Monday, January 21, 2013

Semmelweis was right!

We do a lot of blogging about hand hygiene, and Eli frequently points out the dearth of well-designed studies examining improvement approaches and/or the relationship between HH improvement and reduced infection rates.

In that context, I’d like to recommend this quasi-experimental study from Kathy Kirkland and her colleagues at Dartmouth, published last month in the BMJ Quality & Safety journal. Eli can inform us as to whether it would have “made the cut” for the 2011 Cochrane Review, but I really like this report. The interventions and setting are clearly described, more than one measure of HH is used (direct observation + product use), infection rates are reported over the entire time period, and the discussion section is thoughtful.

There are limitations to the work, but I encourage those who haven’t yet seen it to read it. Mike can add his experience here, as I think his group does more monthly observations, and has had a similar HH journey. I’ve also long been skeptical that rates of 90% can be achieved or sustained in settings where observations are truly clandestine (as I think in most hospitals the HH observers quickly become quite familiar to unit personnel). I’d love to be disabused of this skepticism, if it is indeed misplaced.


Figure from Kirkland KB, et al. BMJ Qual Saf 2012;21:1019-26.

2 comments:

  1. Thanks for highlighting this very nice study Dan. I think the data supporting a link between better hand-hygiene compliance and lower infections is strong. However, we have little evidence supporting specific interventions for improving compliance. Thus, I think future studies should be targeted towards identifying ways to increase HH compliance and sustain that increase.

    To you last point about sustaining HH compliance above 90% being difficult. I think the Joint Commission backing off a specific compliance target in the most recent NPSGs is one indication that your skepticism might be warranted. Although, this wouldn't rule-out outlier hospitals that can achieve high rates of compliance. Finally, since we don't have evidence-based methods to improve HH compliance, perhaps we have to hold off on defining a target compliance rate or a threshold until we have the proper tools to get us there.

    ReplyDelete
  2. I can attest to the fact that presence of the "HH Police" spreads faster than wild fire. The problem is that even with overt observations, we still get low rates. Doesn't really look like Hawthorne is at play. My theory is that for Hawthorne to apply, staff need to care about achieving high rates and more importantly know when they are to clean their hands.

    ReplyDelete