New news on invasive MRSA rates: headed down?

I am at the annual conference of the Society for Healthcare Epidemiology of America, where authentic news was released Sunday, in a very quiet way.

Last fall, a team of authors from the Centers for Disease Control and Prevention published a paper in the Journal of the American Medical Association that for the first time estimated the burden of invasive MRSA disease in the United States. Using data from the Active Bacterial Core Surveillance System — which comprises nine sites around the country and takes in about 14 million people — the team estimated that the annual incidence of invasive MRSA in the US is 94,360 cases, including 18,650 deaths. In an important (and to some researchers controversial) contribution to defining MRSA disease, the team also estimated that 13.7% of the invasive infections were community-associated, 26.6% were hospital-acquired (which they called "hospital-onset"), and by far the largest proportion, 58.4%, were a new category they called "hospital-acquired, community-onset" — the patients acquired the bug while in the hospital, but did not develop disease until after they were discharged. (Cite: Klevens RM et al., Invasive methicillin-resistant Staphylococcus aureus infections in the United States. JAMA, 2007 Oct 17;298(15):1763-71. Full text here.)

The important event Sunday was the first release of an update to that data, adding a second year of analysis. The news: Hospital-acquired infections declined in a small but significant way, by 8.1% — though community-associated and community-onset cases did not change.

The original paper analyzed data from July 20o4, when the ABC system expanded from a pilot program to the full nine sites, through December 2005. The update both reslices that data and adds fresh numbers: It compares calendar year 2005 to calendar year 2006. It finds small downward trends as well in community-onset and community-associated cases, but judges those trends not statistically significant. But Dr. Scott Fridkin, who presented these numbers and was an author on last fall's paper, said the dip in nosocomial cases is statistically robust and a true decline, though adding a third year of data will make the analysis yet more solid.

The decline is intriguing and raises the question: If infection-control interventions in hospitals are pushing down rates of hospital-acquired cases, then why are "hospital-acquired/community-onset" cases not down as well?

As usual with conference papers, the abstract is not online; the CDC says it will be posted today and I'll update when they do. (NB: The data presented at the conference is actually an update of the data in the abstract, and so when the abstract goes up the numbers will not match; the decline presented today is a few percentage points smaller, due to a re-cut of the data after the abstract was written.)

UPDATE 8 April 08: The abstract is now online here.

(Financial disclosure: I am at SHEA for book research and to participate in a session on interactions between researchers and media. In exchange for that session, the organization paid my airfare.)