Editor's note: Didier Pittet, MD, MS, is Professor of Medicine and Director of the Infection Control Programme and WHO Collaborating Centre on Patient Safety at the University of Geneva Hospitals, Switzerland. He is Lead Advisor of the WHO First Global Patient Safety Challenge: Clean Care is Safer Care. We spoke with him about hand hygiene in health care, including how to implement culture change and measure compliance.

Dr. Robert Wachter, Editor, AHRQ WebM&M: Why is hand hygiene so hard?

Dr. Didier Pittet: One of the main reasons is time. When we conducted epidemiological studies to look at why health care workers would not clean their hands appropriately, we realized that they simply had no time for soap-and-water hand washing, which the guidelines called for. Epidemiological studies demonstrated that health care workers had too many opportunities per hour of patient care to clean hands, and they simply could not wash their hands with soap and water in an appropriate time. So that's why we recommended the systematic change to alcohol-based hand rubbing.

RW: Was alcohol-based rubbing a technology that had been available and wasn't being used? Or did the evidence showing that time was the issue allow this new technology to be developed?

DP: The possibility to clean hands with alcohol was available and recommended if you could not reach the washbasin. It was also recommended for multi-resistant bacteria, but with no clear guideline on the application. In other words, it was not widely available, but it was a tool familiar to people working in microbiology laboratories.

RW: Recognizing that a very important step was putting in alcohol-based hand rubs and cutting the time necessary, many hospitals with very convenient hand gels everywhere still have hand hygiene rates of 50%. Once you've solved that big problem of time, or at least made it easier, what other obstacles get in the way?

DP: Of course, the most important was to be able to apply system change, to make sure the alcohol is available at the point of care. Once that change has been made, you have to work on the other elements that will make the behavior change a real success. With a multimodal intervention strategy, you have to use several strategies to change behavior—promoting education using tools that challenge and help health care workers to understand the most important time to perform hand hygiene. And to be honest, the guidelines were not clear—there was no clear indication as to when health care workers should clean their hands. We had to work on that based on epidemiological studies, review the literature, and make sure that we could transform the guidelines into very easy-to-understand cues for action. Performance monitoring and feedback was a third element of the strategy. We monitored the performance of health care workers regularly and gave feedback either at an individual or group level, or sometimes at the department level. We then used cues for action at the point of care using posters or other types of prompts to remind health care workers how and when to perform hand hygiene. The last element of the strategy was to push for a change toward hand hygiene as an essential element of a patient safety culture within the institution.

RW: Let's go to the measurement step. What do you think about how to do it, the role of secret shoppers versus cameras versus other strategies? And then after that, you could give individuals feedback but show them how their colleagues are doing, you could do it at the departmental level, the institutional level. Which of those works best? How do you do the auditing in the first place?

DP: Auditing is performed using infection control practitioners monitoring practices in different wards and different departments by direct observation. Of course, direct observation is sometimes associated with observer bias and even may induce some Hawthorne effect, but people get more and more used to being observed over time, and the Hawthorne effect becomes less important. The benefit of direct monitoring is that you know exactly what you are observing, and when you offer feedback on performance you can explain to health care workers why and how they should modify their behavior and improve. That's a big advantage. The downside is that it is time-consuming and resource-demanding because you need trained and validated health care professionals to monitor the practices.

There are other ways to do it. You can use surrogate markers. One easy way is simply to monitor the amount of alcohol-based hand rub used by different wards or departments in the hospital. The problem with that is you cannot tell if or when people are using the alcohol-based hand rub, if it is appropriately used at the appropriate time, and in the perfect way. But this surrogate marker is relatively cheap and easy to monitor provided you are very well organized in your institution. There are many other ways to monitor performance. Over the coming years, we will have tools available to facilitate monitoring health care worker compliance and also reduce the resources used.

Systems have been developed using cameras in rooms to observe health care workers during their practices. There are systems that are RFID-linked, recognizing the time and exact amount of product used at the point of care. Other systems monitor the gestures of the health care workers, or the amount of time and where and when health care workers are actually cleaning their hands. Of course there are many caveats when you are using those systems. Some may be a problem regarding confidentiality. Another problem arises if you clean your hands half the time using soap and water and half the time using alcohol-based hand rubs. Thus, a system monitoring only the use of alcohol-based hand rub will give only a partial idea of the real performance. To be compared from one hospital to another in particular, or even from one ward to another in the same hospital, all these systems need to be validated.

RW: How do you decide what level of the organization to give information back and then what to do with it? Whether to build in incentives or just use pride and shame and some level of public reporting?

DP: The second step of performance reporting is feedback. In most institutions, it is best to begin with an aggregated approach. That means you are returning the results as an average of a ward, a department, or even a hospital. You can do it at the beginning very easily because usually compliance is relatively low and it gives an incentive to the entire hospital or department to make good progress. Then, progressively, you realize that health care workers would like to get their own personal data. You can start doing some benchmarking, provided you have enough data to allow comparisons. We have described already the methodology for these: you need to monitor 200 opportunities by ward to be able to compare one ward to another and then you repeat monitoring. Later, you will find that people are asking more and more for individual-level feedback. People like to know why they are not compliant with practices. Ideally, you should be able to feed back the performance to each individual health care worker. But we have more than 12,000 employees in our institution—so, how would you like to follow all employees and tell them about their performance! Nevertheless, in the long term, I'm convinced that there will be systems to allow performance feedback at the individual level for large numbers of health care workers.

Sometimes, we have been able to provide individual feedback and to talk them through the process. For example, we say, "Let me tell you how you can improve." "Explain why you did this at that time and forgot to clean your hands," or "Why did you clean your hands when it was not necessary?" Or, "You could have waited during the course of care to perform hand hygiene at a much more appropriate time and saved some of your hand hygiene actions." Average compliance is now very high at our institution according to the Five Moments for Hand Hygiene, the gold standard worldwide for monitoring performance. Now we see health care workers who really want to comply, try to do their best, and are actually performing more hand cleansing actions than necessary, which is not good because it means that they have not completely integrated the Five Moments for Hand Hygiene. At the beginning, we conducted performance feedback on an average aggregated basis in different wards of the hospital for several years. Health care workers are now asking to know not only the average compliance of their department or ward, but also their own compliance with hand hygiene practice because they want to improve. The ideal is to move from a concept of average aggregated performance feedback to individual feedback as much as possible. But in all fairness, I recognize willingly that it is resource demanding.

RW: What have we learned about the use of bigger sticks that might be either public reporting (beyond feedback to individuals or units within an institution) of hand hygiene rates or even using incentives? Do either of those make any sense, and have they been tried?

DP: Yes it makes sense. Our experience has been that benchmarking the aggregated rates between different departments and between different wards in the same departments across the entire institution was extremely useful. Why? Because it's clear that to make sure that you can compare hand hygiene rates or compliance with hand hygiene, one hospital has to compare to another hospital, and you need to make sure that people are really using the same approach to monitor hand hygiene. Unfortunately, most of the time this is not the case, and so it's quite difficult to benchmark one hospital to another. But it is feasible once people have really standardized the monitoring of hand hygiene performance. Nevertheless, benchmarking within the institution is extremely powerful to mobilize the entire department or entire ward to perform better. Of course, once you have the data, you need to decide what you want do with it. I always push for positive rewarding because our experience has been that health care workers in most wards are doing their best. They want to know more. They want to get more feedback. They want to improve.

According to your institutional level of hand hygiene culture, you may be pushed today to make sure that every single ward will compare the rates that they got last month to this month, or those from the last quarter to this quarter. The nomination of hand hygiene "champions" on wards can be very beneficial to help staff change their practices. That would be the best way to make significant improvements. The process of having an accreditator visit your hospital and monitor hand hygiene performance is dangerous. You are tempted to have excellent results at the time of accreditation, but it doesn't tell you that on average, year-long compliance with hand hygiene is good enough.

What I prefer is to recommend that hospitals monitor themselves at an institutional level to know whether the hospital is positioned to become a good or even an excellent hand hygiene hospital. The hand hygiene self-assessment framework tool developed at the World Health Organization is available to make sure that hospitals know where they are. We have much experience with monitoring this tool and it has been used in more than 5000 hospitals worldwide. Some of the data for US hospitals were recently published in the American Journal of Infection Control, and it's clear that many hospitals have succeeded in applying system change; for example, making sure that alcohol-based hand rub is available at the point of care in most hospital rooms and wards. You will realize that from performance monitoring, that education and institutional safety culture have a long way to go.

RW: Can you talk about the impact of hand hygiene? If a hospital is able to successfully get its hand hygiene rate from 50% to 90%, do you have any sense of how many infections are prevented or lives saved or dollars saved?

DP: I admit that it is quite complicated to conduct evidence-based studies to demonstrate the impact of hand hygiene. We performed a literature review (in press) of 48 studies published up to the end of last year. Among these studies, all but two demonstrated that a successful hand hygiene promotion was associated with a reduction of health care–associated infection or cross-transmission of multi-resistant bacteria. However, it's clear that you cannot always monitor all health care–associated infections, and some of the studies looked at targeted outcomes, such as bloodstream infection, pneumonia in the ICU, and ventilator-associated pneumonia. Many studies looked at the overall prevalence of infections. Some studies looked at incidence reduction. Overall, the impact of the most successful study was as high as an 80% reduction. Most reached 40% to 50% reduction, and some achieved a 20% reduction or even a little less. Now the important question is: how much of a decrease should you expect in the infection rate as compared to an increase in the compliance rate? We have no simple answer. Why? Because it has been demonstrated that improving compliance from 20%—and many hospitals have compliance as low as 20%—to 40% or 50% reduces infection. And we can see that increasing compliance from 50% to 70% or 80% is also associated with infection reduction. Then it becomes more difficult to demonstrate a benefit at 80% compliance because the study doesn't have sufficient power.

In most institutions where you are promoting hand hygiene, it is clear that you are also implementing other infection control measures, so it's difficult to say that hand hygiene is actually responsible for the total infections reduced. Similarly, many infection control initiatives today—such as some of the bundles used to prevent catheter-related infections, ventilator-associated pneumonia in the ICU, or surgical-site and urinary tract infections—include hand hygiene performance, hand hygiene monitoring, and hand hygiene facilitated by the use of alcohol-based hand rub. In this case, it becomes very difficult to differentiate between the impact of hand hygiene alone or the impact of hand hygiene within a bundle of care practices. That's why the science is so difficult.

RW: What lessons from hand hygiene are relevant to trying to increase compliance with other safety practices, and where do you think it's different? Are there any areas where the nuances of what you're trying to do in hand hygiene make it problematic to generalize from the hand hygiene experience to let's say trying to decrease falls or medication errors?

DP: Well, that's a fascinating question. I think that our chance in hand hygiene was certainly that the science behind hand hygiene has been founded on an epidemiological study of compliance. If I had to give advice, I would say to start with a really good epidemiological, scientific-based study or studies. Once you have these data, you can design the action and the intervention. What we learned with hand hygiene was that system change was a definite prerequisite to the success of the story, but to change behavior, system change was not sufficient to be implemented as a single measure. To back this up, there are studies where system change has been implemented as the only parameter to change hand hygiene practices, and the intervention was not successful. In other words, in studies where the author actually introduced the alcohol-based hand rub to the bedside and did nothing else, nothing happened.

It's important to realize that system change alone will not produce the desired effect. Hand hygiene requires a multimodal intervention strategy to change practices. For example, when you drive your car, you have to fasten your seatbelt. If you don't have a seatbelt, you won't do it. If you are not afraid of being fined by the police because your seatbelt is not fastened or if you haven't seen an accident with people suffering from not wearing their seatbelts, you will probably not wear your seatbelt. You need a multimodal approach. And this multimodal approach in hand hygiene includes five different elements. I cannot tell you what are the most critical elements in the strategy. But I can tell you it is certain that monitoring and performance feedback have been and still are extremely important; the use of education tools that people understand has been also extremely important. Simplifying the life of the health care workers by having those tools that were easy to implement and attractive was very, very useful. In addition, it was most helpful to secure the support and interest of the CEO, CMO, and the Chief of Nursing to obtain a culture change in the institution.

The other thing I wanted to say is that we had the privilege to drive the hand hygiene promotion campaign worldwide, and we have applied and tested the strategy and monitored the reduction in health care–associated infection rates now in all continents, and it worked—everywhere. It worked for every health care category—for physicians, nurses, auxiliary staff, and students. What we learned from our experience is now spreading all over the world. We are currently in more than 170 of the 194 countries of the United Nations. An important point is to realize that we need to adapt this strategy to different cultures and different countries. Our tools are completely standardized and can be easily adapted to the way in which the strategy is implemented. This is the case today all around the globe. For people to adopt a strategy tailored to their needs, they need to be able to adapt this strategy to their own practices, to their own language, but also to the culture and to the way people are working in their own institutions and sometimes in their own country.

RW: You have emphasized a very nurturing strategy, making it friendly, making it sexy, assuming that these are good people trying to do the right thing. Sometimes people look at hand hygiene and say, it should just be required. It is something that patients have a right to expect and that once the systems are in place and the education has been done, if someone cannot figure out a way to clean their hands they should be fired. What do you think about that line of reasoning?

DP: As you may understand, I've been confronted with this discussion before and have visited some institutions that applied certain strategies in order to really push the change. Before starting with this type of nurturing strategy, you must ensure that the institution is providing all the tools necessary for health care workers to apply the strategy and improve practices in the best way. I have visited too many institutions where the institution is not providing the tools for people to behave very well or as expected. Now without going into too much of the coercive actions that could be promoted at some point, I would say that there are some places and some situations where you have the impression that some health care workers, but they are rare, really do not want to improve and are definitely not helping other health care workers in their own ward, department, or hospital to improve their behavior. So at that point I would say that we should find a way to make sure that those people will not continue to demonstrate such inappropriate behavior. But I would emphasize that these are only rare instances.

Regarding the patient position, I completely agree that patients should have the right to benefit from health care workers practicing appropriate hand hygiene technique every time it's necessary. We have been involving patients in patient participation and/or patient empowerment strategies for some time. From experience, I can say that it is not always easy for the health care workers or even the patients themselves to participate. In the long term, I envisage that patients would play an important role in participating in behavior change, but they should not be responsible for it. They should be partnering with their health care worker for patient safety and certainly not judging health care workers on these simple hand hygiene actions that improve patient safety. I recognize that in some situations, it may be tempting to suggest that behavioral change be pushed by some coercive actions. But, most of the time, I would prefer change to move forward in a partnership between the institution, the health care workers in this institution, and the patients receiving care at this institution.