from ‘strongly disagree’ to ‘strongly agree’. Scoring was reversed, where appropriate, so that a high value represented a highly rated barrier. For example, if an individual strongly agreed that a lack of time hindered their practice, a value of seven would be awarded.

As each barrier item represented a separate potential problem, it was not appropriate to calculate a collective barrier internal reliability value. The acceptability of hand hygiene agents was measured by the item, ‘There is always a hand hygiene product available that will not damage my skin’. Time availabilit y was measured by the item, ‘Sometimes I don’t wash my hands, because I don’t have the time’. Satisfaction with paper towels was measured by the item, ‘I believe that the paper towels available are satisfactory for hand drying’.

Two of the barriers which were measured by the items ‘The number of sinks allows me to wash my hands with ease when I need to’ and ‘The location of the sinks allows me to wash my hands with ease when I need to’ were highly inter-correlated (Cronbach’s alpha 0.87; Spearman’s Rho 5 0.77). These were therefore combined to produce a composite score.

Data analysis

Non-parametric bivariate correlations were calculated in order to establish whether any of the predictive constructs were related to the dependent variables (intention and behaviour). When using the TPB as a theoretical framework, it is recommend that two separate analyses should be conducted to examine the extent to which the constructs predict variance in both behavioural intentions and actual behaviour (Hankins et al., 2000). Hence, two separate hierarchical logistic regression analyses were performed. All of the independent predictors were added into the equations as categorical variables. In order to explain intention, constructs from the TPB were added into the equation on the Ž rst step, followed by personal responsibility and barriers which were added on the second and third steps, respectively. In order to explain hand hygiene behaviour, the construct of intention was added into the equation on the Ž rst step, followed by perceived behavioural control and barriers on the second and third steps, respectively.

Results

Descriptive Ž ndings of the predictive constructs are presented in Table 1. The key points to note are that both the dependent variables and several independent variables were positively skewed with relatively low variance. Therefore, appropriate categorization processes were undertaken (for rationale see Table 2). The responses to subjective norms and the three remaining barriers did not warrant categorization.

Notwithstanding the positive skew of the dependent variables, it should be noted that only a third of respondents (34/104) gave a composite score of seven in response to the items ‘I always like/expect/intend/want to wash my hands’ and only a quarter (26/104) strongly agreed with the statement ‘I always wash my hands before and after contact with each patient’.

EXPLAINING HAND HYGIENE PRACTICE 317

Table 1. Descriptive statistics

Predictors Range Range Mean SD (min) (max)

Intention 2.25 7.00 6.10 1.07 Behaviour 1.00 7.00 5.34 1.67 Attitude 4.25 7.00 6.59 0.69 Subjective norms 1.00 7.00 2.88 1.72 Perceived behavioural control 1.00 7.00 5.65 1.51 Barrier 1—Acceptability of hand hygiene agents 1.00 7.00 2.87 2.02 Barrier 2—Time availability 1.00 7.00 3.17 2.03 Barrier 3—Satisfaction with paper towels 1.00 7.00 3.49 2.18 Barrier 4—Number/location of sinks 1.00 7.00 2.75 1.83 Personal responsibility 1.00 7.00 6.55 1.03

Correlation analysis

Table 3 shows that attitudes, personal responsibility and behaviour were all signiŽ cantly correlated with intention. Attitudes, perceived behavioural control, intention, personal re- sponsibility, time availability and number and location of sinks were all signiŽ cantly corre- lated with behaviour. These two barriers were also signiŽ cantly correlated with perceived behavioural control.

Logistic regression analyses

Explaining intention. The TPB predictors successfully produced a model explaining inten- tion to perform hand hygiene, with 70% of cases correctly classiŽ ed. Personal responsibility contributed a further 2% to the correct classiŽ cation, and barriers a further 7% (Table 4). Correct classiŽ cation rates were 71% for ‘low’ intenders and 84% for ‘high’ intenders. The overall correct classiŽ cation rate was 79%, thus the model correctly predicted the majority of variance in intention to perform hand hygiene behaviour. Therefore, the overall predictive validity of the model was high (c2 57.95, df 5 35, p , 0.01).

Both attitudes (p , 0.05) and personal responsibility (p , 0.01) were signiŽ cant predic-

Table 2. Categorization of dependent and independent variables

Variable Score Label Numbers in group (%)

Intention to wash hands above 6 ‘high’ 61 6 and below ‘low’ 39

Behaviour 6 and above ‘high’ 63 below 6 ‘low’ 37

Attitudes 7 ‘high’ less than 7 ‘low’

Perceived behavioural control 6.5–7 ‘high’ 1–6 ‘low’

Personal responsibility 7 ‘high’ less than 7 ‘low’

Number/location of sinks 4–7 ‘high’ 2–3.5 ‘medium’ 1–1.5 ‘low’

318 E. A. JENNER ET AL.

T a

b le

3. S

im pl

e (S

pe ar

m an

’s R

ho )

co rr

el at

io ns

be tw

ee n

pr ed

ic to

rs (n

5 10

4)

A tt

it u

d e

S u

bj ec

ti ve

P er

ce iv

ed In

te n

ti on

B ar

ri er

1— B

ar ri

er 2—

B ar

ri er

3— B

ar ri

er 4

— P

er so

n al

n or

m s

b eh

av io

u ra

l A

cc ep

ta b

ili ty

of T

im e

S at

is fa

ct io

n N

u m

be r/

re sp

on si

b ili

ty co

n tr

ol h

an d

hy gi

en e

av ai

la bi

lit y

w it

h pa

p er

lo ca

ti on

ag en

ts to

w el

s of

si n

ks

A tt

it u

d e

S u

b je

ct iv

e 2

0. 0

1 n

or m

s P

er ce

iv ed

0. 1

6 2

0. 04

be h

av io

ur al

co n

tr ol

In te

n ti

on 0.

3 0*

* 0.

03 0.

11 B

ar ri

er 1

2 0.

0 9

0. 09

0. 06

0. 00

B ar

ri er

2 2

0. 1

4 2

0. 09

2 0.

37 **

2 0.

18 0.

05 B

ar ri

er 3

2 0.

0 4

0. 16

2 0.

12 0.

14 0.

24 *

2 0.

08 B

ar ri

er 4

2 0.

1 7

2 0.

02 2

0. 24

* 2

0. 11

0. 28

** 0.

11 0.

1 3

P er

so n

al 0.

4 1*

* 2

0. 02

0. 22

* 0.

42 **

0. 02

2 0.

19 *

0. 0

1 2

0 .0

8 re

sp on

si bi

lit y

B eh

av io

ur 0.

2 5*

* 0.

02 0.

35 **

0. 55

** 2

0. 01

2 0.

40 **

0. 0

3 2

0 .3

0* *

0. 36

**

** p

, 0

.0 1;

* p

, 0

.0 5.

EXPLAINING HAND HYGIENE PRACTICE 319

Table 4. Hierarchical logistic regression to predict intentions

Predictors Chi-square Correct classiŽ cation D classiŽ cation (%) (%)

TPB predictors 20.28 69.9 – Personal responsibility 29.12 71.8 1.9 Barriers 1–4 57.95 78.6 6.8

Table 5. Predictors of intention to perform hand hygiene

Predictors B1 SE1 Wald test SigniŽ cance statistic

Attitude 2 2.16 0.88 5.99 0.014* Subjective norms – – 5.41 0.943 Perceived behavioural control 0.98 0.84 1.35 0.246 Personal responsibility 2 2.65 0.95 7.76 0.005** Acceptability of hand hygiene agents – – 4.20 0.650 Time availability – – 8.33 0.215 Satisfaction with paper towels – – 9.57 0.144 Number/location of sinks – – 0.83 0.661

** p , 0.01; *p , 0.05. Note. 1 Unstandardized beta (B) values and standard error values are not given when the predictors have a degree of freedom value of more than 1.

tors of intention. However, neither subjective norms nor perceived behavioural control signiŽ cantly predicted intention (Table 5).

Explaining behaviour. Intention was a strong predictor of behaviour, with 79% of cases being correctly classiŽ ed. Perceived behavioural control did not contribute further to the expla- nation of behaviour but barriers contributed an additional 10% (Table 6). Correct classiŽ cation rates were 79% for ‘low’ hand washers and 92% for ‘high’ hand washers. The overall correct classiŽ cation rate was 87%, thus the model correctly predicted the majority of variance in performing hand hygiene behaviour. Therefore, the overall predictive validity of the model was high (c2 80.22, df 5 22, p , 0.001).

Both intention (p , 0.001) and perceived behavioural control (p , 0.05) were signiŽ cant predictors of behaviour. Number and location of sinks approached signiŽ cance (p 5 0.059) (Table 7).

A diagrammatic representation of the summative predictive model based on the Ž ndings is presented in Figure 2. This shows that behaviour is predicted by intentions and perceived

Table 6. Hierarchical logistic regression to explain behaviour

Predictors Chi-square Correct classiŽ cation D classiŽ cation (%) (%)

Intention 32.3 78.6 – Perceived behavioural control 45.2 77.7 – Barriers 1–4 80.2 87.4 9.7

320 E. A. JENNER ET AL.

Table 7. Predictors of hand hygiene behaviour

Predictors B1 SE1 Wald test SigniŽ cance statistic

Intention 2 4.53 1.15 15.58 0.000** Perceived behavioural control 2 2.58 1.04 6.10 0.014* Acceptability of hand hygiene agents – – 4.89 0.558 Time availability – – 9.78 0.134 Satisfaction with paper towels – – 4.17 0.653 Number/location of sinks – – 5.65 0.059

**p , 0.001; *p , 0.05. Note. 1Unstandardized beta. (B) values and standard error values are not given when the predictors have a degree of freedom value of more than 1.

behavioural control, whilst intention is predicted by attitudes and personal responsibility, but not subjective norms or perceived behavioural control as posited in the proposed predictive model. The effect of the two speciŽ c barriers, namely time and number and location of sinks, on behaviour may be mediated by their effect on perceived behavioural control.

Discussion

This study has demonstrated the value of using the TPB as a basis for modelling hand hygiene practice. However, the addition of the speciŽ c barriers and the construct of personal responsibility further enhanced the predictive validity of the model.

Attitudes were shown to be a signiŽ cant predictor of intention to perform hand hygiene. Although the values attached to the construct were already high, only a third of respondents

FIG. 2. Summative predictive model.

EXPLAINING HAND HYGIENE PRACTICE 321

(34/104) said they always liked/expected/intended/wanted to wash their hands before and after contact with each patient. Given this Ž nding, it is important to consider factors that can affect attitude formation. Fazio and Zanna (1981) have shown that attitudes formed through direct personal experience are better predictors of behaviour, and later work by Fazio (1986; 1990) showed that direct personal experience results in the formation of attitudes which are not only more readily accessible but are also stronger. The strength of health care workers’ attitudes to hand hygiene may therefore be in uenced by the fact that they do not see the direct effect of hand washing immediately, if at all. As Jenner et al. (1999) point out, ‘the interval between someone failing to wash their hands and a patient developing an infection makes it unlikely that the two events will be naturally associated’ (p. 95). This time lag makes it difŽ cult to reinforce by demonstration that a given action results in a predicted conse- quence. However, Bligh (1972) has shown that teaching through demonstration and experi- ential learning often results in higher information retention rates by students. Therefore, given the limited success of lectures or discussions in achieving long-term behavioural change (Courington et al., 1991; Gould et al., 1996), it may be worth adopting a more practical approach to teaching the importance and value of hand hygiene to shape students’ attitudes. Our current research addresses this point and evaluates the efŽ cacy of experiential learning in helping to maintain early levels of enthusiasm for hand hygiene practice amongst health care students. During laboratory classes students performed Ž ngertip impressions on culture plates before and after use of either soap or antiseptic to wash their hands. Students then inspected their culture plates and compared microbial growth in the absence of hand washing (‘before’) and after use of the two types of cleaning agent. The longitudinal follow-up to determine whether this intervention can serve to maintain health care students’ beliefs in the importance of hand washing is currently underway. This is timely given that the Comptroller and Auditor General (2000) has recently emphasized the need to implement education strategies to enhance the appropriate practice of hand hygiene in his independent audit on the costs of hospital-acquired infection.

Contrary to predictions arising from the theory, however, neither subjective norms nor perceived behavioural control were found to predict hand hygiene intentions. The model hypothesized that speciŽ c barriers would be predictive of behavioural intentions. Although the addition of these contributed an extra 7% to the correct classiŽ cation rate, none were signiŽ cantly correlated to or predictive of intention.

However, the additional construct of personal responsibility was shown to be a signiŽ cant predictor of intention to perform hand hygiene practice. This construct added 2% to the correct classiŽ cation rate over and above the 70% explained by the TPB constructs. Previous research has consistently shown that those who attribute their behaviour to internal causes are more inclined to adopt healthy behaviours (King, 1982). Conversely, other researchers have shown that those who attribute externally are poorer at adhering to recommendations (Davison & Valins, 1969; Storms & Nisbett, 1970). Clearly then, ways need to be found to enhance and maintain health care workers’ sense of personal responsi- bility regarding their role in the prevention of cross-infection through appropriate hand hygiene practice. Indeed, as Jenner et al., (1999) state, ‘it is part of the management ethos of the 90’s that people should “take ownership” of rules because it is perceived that they are more likely to be accepted and internalised, and therefore complied with’. They further point out that, ‘empirical-rational or normative-reeducative strategies are to be preferred to power- coercive strategies’ (p. 99).

It has been suggested that patients should be encouraged to ask their carers to wash their hands (Handwashing Liaison Group, 1999). Strategies tested in this country include asking patients to hold up yellow cards (Sen et al., 1999) and asking them to enquire verbally, ‘Did

322 E. A. JENNER ET AL.

you wash your hands?’ (McGuckin et al., 2001). However, Gillespie (2001) argues that, ‘clinical colleagues might view patient empowerment as a potentially detrimental in uence on an open and co-operative doctor–patient relationship’ (p. 298). He does, however, commend the use of signposts that continually direct and remind health care workers to decontaminate their hands. These have been used by Ndawula and Cutter (2001) to empower patients.

The relative efŽ cacy of various message-framing techniques as described by Rothman et al. (1993) should perhaps be tested. While the research reported here indicates that ‘intern- ally-oriented’ messages are likely to be most efŽ cacious, analysis of existing hand hygiene posters indicates that those based on this type of message are currently in the minority (in preparation). The new promotional material designed by Rao et al., (2002) might have greater impact if the content were framed differently. In designing future posters, appeals should be directed towards invoking health care workers’ sense of personal responsibility for their own hand hygiene practice.

The signiŽ cant predictors of hand hygiene behaviour were found to be intention and perceived behavioural control. Only a quarter (26/104) of those responding said that they always washed their hands before and after contact with each patient. The theoretical model used included four speciŽ c barriers to hand hygiene practice, which may impinge upon perceived behavioural control. Two of these, namely time availability and number and location of sinks, were signiŽ cantly correlated with both behaviour and perceived behavioural control. This suggests that the effect of these barriers on behaviour is mediated by perceived behavioural control (Figure 2) and there is clearly a need to explore the nature of the exact relationship. The direction of the correlations means that increased time and sinks made it more likely that health care workers would wash their hands. In the hospital in which this study was conducted, the infection control team had already attempted to address time constraints by ensuring that a bottle of alcohol hand gel was available on every bedside locker. This was in addition to liquid soap and two different kinds of antiseptics being available for hand washing at the sinks. Yet still it appears that perceptions about a lack of time persist even where attempts to remedy the problem have been made. The introduction of new products in the clinical setting needs to be accompanied by demonstrations to show how easy, quick and effective it is to use alcohol-based products for hand hygiene.

In relation to the number and location of sinks, the design of the wards in the hospital where this study was conducted was far from ideal. However, other researchers have reported that hand hygiene adherence was poor even where facilities for hand washing were more than satisfactory (Gould, 1994; Tibballs, 1996). Indeed, Preston et al. (1981) found no improve- ment in hand hygiene behaviour when a hospital moved to new premises with a sink provided at every bed space. Therefore, although the facilities for hand washing must obviously be provided, changing the design of the ward layout in relation to the number and location of sinks would probably have limited impact on changing behaviour.

As regards the other two predictors, neither hand hygiene agents nor paper towels were perceived as barriers to the performance of the behaviour. The reason for this may be because there was not only a choice of hand hygiene agents available, but the paper towels provided at all sinks were of the very soft variety.

Conclusions

This study has explored a theoretical framework to determine factors that predict hand hygiene. Potential targets for changing hand hygiene behaviour have been identiŽ ed so that interventions can be designed to increase health care workers’ adherence to hand hygiene guidelines. This is essential if the prevalence of hospital-acquired infections is to be reduced.

EXPLAINING HAND HYGIENE PRACTICE 323

Since it has been shown that the spread of hand-borne hospital pathogens which are present in endemic proportions can be brought under control by ‘even small increases in the frequency of effective handwashes’ (Cooper et al., 1999, p. 131), it is imperative that ways are found to increase adherence to this practice, which is the cornerstone of any infection prevention programme.

It is clear, however, that the study has some limitations. While the current cross-sectional design does not allow causation to be determined, nonetheless such exploratory work is a necessary stage in identifying factors appropriate for study in follow-up longitudinal studies. The real value of this work lies in its ability to inform the design of more rigorous research. The work described here is useful in that it suggests areas, issues and measures that researchers may wish to concentrate upon in future longitudinal studies. It is particularly useful to have such exploratory data available as a basis for grant applications to fund such resource-intensive work.

Self-report measures could be a limitation as they may result in over-estimation of intention and behaviour and hence result in data being positively skewed (Abraham et al., 1999). Indeed, this is what Tibballs (1996) found. However, the study reported here showed that a majority of health care workers admit to sub-optimal levels of hand hygiene practice. Only a quarter of respondents said that they would always wash their hands before and after contact with each patient. It is likely that the reality is even worse. This Ž nding also suggests that the sample was not biased towards more adherent health care workers.

Despite attempts to recruit staff from other disciplines, i.e. medicine and therapy services, the composition of the sample consisted almost entirely of female registered nurses. The difŽ culty in recruiting doctors to participate in studies has been reported by others (Firth-Cozens, 2001).

Two attempts were made to enhance the return rate of the questionnaire by redistribu- tion and extending the date for return. Although the sample size was sub-optimal, it was within the expected range. In an analysis of nursing studies published between the 1950s and the 1980s, Browne et al. (1984) found that the average sample size was under 100 subjects in all four decades. This has been conŽ rmed in a more recent analysis conducted by Moody et al. (1988).

Despite these limitations, this study has achieved its aims. Using an approach which integrates health behaviour theories and existing research Ž ndings, the study provides a framework which may be used as a building block for both theoretical developments and practical interventions. SpeciŽ cally, it identiŽ es three key targets: attitudes, perceived be- havioural control and personal responsibility. Pointers for future research have been pre- sented.

Acknowledgements

Thanks are extended to Jacqui Murray-Leonard, who helped with the administration of the questionnaire and all health care workers who completed it. We are also grateful to Dr Elena Kulinskaya for her statistical advice throughout the preparation of this paper.

References

ABRAHAM , C., CLIFT, S. & GRABOWSKI, P. (1999). Cognitive predictors of adherence to malaria prophylaxis regimens on return from a malarious region: a prospective study. Social Science and Medicine, 48(11), 1641–1654.

AJZEN, I. (1985). From intention to actions: a Theory of Planned Behaviour. In: J. KUHL & J. BECKHAN (Eds), Action-control: from cognition to behaviour (pp. 11–39). Berlin: Spinger-Verlag.

324 E. A. JENNER ET AL.

ALBERT, R. K. & CONDIE, F. (1981). Hand washing patterns in medical intensive care units. New England Journal of Medicine, 304(24), 1465–1466.

ALVARAN, M. S., BUTZ, A. & LARSON, E. (1994). Opinions, knowledge, and self-reported practices related to infection control among nursing personnel in long-term care settings. American Journal of Infection Control, 22(6), 367–370.

ARAF, S., BAILOOR, S., CHEUNG, I., CROUCH, J., KELLY, M., NWUME, H., THO, H. S. & SEDGWICK, J. (1999). Efforts at improving handwashing compliance should be focussed on consultants, the worst culprits. Bmj.com. 12 October.

AYLIFFE, G. A. J. (1986). Nosocomial infection: the irreducible minimum. Infection Control, 7 (Suppl. 2), 91–95. AYLIFFE, G. A. J., BABB, J. R. & QUORAISHI, A. H. (1978). A test for ‘hygienic’ hand disinfection. Journal of Clinical

Pathology, 31, 923–928. BARTZOKAS, C. A., WILLIAMS, E. E. & SLADE, P. D. (1994). A psychological approach to hospital-acquired infections.

Studies in Health and Human Sciences, 24. Edwin Mellen Press. BLACK, R. E., DYKES, A. C., ANDERSON, K. E., WELLS, J. G., SINCLAIR, S. P., GARY, G. W., HATCH, M. H. &

GANGAROSA, E. J. (1981). Handwashing to prevent diarrhea in day-care centers. American Journal of Epidemiology, 113(4), 445–451.

BLIGH, D. A. (1972). What’s the use of lectures? New Barnet, UK: D. A. & B. BLIGH. BROWN, J. S., TANNER, C. A. & PADRICK, K. P. (1984). Nursing’s search for scientiŽ c knowledge. Nursing Research, 33,

26–32. BUTZ, A. M., LAUGHON, B. E., GULLETTE, D. L. & LARSON, E. L. (1990). Alcohol-impregnated wipes as an alternative

in hand hygiene. American Journal of Infection Control, 18, 70–76. CHIEF MEDICAL OFFICER (2002). Getting ahead of the curve. A strategy for combating infectious diseases (including other

aspects of health protection). London. Department of Health. COMPTROLLER AND AUDITOR GENERAL. NATIONAL AUDIT OFFICE (2000). The management and control of hospital

acquired infection in acute NHS trusts in England. London: TSO 17 February. HC 230 Session 1999–00. CONLY J., HILL, S., ROSS, J., LERTZMAN, J. & LOUIE, T. J. (1989). Handwashing practices in an intensive care unit: the

effects of an educational program and its relationship to infection rates. American Journal Infection Control, 17(6), 330–339.

CONNER, M. & NORMAN P. (Eds) (1995). Predicting health behaviour: research and practice with social cognition models. Buckingham: Open University Press.

COOPER, B. S., MEDLEY, G. F. & SCOTT, G. M. (1999). Preliminary analysis of the transmission dynamics of nosocomial infections: stochastic and management effects. Journal of Hospital Infection, 43(2), 131–147.

COURINGTON, K. R., PATTERSON, S. L. & HOWARD, R. J. (1991). Universal precautions are not universally followed. Archives of Surgery, 126, 93–96.

DAVISON, G. C. & VALINS, S. (1969). Maintenance of self-attributed and drug-attributed behaviour change. Journal of Personality and Social Psychology, 11, 25–33.

DOH/PHLS (1995). Hospital infection control. Guidance on the control of infection in hospitals. Prepared by the Hospital Infection Working Group of the DOH and PHLS. London: PHLS.

DUBBERT, P. M., DOLCE, J., RITCHER, W., MILLER, M. & CHAPMAN, S. W. (1990). Increasing ICU staff handwashing: effects of education and group feedback. Infection Control and Hospital Epidemiology, 11, 191–193.

EMMERSON, A. M., ENSTONE, J. E., GRIFFIN, M., KELSEY, M. C. & SMYTH, E. T. M. (1996). The second national prevalence survey of infections in hospitals—overview of the results. Journal of Hospital Infection, 32, 175–190.

FAZIO, R. H. (1986). How do attitudes guide behaviour? In: R. M. SORRENTINO & E. T. HIGGINS (Eds), The handbook of motivation and cognition: foundations of social behavior (pp. 204–243). New York: Guilford Press.

FAZIO, R. H. (1990). Multiple processes by which attitudes guide behaviour: the MODE model as an integrative framework. Advances in Experimental Social Psychology, 23, 75–109.

FAZIO, R. H. & ZANNA, M. P. (1981). Direct experience and attitude-behaviour consistency. Advances in Experimental Social Psychology, 14, 161–202.

FIRTH-COZENS, J. (2001). Why I study stress in doctors. The Psychologist, 14(4), 202–203. FISHBEIN , M. (1967). Attitude and the prediction of behaviour. In: M. FISHBEIN (Ed.), Readings in attitude theory and

measurement. New York: John Wiley. GARNER, J. S. & FAVERO, M. S. (1985). Guideline for handwashing and hospital environmental control. Atlanta, GA:

Centers for Disease Control, U.S. Department of Health and Human Services. GILLESPIE , T. (2001). Patient empowerment to improve handwashing compliance by healthcare workers. Journal of

Hospital Infection, 49, 298–299. GOLDMANN, D. & LARSON, E. (1992). Handwashing and nosocomial infections. New England Journal of Medicine, 327,

120–122. GOULD, D. (1994). Nurses’ hand decontamination practice: results of a local study. Journal of Hospital Infection, 28(1),

15–30.

EXPLAINING HAND HYGIENE PRACTICE 325

GOULD, D., WILSON-BARNETT, J. & REAM, E. (1996). Nurses’ infection-control practice: hand decontamination, the use of gloves and sharp instruments. International Journal of Nursing Studies, 33(2), 143–160.

GRUBER, M., BEAVERS, F. E., JOHNSON, B., BRACKETT, M., LOPEZ, T., FELDMAN, M. J. & VENTURA, M. (1989). The relationship between knowledge about AIDS and the implementation of universal precautions by registered nurses. Clinical Nurse Specialist, 3, 182–185.

HANDWASHING LIAISON GROUP (1999) [Editorial]. Handwashing: a modest measure-with big effects. British Medical Journal, 318, 686.

HANKINS, M., FRENCH, D. & HORNE, R. (2000). Statistical guidelines for studies of the Theory of Reasoned Action and the Theory of Planned Behaviour. Psychology and Health, 15, 151–161.

HEENAN, A. (1992) Handwashing practices. Nursing Times, 88(34), 70. HILL, D., GARDENER, G. & RASSABY, J. (1985). Factors predisposing women to take precautions against breast and

cervix cancer. Journal of Applied Social Psychology, 15, 59–79. HOWELL, D. C. (1992). Statistical methods for psychology, 3rd edition. California: Duxbury Press. JENNER, E. A., MACKINTOSH, C. M. & SCOTT, G. M. (1999). Infection control—evidence into practice. Journal of

Hospital Infection, 42(2), 91–104. JENNER, E. A., WATSON, P. W. B., MILLER, L., BONIFACE, D. R. & JONES, F. (2000). A survey of student nurses’

attitudes towards hand hygiene practice. Abstract of poster presented at the British Psychological Society, Division of Health Psychology Annual Conference, University of Kent at Canterbury, 6–8 September. Proceedings of the British Psychological Society, 9(1), 68.

KAPLAN, L. M. & MCGUCKIN, M. (1986). Increasing handwashing compliance with more accessible sinks. Infection Control, 7(8), 408–410.

KHAN, M. U. (1982). Interruption of shigellosis by hand washing. Transactions of the Royal Society of Tropical Medicine and Hygiene, 76(2), 164–168.

KING, J. B. (1982). The impact of patients’ perceptions of high blood pressure on attendance at screening: An extension of the health belief model. Social Science and Medicine, 16, 1079–1092.

KOLARI, P. J., OJAJARVI, J., LAUHARANTA, J. & MAKELA, P. (1989). Cleansing of hands with emulsion—a solution to skin problems of hospital staff? Journal of Hospital Infection, 13, 377–386.

KRETZER, E. K. & LARSON, E. L. (1998). Behavioral interventions to improve infection control practices. American Journal of Infection Control, 26, 245–253.

LARSON, E. L. (1995). APIC guideline for hand washing and hand antisepsis in health care settings. American Journal of Infection Control, 23, 251–269.

LARSON, E. L. & KILLIEN, M. (1982). Factors in uencing handwashing behavior of patient care personnel. American Journal of Infection Control, 10(3), 93–99.

LARSON, E. L. & LARSON, E. B. (1983). In uence of a role model on handwashing behavior. Abstract of a paper presented at the Tenth Annual Educational Conference, Association of Practitioners in Infection Control, San Diego. American Journal of Infection Control, 11(4), 146.

LARSON, E. L., MCGINLEY, K. J., GROVE, G. L., LEYDEN, J. J. & TALBOT, G. H. (1986). Physiologic, microbiologic and seasonal effects of handwashing on the skin of health care personnel. American Journal of Infection Control, 20(2), 65–72.

LOHR, J. A., INGRAM, D. L., DUDLEY, S. M., LAWTON, E. L. & DONOWITZ, L. G. (1991). Hand washing in pediatric ambulatory settings. American Journal of Diseases in Childhood, 145(10), 1198–1199.

MACKINTOSH, C. A. & HOFFMAN, P. N. (1984). An extended model for transfer of microorganisms via the hands: differences between organisms and the effect of alcohol disinfection. Journal of Hygiene, 92, 345–355.

MARCIL, W. M. (1993). Handwashing practices among occupational therapy personnel. American Journal of Occu- pational Therapy, 47(6), 523–528.

MAYER, J. A., DUBBERT, P. M., MILLER, M., BURKETT, P. A. & CHAPMAN, S. W. (1986). Increasing handwashing in an intensive care unit. Infection Control, 7(5), 259–262.

MCGUCKIN, M., WATERMAN, R., STORR, J., BOWLER, I. C. J. W., ASHBY, M., TOPLEY, K. & PORTEN, L. (2001). Evaluation of a patient-empowering hand hygiene programme in the UK. Journal of Hospital Infection, 48, 222–227.

MOODY, L. E., WILSON, M. E., SMYTH, K., SCHWARTZ, R., TITTLE, M. & VANCOTT, M. L. (1988). Analysis of a decade of nursing practice research: 1977–1986. Nursing Research, 37, 374–379.

MORTIMER, E. A., WOLSKY, E. & HINES, D. (1966). The effect of rooming-in on the acquisition of hospital staphylococci by newborn infants. Pediatrics, 37, 605–609.

NDAWULA, E. & CUTTER, M. (2001). [Letter]. ‘Road signs’ approach to hand hygiene. Journal of Hospital Infection, 48, 242–246.

NEWMAN, J. L. & SEITZ, J. C. (1990). Intermittent use of an antimicrobial hand gel for reducing soap induced irritation in health care personnel. American Journal of Infection Control, 18, 194–200.

326 E. A. JENNER ET AL.

NORMAN, P. & SMITH, L. (1995). The Theory of Planned Behaviour and exercise: an investigation into the role of prior behaviour, behavioural intentions and attitude variability. European Journal of Social Psychology, 25, 403–415.

OPPENHEIM , A. N. (1992). Questionnaire design: interviewing and attitude measurement, 2nd edition. London: Pimter Publications.

PITTET, D., HUGONNET, S., HARBARTH, S., MOUROUGA, P., SAUVAN, V., TOUVENEAU, S. & PERNEGER, T. V. (2000). Effectiveness of a hospital-wide programme to improve compliance with hand hygiene. The Lancet, 356 (9238), 1307–1312.

PLOWMAN, R., GRAVES, N., GRIFFIN, M., ROBERTS, J. A., SWAN, A. V., COOKSON, B. & TAYLOR, L. (1999). The socio-economic burden of hospital acquired infection. Executive summary. London: Public Health Laboratory Service.

PORTER, S. R., EL-MAAYTAH, M., AFONSO, W., SCULLY, C. & LEUNG T. (1995). Cross-infection compliance of UK dental staff and students. Oral Diseases, 1(4), 198–200.

PRESTON, G. A., LARSON, E., STAMM, W. E., SCHAECHTER, M., MEDOFF, G. & SCLESSINGER, D. (1981). The effect of private rooms on patient care practices: colonization and infection in an intensive care unit. American Journal of Medicine, 70, 641–645.

RAO, G. G., JEANES, A., OSMAN, M., AYLOTT, C. & GREEN, J. (2002). Marketing hand hygiene in hospitals—a case study. Journal of Hospital Infection, 50, 42–47.

ROSENSTOCK, I. M. (1966). Why people use health services. Millbank Memorial Fund Quarterly, 44, 94–124. ROTHMAN, A. J., SALOVEY, P., TURVEY, C. & FISHKIN, S. A. (1993). Attributions of responsibility and persuasion:

increasing mammography utilization among women over 40 with an internally oriented message. Health Psychology, 12(1), 39–47.

SEMMELWEIS, I. P. (1861). The etiology, the concept and the prophylaxis of childbed fever. In: F. P. MURPHY (translation, 1981). Birmingham, Alabama: Classics of Medicine Library.

SEN, R., KEANEY, M., TRAIL, A., HOWARD, C. & CHADWICK, P. (1999). Healthcare workers washed their hands on only a third of occasions. British Medical Journal, 319, 21 August.

SHAHID, N. S., GREENOUGH, W. B., SAMADI, A. R., HUQ, M. I. & RAHMAN, N. (1996). Handwashing with soap reduces diarrhoea and spread of bacterial pathogens in a Bangladesh village. Journal of Diarrhoeal Disease Research, 14, 85–89.

STANTON, B. F. & CLEMENS, J. D. (1987). An educational intervention for altering water sanitation behaviors to reduce childhood diarrhea in urban Bangladesh. II. A randomized trial to assess the impact of the intervention on hygienic behaviors and rates of diarrhea. American Journal of Epidemiology, 2, 292–301.

STORMS, M. D. & NISBETT, R. E. (1970). Insomnia and the attribution process. Journal of Personality and Social Psychology, 16, 319–328.

TIBBALLS, J. (1996). Teaching hospital medical staff to handwash. Medical Journal of Australia, 164, 395–398. UK HEALTH DEPARTMENTS (1998). Guidance for clinical health care workers: protection against infection with

blood-borne viruses. Recommendations of the Expert Advisory Group on AIDS and the Advisory Group on Hepatitis. Wetherby: DOH.

VAN DE MORTEL, T. & HEYMAN, L. (1995). Performance feedback increases the incidence of handwashing by staff following patient contact in intensive care. Australian Critical Care, 8(2), 8–13.

VOSS, A. & WIDMER, A. F. (1997). No time for handwashing!? Handwashing versus alcoholic rub: can we afford 100% compliance? Infection Control and Hospital Epidemiology, 18(3), 205–208.

WEEKS A. (1999). [Letter]. Why I don’t wash my hands between each patient contact. British Medical Journal, 319, 518.

WILLIAMS, E. E. (1987). Psychological contributions to the control of hospital acquired infections. Unpublished PhD, University of Liverpool.

WURTZ, R., MOYE, G. & JOVANOVIC, B. (1994). Handwashing machines, handwashing compliance, and potential for cross-contamination. American Journal of Infection Control, 22, 228–230.

ZIMAKOFF, J., KJELSBERG A., LARSON, S. O. & HOLSTEIN, B. (1992). A multi-centre questionnaire investigation of attitudes toward hand hygiene, assessed by the staff in Ž fteen hospitals in Denmark and Norway. American Journal of Infection Control, 20, 58–64.