10 August 2010

News break: Hospital-acquired MRSA trending down - but why?

There's good news today in the Journal of the American Medical Association: A 4-year study by the CDC and its partners in the Active Bacterial Core Surveillance System reports significant declines in invasive MRSA infections contracted in hospitals. The study, which covers 2005 through 2008, finds a decline of 9.4% per year among infections that were contracted in hospitals and also diagnosed there, and a parallel decline of 5.7% per year in what the CDC calls "hospital-acquired community-onset" infections, ones that were acquired in the hospital but didn't become evident until after the patient was discharged. Overall, the decline over the study period of hospital-onset infections was 28%, and the decline in hospital-acquired community-onset infections was 17%.

MRSA is the leading organism in the vast national epidemic of hospital-acquired infections (HAIs), which conservatively sicken 1.7 million Americans per year and kills 99,000 of them. (Those numbers date back a decade to an Institute of Medicine report, and have been challenged by Consumers' Union as an underestimate.) So any solid indication that the epidemic is decreasing is good news. And the CDC study is a solid indication, built on a population-based survey that covers about 15 million people in 9 geographical areas.

So it's a great pity that we don't really know why MRSA has declined in this fashion. The study can't tell us. And because we don't know, we'll find it harder than it ought to be to keep the trend going in the appropriate direction.

Here's the problem: Though it is about healthcare infections, this study doesn't use data from hospitals. The study itself says: "National data describing changes in incidence in US healthcare institutions are not available." The data that hospitals report on infections that occur within their walls or result from their actions, contained in the CDC's National Healthcare Safety Network,  is voluntary, partial and anonymous; in fact, to participate, hospitals are guaranteed confidentiality. The only surveillance systems in the US where hospitals are not anonymous are the various states where legislators, out of exasperation or in response to citizen pressure, have passed laws mandating that infections be reported.

So the declines in MRSA incidence that are reported in this study can't be linked to specific practices — and that's important, because for more than a decade, American healthcare has been locked in a ferocious argument over the best way to reduce MRSA and other HAIs in hospitals.

On the one hand, there are institutions such as the Pittsburgh VA (in a project partially funded by the CDC and since adopted across the entire VA) and Evanston Northwestern Healthcare (now called Northshore University Health System) that follow some variant of "active surveillance and testing" or simply "search and destroy," which tests incoming patients for MRSA carriage and isolates and treats them until they are clear. On the other hand, there are institutions that reject "search and destroy" as too MRSA-specific (and too dependent on expensive rapid-test technology) and opt instead for broader infection-control programs with special emphasis on hand hygiene and antibiotic stewardship. (This paper by physicians from Virginia Commonwealth University summarizes the issues well.) The patients whose data ended up in the JAMA CDC study might have attended hospitals that followed either of these paths, or neither. There's no way to know.

In addition, a significant proportion of the decline in the CDC study fell into the category of bloodstream infections — which are now also being targeted by the checklist approach espoused by Macarthur Fellow Dr. Peter Pronovost and New Yorker writer and surgeon Dr. Atul Gawande, and adopted patchily across the US. Plus, there's a further confounder: Since 2009, the Center for Medicare and Medicaid Services has been applying a carrot-and-stick approach — refusal to reimburse for the extra care needed — to certain preventable hospital-caused conditions, including central-line associated bloodstream infections (which are caused by a variety of organisms including MRSA). How successful that has been, or how much influence it has exerted, has not been assessed.

So, to recap: MRSA appears to be declining in hospitals; that's good. From this study, we can't say why: That's frustrating. And, one more point: If we had truly accountable, truly transparent hospital reporting for preventable infections and other medical errors, we would not be in this data fog. Surely it's past time to clear the air.

Kallen AJ, Mu Y, Bulens S et al. Health Care–Associated Invasive MRSA Infections, 2005-2008. JAMA. 2010;304(6):641-647. doi:10.1001/jama.2010.1115
Accompanying editorial:
Perencevich EN, Diekema DJ. Decline in Invasive MRSA Infection: Where to Go From Here? JAMA. 2010;304(6):687-689. doi:10.1001/jama.2010.1125


Thomas said...

Maryn -- You can't tell from this report, but surely we can guess . . . well, at least I'd guess. . . that at least a significant part of the decline is linked to hospitals' renewed emphasis on basic hygiene, cleaning, hand-sanitizing, etc. spurred by the resistant S.aureus outbreak, and other hospital-acquired infections. What do you think?

Maryn McKenna said...

Thomas: The problem is, it's a guess (and, another guess, could also be due to a change in the behavior of the organism, as the JAMA editorial suggests). It's a reasonable assumption — but I've not yet met a hospital that is willing to invest in a new program on the basis of a "reasonable assumption." We need better data.

Maryn McKenna said...

Because - sorry, hit Send too soon - what if within"guess" you have a hospital that does basic hygiene very well? Or "search and destroy" very badly? Which is doing a better job then? My point is, we need better, more transparent data, so that we can compare "what works" between institutions and really know what is better.