A large-scale assessment of hand hygiene quality and the effectiveness of the “WHO 6-steps”

Failed hand hygiene of clinical staff is the major contributor to Healthcare-Associated Infections (HAIs) which occur in 7.1% of hospital admissions in the Western world [1]. Each year around 150,000 deaths in Europe and 100,000 in the USA are attributed to HAIs [2]. European and US standards have been developed in accordance with the World Health Organization’s (WHO) recommendations, defining 5 crucial moments of hand hygiene and 6 practical hand rubbing steps through which alcohol-based hand rub solutions have been proven effective [3].

Several studies have investigated staff compliance to hand hygiene guidelines, however, most involved relatively small number of subjects, typically 50–500 [4–8]. Studies of hand hygiene in hospitals have focused on hand washing standards [9–11], improving hand rubbing technique [12, 13], or the composition of the hand rub and scrub [14–18], the attitude of staff regarding compliance to hand hygiene moments [19, 20], the impact of finger rings, wrist watches and other accessories upon hand hygiene quality [8, 21], and strategies of monitoring the compliance to the 5 moments [22]. Consensus guidelines and recommendations regarding hand hygiene have been produced through the WHO and descriptions on how they can be introduced have been published in [23, 24]. Automated audits of hand hygiene have recently been trialled [25–27], but these are limited to monitoring compliance to the 5 moments rather than the quality of application as intended via the 6 steps.

The National University Hospital (NUH) of Singapore is a 1000-bed tertiary referral center. Hand hygiene training, assessment and monitoring have been a core activity since 2007, but the focus like in most other institutions has been on compliance with the 5 moments, which has improved from 15% to 69% based on up to 800 observations/month (unpublished). Less attention has been given to the technique and quality of hand hygiene [28].

In May 2011, NUH undertook a program to coincide with the WHO World Hand Hygiene Day within the frame of WHO Save Lives: Clean Your Hands campaign. Building from past subjective assessments of hand hygiene quality using Ultra Violet (UV)-marked alcoholic hand rub (Schülke Optik; Schülke & Mayr GmbH, Norderstedt, Germany) and regular desk top UV lights, we established a process for blinded assessment and storage of digital images for subsequent analysis. In this study digital images were used to analyze hand hygiene quality and therefore the effectiveness of a training program based on WHO’s 6 steps approach.

After the week-long assessment period, all images were reanalyzed applying the same definitions as used on site. This resulted in 3349 passes (72%) consistent with a good level of inter rater concordance between the on-site team and the later re-evaluation. The amount of overruled on-site decisions added up to approximately 8.5% of the total cases. The kappa concordance test between on-site evaluation and re-evaluation gave κ=0.80 (95% CI 0.78–0.82). We used the post-evaluation metrics for further analysis.

Pass rates by staff occupation type and demographics are summarized in Table 1. Figure 1 shows the rate and confidence interval of satisfactory hand washing, plotted against occupation and gender. Similarly, in Figure 2, rates and confidence intervals are arranged by age groups and gender.

The overall pass rate was 72% (95% CI 71%–73%). Female staff performed significantly better in almost all job and age categories, with an overall pass rate of 75% (95% CI 73%–77%) versus 62% (95% CI 59%–65%) that of males.

Risk indicators (RI) for failed hand rub application were identified and listed in Table 2. For these risk groups, a higher incidence of unsatisfactory hand washing was found. Identified RI’s are supported by non-overlapping confidence intervals (Figure 3) and statistical significance.

No.	Hypothesis tested	Incidence	Odds	Significance
		higher by	ratio
1	Male staff have higher incidence of unsatisfactory hand washing than females	51% (95% CI 32%–62%)	1.82 (95% CI 1.49–2.22)	p<0.001
2	Environmental service workers have higher incidence of unsatisfactory hand washing than all other job categories counted together	32% (95% CI 21%–45%)	1.49 (95% CI 1.14–1.96)	p<0.001
3	Staff aged –40 and 60+ have higher incidence of unsatisfactory hand washing than those aged 40–60	21% (95% CI 11%–30%)	1.29 (95% CI 1.05–1.58)	p=0.001

These risk indicators were found significant and relevant. Corresponding confidence intervals are shown in Figure 3.

We classified all individuals into three quality groups based on whether they made no mistakes, small mistakes only, or big mistakes. Figure 4 shows the distribution of these groups defined by gender and occupation. Among those who failed the first assessment, we further distinguished three categories: those who made mistakes on the palmar side of the hand, on the dorsal side, or both. The distribution of these categories, within the gender and occupation groups is presented in Figure 5.

The most frequently missed sites were identified on the dorsal side of fingers in the proximity of the nails (16%), on the thenar eminence and in the proximity of the wrist crease (18%). Figure 6 shows example images taken during the exercise.

This study evaluated the effectiveness of the WHO 6-step hand rubbing technique, involving a large number of health care workers and a novel screening technology.

A key advantage of the applied methodology is the ability to evaluate the actual outcome of hand washing and in this situation, the educational process. Identifying failures in hand hygiene coverage represents an opportunity to enhance one’s infection prevention efforts. Identifying certain risk groups and typical mistakes regarding the application patterns allows for the design of specific education programs and also helps to left the use of a hand rubbing protocol. Another advantage of the methodology is the ability to easily involve several thousand health care staff in the investigation, a major increase compared to the 465 subjects of the largest similar study [8]. Instead of applying restricted observation of nurses and physicians only, as most similar studies do, our investigations involved all staff of a single hospital because all health workers have the potential to contribute to cross-infection [31].

Previous studies have found a higher incidence of unsatisfactory hand hygiene in males, but this difference is generally insignificant possibly due to small samples size. A recent study culturing glove juice found the prevalence of Enterobacteriaceae higher in males by 55% (23.8% vs. 15.4%, p=0.156) [8], similar to our findings (Table 2). The differences we found between various occupational groups were similar to non-significant trends observed by others [32–34], but in our study statistical significance was evident.

Our study did not apply microbiological validation but assumed that bright areas of the hand were clean, while dark spots could contain potentially dangerous pathogens. The general intention in hand hygiene is to fully cover the surface of both hands with antiseptic solution. Our methodology monitors this information. Microbiological techniques provide an alternative and arguably more useful assessment but are more complex and expensive to undertake particularly on this scale. All commercial solutions must undergo specific tests for efficacy against germs: they are approved as bactericidal (EN 1276), fungicidal (EN 1650), or sporicidal (EN 13704) solutions. These tests provide the evidence that once the surface is treated appropriately, efficacy is guaranteed. Occasionally microbiological studies have been employed for randomized hand hygiene assessment, however, always on a small scale [5, 8]. Nevertheless, sampling and cultivation are susceptible to environmental effects and cross-contamination, distorting the results of the evaluation. In this respect our methodology provides a simpler alternative to culture techniques.

The main contribution of our study is not the monitoring methodology itself, but the statistical findings of the hand hygiene quality assessment. Microbiological validation of our monitoring and assessment methodology is an area for future research.

The setting of our study is likely to have actually overestimated the real life quality of hand rub application given that it was a specific and overtly observed assessment. The generalizability of our findings therefore has some limitations however monitoring staff in everyday practice would carry significant challenges. The fact that a single measurement was taken of each subject, and the deviation between on-site and subsequent evaluation, may slightly distort the statistical data.Similarly, another limitation of the study is that it was undertaken at a single institution.

This study has demonstrated that a high failure rate in hand hygiene coverage among health care workers can occur even when observed immediately after a specific training programme of the WHO’s 6 steps. It is a significant and generally underrated risk factor since imperfect hand hygiene technique may endanger the effectiveness of the entire hand hygiene routine.

We established an environment to observe and evaluate the effectiveness of the 6 step hand washing procedure, which—when applied at the times of the WHO 5 moments—is considered the most effective way to prevent the transmission of HAIs. The quality of hand rub application was analyzed on 4642 hospital staff immediately after a personalized education program. High failure rates suggest that the 6 steps, as advocated by the WHO, may be too complex to expect staff to comply with routinely. Our findings have revealed that besides effective monitoring of compliance with the five moments, there is a strong need for improved, targeted educational efforts on handhygiene technique.