Healthcare Analytics, Population Health Management, Healthcare Big Data

Featured Articles News

How data analytics helped measure, increase hand hygiene

By Jennifer Bresnick

- While the Centers for Disease Control asserts that hospital-acquired infections are on the decline thanks to financial quality penalties and a closer focus on patient safety, the numbers are still staggeringly high, with 722,000 cases and 75,000 deaths in 2011.  Despite the omnipresence of antibacterial gel dispensers and reminders for staff and visitors to keep clean, a lack of adequate hand hygiene is still a major culprit in the spread of superbugs and deadly infections.

Greenville Health System in South Carolina didn’t want its staff members to contribute to the high toll that poor hygiene imposes on patients, so Dr. Tom Diller decided to figure out a better way to measure and increase compliance.  Dr. Diller explained to HealthITAnalytics how he helped to create a benchmarking system at Greenville, based on the World Health Organization’s Five Moments for Hand Hygiene guidelines, to flag setting-specific opportunities to improve patient safety.

What was the impetus for starting this study?

If you look at hospitals across the country, only about 40 to 50 percent of hand hygiene is actually done well, and that’s even after fairly intense efforts.  The more progressive organizations have tried to measure hand washing, either through direct observation, which has been the gold standard up to this point in time, or through various methods that don’t work very well, like patient surveys asking if the clinician washed their hands.  They might check how often the dispensers go dry as a measure of how much product has been used, but that also doesn’t work very well.

The dilemma with direct observation is that it’s very resource intensive, and it’s very difficult to be really accurate.  At Greenville, five years ago before we did this study, the unit manager would do their own survey on a monthly basis, report the numbers, and it would all kind of flow into corporate and we’d have a number.  The problem is that the unit managers were reporting 90% to 95% compliance every month, because the doctors and nurses knew they were being observed, so they washed their hands a whole lot more often than they might have if no one was breathing down their necks about it.   No one wanted to look bad – none of the auditors wanted their unit to look bad, either – so it wasn’t an accurate picture of what was really going on.

READ MORE: Big Data Shows Gender-Based Medical Error, Patient Safety Patterns

So what we did instead was what I call a “super-secret shopper.”  Nobody knew that we had trained people roaming the hospital measuring hand hygiene.  What we found was a rate of about fifty-three or fifty-four percent, which is consistent with numerous other reports from across the country of when people have done this type of observation.

What are the advantages of using an electronic monitoring system, and how do they work?

There are some clear advantages to using electronic monitoring systems instead of these manual processes.  It’s much less resource intensive, because you don’t have to recruit somebody to follow people around, watching them.   Secondly, it’s not intrusive in the patient care at all. .  And third is that you can get much more real-time results.  With the old way of doing it, we compiled things on a monthly basis.  We’d send out what their hand hygiene rate was in the last month.  Well, that doesn’t really help drive change, but in these electronic monitoring systems you can get it by day or by week, and you can make changes more quickly.

Most systems on the market at this point use a methodology that is what I’ll call “entry and exit.”  There’s usually some sort of a chip in the badge of the individual, and as they walk through the door it activates and then they’re able to electronically determine whether or not they actuated the soap or the gel or foam while they’re in the room.  But there’s a huge problem with that because it’s not evidence-based practice.

WHO has the Five Moments of Hand Hygiene, and the CDC has published their own method that’s very similar to it.  The issue is that once you enter the room, there are significant opportunities for additional hand hygiene.  For example, an issue that occurred in our hospital in Greenville was an outbreak of a fairly serious infection of Vancomycin-resistant enterococcus (VRE).  We found that it was actually present in a few rooms on the privacy curtains.  So if a person had entered the room, washed their hands, and reached up and grabbed the privacy curtain and pulled it shut, they now were re-contaminated and that’s how they were actually spreading the infection.

READ MORE: UPenn Uses Machine Learning, EHRs to Target Severe Sepsis

So the Five Moments or the CDC’s methodology really requires you to wash your hands or clean your hands before you touch the patient, before you do a sterile procedure, after you touch the patient, or after you touch something in the patient’s room that could be contaminated.  Those opportunities are not measured by this entry-and-exit thing at all.  There have been studies done trying to correlate entry and-exit with the Five Moments.  They haven’t been very successful.  So the reason for all this research was to develop an electronic monitoring system that will accurately measure the five moments of hand hygiene, because that’s what we were teaching our staff to do.

How did you identify the benchmarks for hand hygiene?

For our original benchmark study, we used a sampling methodology that had been developed by the World Health Organization.  So basically, what would happen is a very well trained observer would go up to a nurse or up to a physician or a nurse and say to them, “I’m here doing observations on hand hygiene.  I’m going to follow you for the next five or ten minutes.  Just do whatever you normally do.  I’m going to try to stay out of the way, but I am going to come into the room with you.”

They weren’t really collecting compliance rates.  They were looking at how many times the healthcare worker should have washed their hands and cleaned their hands in the process of care.  Those observations because a formula that was based on the patient census of that unit, and the nurse-patient ratio for that unit.  It was an algorithm to predict that denominator: the number of hand hygiene opportunities that should be going on in that unit for that day.  What we had for the numerator then was every time somebody actually actuated the soap or gel or foam. So we could calculate an index at that point.

After that, we decided we needed to validate the results of that observational study.  To do that, put video cameras in twelve patient rooms.  Patients would give consent for us to be able to do the study, and basically if a patient was assigned to that particular room and gave consent, then we would turn on the video camera and record the care 24 hours a day.  We could look at the videotape and compare that data to what the human observers had found.  We found a very high degree of correlation between those two measures, so we knew that our metrics from the original benchmark study were accurate and so that we could go forward and use this now as a denominator and actually extrapolate this and use it in other facilities.

READ MORE: How Big Data Velocity Informs Population Health, Patient Safety

We also have a lot of information about the numerator, too, and we have a third paper in the works that will be published at some point down the road.  Right now, there’s only one electronic hand hygiene system that is based on our scientifically validated results using the WHO Five Moments protocols, which is the DebMed GMS system. So it’s the only one on the market.  All the others on the market, all the other electronic ones, measure only the entry and exit, which is not completely accurate.

Continue to site...