Shocking news: incenting physicians towards one goal has negative impacts on other goals

I am so surprised by this shocking, impossible to predict, outcome:

AMNews: March 17, 2008. CMS metric may prompt excessive antibiotic use ... American Medical News

...A new study says physicians are 39% more likely to misdiagnose hospital patients as having community-acquired pneumonia due to the high-stakes environment fostered by mandatory public reporting of quality measures -- in this case, whether pneumonia patients got antibiotics within four hours of arriving at the hospital.

The results, published in the Feb. 25 Archives of Internal Medicine, are similar to those found in a Chest study published last year and echo many physicians' complaints about the measure of initial antibiotic timing, known as door-to-needle time. A February 2007 Infection Control and Hospital Epidemiology study tied excessive use of antibiotics encouraged by the performance metric to a severe outbreak of Clostridium difficile at a small rural hospital...

I trust my sarcasm is sufficiently dripping.

The CMS pay-for-performance programs will be eventually recognized as the most harmful health care innovation since Evaluation and Management coding killed primary care in the 1990s.

Alas, that recognition is still 18 months away -- and here I'm being atypically optimistic. E&M coding, for example, has now been a 13 year disaster.

There are many ways to improve health care quality that have been shown work and to be relatively free of perverse consequences. "Pay-for-performance", however, is an ideologically driven program that was reasonable to experiment with -- before the evidence of failure emerged. Now it needs to die -- but it won't until the Bush team leaves CMS.