Here's where I have trouble with Faraday's logic.
First, many hospitals around the country have, in fact, gotten their surgical site infection rates down to zero and held them there, through extreme diligence and monitored protocols. For example, this 2010 New York State Department of Health report shows a number of hospitals with zero colon, hip replacement, and coronary bypass graft surgical site infections.
Second, these penalties don't cost hospitals very much, in part because the severity of the patients' underlying illness puts them into a higher payment code already. According to a May, 2010 paper in the American Public Health Association Journal Medical Care, CMS estimated avoided spending of a mere $20 million to $50 million a year, or no more than $12,500 per hospital, for all and any hospital-acquired adverse events.
Third, patients were identified as having a prior infection not from their medical record or physician report, but by self-report, which is prone to error. They simply answered a question about whether they recalled ever having a history of a skin infection, such as cellulitis, a skin abscess, or wound infection that required antibiotic treatment.
I'm not sure most patients could remember that. And as for treatment, the question wasn't specific enough to distinguish an over-the-counter antibiotic product like Neosporin from one that required a physician's prescription. That masked any degree of severity the patients might recall.
Fourth, the sample size, culled from patients undergoing surgery at the University of Maryland Medical Center and Johns Hopkins Hospital, was only 613, and of those, 135 reported a prior skin infection requiring antibiotic treatment and 24 developed a surgical site infection or died of infection within six months. Not a very big sample.