Monitoring hand hygiene adherence serves multiple functions: system monitoring, incentive for performance improvement, outbreak investigation, staffing management, and infrastructure design.60,648,651,663,666,670,676,684,686,713,714 It has to be kept in mind, however, that hand hygiene performance is only one node in a causal tree leading to the two major infectious outcomes: HCAI and health care-associated colonization with multi-resistant microorganisms. As a process element in this causal chain, hand hygiene performance itself is influenced by many factors, not least the structural aspects related to the quality and availability of products such as alcohol-based handrub at the point of care.
The correct moment for hand hygiene is usually termed “opportunity”. According to an evidence-based model of hand transmission,1,885 the opportunity corresponds to the period between the moment in which hands become colonized after touching a surface (either environment or patient) and the moment in which hands touch a receptor surface. This transition can potentially result in a negative infectious outcome. Opportunities constitute the denominator in the calculation of compliance with optimal hand hygiene. As a consequence, measurement technologies and methods can be divided into two main categories: those with a measured denominator, and those without.
An ideal indicator of hand hygiene performance would produce an unbiased and exact numerical measure of how appropriately HCWs practise hand hygiene so that its preventive effect on negative infectious outcomes is maximized. Ideally, such an indicator implies a technology that does not interfere with the behaviour of those observed, assesses the microbiological outcome of each hand cleansing action in real time, and reliably captures each moment requiring hand hygiene even during complex care activities. Furthermore, the method used should not require excessive staffing time and other incurred costs to provide sufficient data to exclude selection bias and underpowering. Bias and insufficient sample size represent the two major threats to meaningful monitoring outputs (see Part III, Section 1.1 below).
Today, such an ideal method does not exist. All current measurement approaches produce approximate information on real hand hygiene performance, each with certain advantages and disadvantages (Table III.1.1).
Hand hygiene performance in health care can be monitored directly or indirectly. Direct methods include direct observation, patient assessment or HCW self-reporting. Indirect methods include monitoring consumption of products, such as soap or handrub, and automated monitoring of the use of sinks and handrub dispensers.
1.1. Monitoring hand hygiene by direct methods
Detection of hand hygiene compliance by a validated observer (direct observation) is currently considered the gold standard in hand hygiene compliance monitoring.58 It is the only method available to detect all occurring hand hygiene opportunities and actions and to assess the number of times and appropriate timing when hand hygiene action would be required in the sequence of care. Observations are usually performed by trained and validated observers who observe care activity directly and count the occurring hand hygiene opportunities and determine the proportion being met by hand hygiene actions. It is essential that hand hygiene opportunities, indications, and actions are clearly defined (see Part III, Section 1.2). The validation of observers is essential for the quality of observation data (see under 1.2.3).
Opportunities for hand hygiene action using alcohol-based handrubs can be distinguished from those requiring handwashing with soap and water. If pre-established in the selected methodology, direct observations allow to collect more detailed information. This can comprise glove use, handrubbing technique, application time, and other quality parameters that affect hand hygiene efficacy such as the wearing of jewellery and fingernail status (see Part I, Sections 23.4 and 23.5). Whereas routine monitoring needs to be kept simple and straightforward, observations for research purposes can be even more detailed. A major drawback of direct observation is the large effort required (trained and validated staff and many working hours). For example, with a typical average density of 10 hand hygiene opportunities per hour, a total observation time of 80 hours is required to obtain 500 opportunities.
Causes of potential bias arising from hand hygiene direct observation are listed in Table III.1.2. The most important are observation, observer, and selection bias. Observation bias is generated by the presence of an observer who influences the behaviour of the observed HCWs towards a higher compliance or by an increased attention to the topic under study. In a recent study, compliance found to be 45% with overt observations was in reality only 29% when observations were covert.736 Observation bias can also induce increased recourse to hand hygiene action at inappropriate times during the sequence of care, i.e. not associated with true improvement in compliance. If observational surveys are conducted periodically, this bias would be equally distributed among all observations.831 Observation bias might be eliminated by keeping observations covert. Such observations, however, are not recommended in conjunction with promotional interventions because they can induce mistrust in the observed HCWs. Furthermore, hiding the true reason for the presence of an observer can hardly be maintained in the case of repeated observations. If a baseline observation is covert, then the results of overt follow-up observations would be confounded by the change in method. The observation bias can also be attenuated by desensitizing HCWs through the frequent presence of observers or an unobtrusive conduct during observation sessions. Some investigators call this effect the “Hawthorne effect” following ergonomic studies in the early 20th century at the Hawthorne factory of Western Electrics in the USA.334,810,1032,1033 On the other hand, this effect can be used deliberately to stimulate hand hygiene compliance in a promotional intention, rather than to obtain objective quantitative results.334,810,1033 Obtaining a sustained and never-ending Hawthorne effect associated with improved compliance with hand hygiene and decreased infection and cross-transmission rates could certainly represent an ideal perspective.810
Observer bias refers to the systematic error introduced by inter-observer variation in the observation method (Table III.1.2). To reduce this bias, observers have to be validated. It is noteworthy that even the same observer can unconsciously change his/her method over time.
Selection bias results from systematically selecting HCWs, care settings, observation times, or health-care sectors with a specific hand hygiene behaviour. In practical terms, this bias can be minimized by randomly choosing locations, times during the day, and HCWs.
Another threat to meaningful hand hygiene compliance results is the inclusion of a small sample size. In a comparative quantitative analysis of hand hygiene performance during two different periods, a large enough sample is needed to exclude the influence of chance. A sample size calculation should therefore be performed at the design stage of every hand hygiene monitoring scheme. For example, to show a difference between 40% and 60% compliance in two different measurements with a power of 90% and an alpha error of 5%, twice 140 (140×2) opportunities have to be observed. The sample size increases to twice 538 (538×2) opportunities when a difference between 40% and 50% is to be detected. Another more innovative statistical approach for measuring improvement over time and determining whether statistical improvement has really occurred is described in Appendix 4. However, because this method has not yet been applied to the analysis of hand hygiene data, further research is needed to consolidate its use in this field.
If hand hygiene monitoring is used for comparison between health-care sectors or periods, confounding factors should be included in the dataset and corrected for by stratification, adjustment, or by keeping them unchanged between the monitoring sets. Typical confounders in this field are professional category, time of day, and health-care setting. Critical reviews of observation methods have been published.809,1034,1035
Patients could be observers of HCWs’ hand hygiene compliance. In two studies, patients were encouraged to find out if HCWs had washed their hands before patient contact.804,805 Patient monitoring of hand hygiene compliance is not well documented, however, and has never been objectively evaluated.1036 Patients may not feel comfortable in a formal role as observers and are not always physically or mentally able to execute this task.737,1037
Self-assessment by HCWs can be carried out. It has been demonstrated, however, that self-reports of compliance do not correlate well with compliance measured by direct observation, and self-assessment markedly overestimates compliance with hand hygiene.218,220,666,667,676,733
1.2. The WHO-recommended method for direct observation
Observation is a sophisticated activity requiring training, skill and experience. Observers have to be aware of the multiple potential biases introduced with the observation process and they can help to minimize these by gaining a full understanding of the methodology. A stringent adherence to the same methodology over space and time is required.
WHO proposes a standardized hand hygiene observation method based on an approach validated through several studies.60,652,686,738 All relevant theoretical and practical aspects related to this method are detailed in the Hand Hygiene Reference Technical Manual that is included in the Implementation Toolkit (available at http://www.who.int/gpsc/en/). An “Observation form” for data collection, consistent with the proposed method and including concise user instructions, is also availabletogether with a “Compliance calculation form” to facilitate the immediate performance feedback. Observation of hand hygiene practices is an essential component of the WHO Hand Hygiene Improvement Strategy (See Part I, Sections 21.2 and 21.3).
1.2.1. Profile and task of observers
The task of observers is to observe HCWs during their usual care activity and to assess their compliance with the recommended indications for hand hygiene. To be able to accomplish this task, observers have to be able to understand the logic of care. Ideally, they have training and experience in patient care as professionals.
1.2.2. Training of observers
Observers have to be trained according to the principles of “My five moments for hand hygiene” and, ideally, have become excellent monitors of the application of hand hygiene during health-care delivery. Their excellence should be confirmed through observations performed by a senior observer, if feasible, depending on the setting. They have then to be instructed in hand hygiene observation according to the present methodology. This should take a relatively short time if they have already proved to be proficient in the application of the five moments.
1.2.3. Validation of observers
Once knowledgeable in the use of the observation form and process, observers must be validated either by parallel observation jointly with a confirmed observer, or by being tested through the use of the WHO Training Film included in the WHO Implementation Toolkit (available at http://www.who.int/gpsc/en/). In the first case, two observers engage in an observation session during a real-life care situation and each completes an observation form separately while observing the same HCW and the same care sequence. Results are then compared and discordant notifications discussed. This process is repeated until concordance is reached in the number and nature of each occurring hand hygiene opportunity. It is recommended that the person in charge of validation remains the same for all new potential observers in a given setting. It is advisable to perform validation in each care setting that is to be monitored by the future observer. The WHO Training Film provides visual examples of the five moments for HCWs and observers. Observers can be trained and tested through the use of the scenarios, which include different sequences of health care where hand hygiene is necessary. Observers are asked to complete the form while watching the film, and the trainer can then judge their performance by comparing the results with the those provided in a slide show presentation that accompanies the film. The subsequent discussion is usually very valuable for learning purposes. If a time grid of opportunities can be established in a scenario, kappa statistics can be calculated to quantify the level of coincidence between two observers.
1.2.4. Understanding the five moments for hand hygiene
The concept of “My five moments for hand hygiene” has been created as a robust framework for understanding, training, measuring, and communicating hand hygiene performance.1 Understanding this concept (see Part I, Section 21.4) is a prerequisite for any future observer. It is a simple concept that should not leave any knowledge gap between the insight of observers and observed HCWs once they are adequately trained in hand hygiene. It is essential, however, that local specificity related to the application of the “five moments” is established and known by everyone. For example, the delimitation of the patient zone in a given setting needs to be specifically determined.
Health-care activity must be imagined as a succession of tasks during which the HCWs’ hands touch different types of surfaces prior to and after patient contact. Each contact is a potential source of contamination for HCW’s’ hands.
A crucial point specific to observations is the distinction between indications and opportunities, which is more extensively described in the Hand Hygiene Reference Technical Manual. The indication is the reason why hand hygiene is necessary at a given moment to effectively interrupt microbial transmission during care, and it corresponds to precise moments in patient care. Very close to the concept of indication, the term opportunity is much more relevant to the observer: it determines the need to perform the hand hygiene action, whether the reason (the indication that leads to the action) be single or multiple. From the observer point of view, the opportunity exists whenever one of the indications for hand hygiene occurs and is observed. Several indications may arise simultaneously and create a single opportunity. Very importantly, the opportunity constitutes the denominator for calculating compliance, i.e. the proportion of times that HCWs perform hand hygiene action of all observed moments when this was required.
For this purpose, hand hygiene action is defined as either rubbing hands with an alcohol-based handrub accepted by the institution or handwashing with soap and water. Neither the duration nor other quality aspects of hand hygiene such as the quantity of product used, glove use, length of fingernails, or the presence of jewellery are assessed.
It is important to understand that hand hygiene actions not corresponding to an opportunity, and therefore “additional” and not required, should not be taken into account by the observer.
1.2.5. Understanding the observation form
Observations are noted on a paper form using a pencil and rubber. Each form represents a separate observation session. Experience shows that this material is ergonomic for observations. The surface of a sheet of paper provides the necessary overview of the past evolution of observed activity in several, simultaneously observed HCWs. Using a pencil and an eraser, errors can easily be corrected.
The form has three main sections: 1) a header contains information on the institutional level (country, city, hospital, site identity); 2) a second header contains information on the session (observer identity, date, start and end time, duration, period number, session number, form number, department, service name, ward name); and 3) four columns below the header represent the sequence of actions for different HCWs observed during the same session. Each column is usually dedicated to one HCW and therefore the form can include up to four HCWs. Alternatively, in situations with low activity, each column can be dedicated to a different professional category and therefore the HCWs belonging to the same professional category can be grouped within one column. This method can be practical when the observer chooses to observe more than four HCWs during the same session. This results, however, in a loss of the possibility to calculate a per person density of hand hygiene opportunities and individual feedback after the session. The header of each column contains information about the observed HCW (professional category, code, number). The rest of the column consists of equal blocks that are incrementally numbered from 1 to 8 from top to bottom. Each block represents one of the sequentially occurring opportunities for hand hygiene. For each opportunity, the observer notes in the corresponding block all the applicable indications and if hand hygiene was executed by handrubbing, handwashing or missed.
1.2.6. Determining the scope of an observation period
Before starting an observation period, the investigators and project coordinators must determine the scope of observations. Possible scopes are listed in Table III.1.3. If the scope is to build a comparison between two or more observation periods to assess the evolution of hand hygiene compliance over time, special attention should be paid to control for the potential confounding factors. This can be achieved by predefining a target number of opportunities by profession, wards, and time of day. To minimise inter-observer variability, the observer or the team of observers should remain the same across the different periods of the project. The best unit for calculation is the denominator, i.e. opportunities for hand hygiene, because this will directly influence the results.
184.108.40.206. Selection of location and time
A representative mix of wards and time of day should be sought. Naturally, observers tend to undertake their activity at times and in locations with a high density of care to gather a higher number of opportunities more quickly. Observers have to be aware that changing the method of selecting time and location for observations between observation periods can lead to bias because there is usually an association between density of opportunities and compliance. Therefore, we suggest to establish a rough location plan and timetable ahead of planned observations that will be remain stable over observation periods.
220.127.116.11. Selection of HCWs
Once location and time are determined, observers have to choose the HCWs to be observed during a session. Selection bias should be minimized by choosing at random. In the case of repeated observation periods in particular, observers may know the intrinsic performance of individual HCWs and this could easily influence the overall observation result by always selecting HCWs with extreme behaviour.
18.104.22.168. Starting, continuing, and concluding an observation session
Once a health-care situation is identified, the observer may introduce himself/herself by indicating unobtrusively the scope of his/her presence. The way in which this introduction is handled depends on local social and medical culture. A balance should be sought between increased observation bias through a too overt presence and inducing the feeling of being cheated in the observed by pretending to be there for another scope. This includes also a discreet positioning of the observer.
After completing the form header, each observed opportunity is noted on the form (see above). Only opportunities for which the entire time between the two delimiting hand-to-surface exposures can be observed are noted.
During the observation session, the observer must not interfere with observed staff. The session should be concluded after 20 minutes ± 10 minutes according the duration of care activity. The observer may want to give feedback to the observed HCW(s) about the observed hand hygiene performance. This depends on the scope of the observation, but it was found to be very efficient and appreciated by HCWs.
Following data entry (Epi Info databases for entering data collected according to the WHO-recommended method for direct observation are available), the simplest form of results is the overall compliance. This is calculated by dividing the number of observed hand hygiene actions performed when an opportunity occurs, by the total number of opportunities. It has been found useful to stratify compliance by institutional sector, professional category, and indication (moment) for hand hygiene using the ‘My five moments for hand hygiene’ as strata.1
1.2.8. Reporting of results
Feedback of results to those concerned is a very powerful promotional tool and should firstly address groups with a strong internal identity. A short delay between observation activity and reporting of results might increase the effect of feedback. Continual feedback of unchangingly bad results without any intervention should be avoided, as it may lead to “desensibilization” and demotivation.
Special attention should be given to the potentially low number of observed opportunities when using percentages to report compliance. Low numbers occur especially with stratified results. It is good practice to calculate 95% confidence intervals and include these in graphics. For instance, for 30 opportunities with a compliance of 50%, the confidence interval would stretch from 31% to 69% compliance. With 100 opportunities and 50% compliance, the confidence interval would shrink to 40–60%, and for 200 and 50% compliance opportunities to 43–57%. Finally, observations can be reported to HCWs directly after each session, which produces an immediate impact. For statistical methods to measure hand hygiene compliance over time see also Appendix 4.
1.3. Indirect monitoring of hand hygiene performance
In the quest for less expensive monitoring approaches, experts have used the consumption of hand hygiene products such as paper towels,1038 alcohol-based handrub or liquid soap60,334,429,486,489,713,803,852 to estimate the number of hand hygiene actions. To make these monitoring techniques more meaningful, the quantity of handrub was translated into a number of hand hygiene actions by using the average amount per action as a divider. The missing denominator of the need for hand hygiene actions was either ignored by only following the evolution over time, or substituted by a surrogate measure such as patient days or workload indicators drawn from a computerized database of nursing activities.851
Some studies60,334,486 have shown that the consumption of products used for hand hygiene correlated with observed hand hygiene compliance, whereas others have not.1039 Thus, the use of this measure as a surrogate for monitoring hand hygiene practices deserves further validation. Other studies found that feedback based on measured soap and paper towel consumption did not have an impact on hand hygiene.802,1038
Methods based on product consumption cannot determine if hand hygiene actions are performed at the right moment during care or if the technique is correct. The advantages, however, are that they are simple, can be continuous, and provide a global picture that remains unaffected by selection or observer bias and, most likely, observation bias. The amount of alcohol-based handrub used by health-care settings has been selected as one of the indicators. Nevertheless, it has to be considered that this measure may not exactly reflect the product consumption by HCWs, but could include the amount used by visitors or patients, especially if the dispensers are located also in public areas of the health-care setting and they are wall-mounted.
1.4. Automated monitoring of hand hygiene
The use of sinks and handrub dispensers can be monitored electronically.699,710,852,986 Systems that are even able to identify HCWs when using a sink or a handrub dispenser are under development. These methods allow precise quantitative results on hand hygiene activity to be obtained, with the only costs being the installation and maintenance of the system. Changes over time can be assessed. Some studies have attempted to measure the need for hand hygiene by monitoring patient room entries and linking each entry to the use of a sink or a handrub dispenser. For the moment, no comparative studies exist to validate the appropriateness of electronic detection of hand hygiene opportunities.
Wireless devices placed inside handrub or soap dispensers can provide useful information regarding patterns of hand hygiene frequency. A recent study evaluated wireless devices that were placed inside handrub dispensers on a general medical ward and in a surgical intensive care unit.1040 During a 3-month trial period, 17 304 hand hygiene episodes using handrub were recorded on the medical ward for a rate of 9.4 hand hygiene episodes/patient-day. A total of 50 874 hand hygiene episodes using handrub were recorded in the ICU for a rate of 47.7 hand hygiene episodes/patient-day. Average usage was highest between 10:00 and 19:00; the lowest was at 05:00. By mapping the location of each device, it was observed that dispensers located in rooms with patients on contact precautions were used significantly less often than those located in other rooms on the ward (P = 0.006).
All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: tni.ohw@sredrokoob). Requests for permission to reproduce or translate WHO publications – whether for sale or for noncommercial distribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; e-mail: tni.ohw@snoissimrep).
World Health Organization, Geneva
WHO Guidelines on Hand Hygiene in Health Care: First Global Patient Safety Challenge Clean Care Is Safer Care. Geneva: World Health Organization; 2009. 1, Hand hygiene as a performance indicator.