Can you tell a hospital is safe by its "broken windows"?

This past weekend there were many discussions of James Q. Wilson's "broken windows" theory. Dr. Wilson, unfortunately, passed away this past week. The theory and research suggest that perception of a safe neighborhood prevents crime. If people feel they're in a safe place, they are less likely to commit a crime. If, however, they feel the neighborhood is unsafe or crime-ridden, they're more likely to commit crime. Thus, under this theory, police forces should arrest and prosecute even the smallest crimes, such as graffiti, and cities should quickly repair broken windows.

I lived in Rudy Giuliani's New York City during the implementation of this strategy, but I'm not here to defend his law enforcement policies one way or another, since I'm not an expert. Crime did fall, but it might have been for other reasons.  What I'm more interested in is if there could be an analogous theory in hospitals?  Is there a safe hospital theory?  It made me wonder if clinicians in safer or cleaner hospitals are more apt to practice hand hygiene or have higher compliance with CLABSI checklists.

I'm not aware of much data in this regard.  Two of the better analyses were done by Pat Stone's group at Columbia (I was a co-author). Looking at data from 415 ICUs in 250 hospitals they found that there was no convincing evidence of a cross-over effect between CLABSI and VAP; that is compliance with the CALBSI Bundle elements was never associated with a decrease in VAP rates.  In a separate paper, they found that compliance with the VAP bundle did not lower CLABSI rates. So for at least two device infections, there appears to be no such thing as a safe hospital. Lankford et al. in EID (2003) hypothesized that hand hygiene would increase after construction of a shiny new hospital. It actually decreased from 53% to 23%.  Hopefully Mike and Dan can add to this list of studies.

There has been a lot more research on what makes a quality hospital outside of infection prevention.  Twenty years ago, there was an important study in Medical Care that looked at disease-specific mortality in acute myocardial infarction, congestive heart failure, pneumonia, stroke, obstructive lung disease, or gastrointestinal hemorrhage in 30 hospitals. They found little correlation between disease-specific mortality rates within each hospital. So, MI mortality was not correlated with CHF mortality, even if they were likely to be treated by the same physicians and nurses. If mortality isn't a quality indicator, one wonders if other quality indicators have any relevance.

And what is a post without a non-scientific anecdote? Several years ago, when I was the hospital epidemiologist at a large hospital in Baltimore, we had high rates of MDR-Acinetobacter infections. This led our group to conduct a pilot study looking at the impact of universal gown and gloves in ICU-settings. At the time a new Chief Medical Officer, who happened to be a pulmonary-critical care specialist, started attending in our ICUs. He was struck at the level of gown/glove compliance that he saw and declared that he had never seen such a safe hospital. This actually meant something, since he had just moved from Barnes Jewish Hospital in St. Louis; home to one of my heroes, Vicki Fraser.  Were we really safer than BJH? I don't know, but the sight of all of those gowns and gloves did make it appear that we were really trying our best to be safe. Soon after, our CLABSI rates fell drastically, after a lot of effort sure, but the culture was changing. Maybe there is something in this theory that applies to hospitals after all? Too bad there appears to be no such thing as a "safe" hospital.