Though federal officials insist that scientifically valid formulas will be used, some hospital officials worry formulas won’t adequately adjust for, say, the patient’s socioeconomic or health literacy status, or a poorer, inner-city hospital’s lower level of resources.
“Using process measures will always be frustrating,” Mitchell says. He acknowledged that it’s like looking for lost car keys under the lamppost because that’s where the light shines, “when what we really want to know is something that’s outside that light circle. We’re getting there. We are getting bigger and bigger lightbulbs, for example, with electronic health records, which can capture more of this information.”
There also is the concern expressed by some providers about “teaching to the test”—that by focusing on just these 12 process measures, clinicians may spend inordinate time improving just those measures and ignoring care that is equally or even more relevant to that patient’s survival.
“This is Willie Sutton,” Mitchell says. “Right now, you go where the money is. The money is in declaring those results and getting them any way you can.”
Using metrics with time deadlines to rate care sometimes results in unintended consequences.
One classic example was one Hospital Compare measure that required clinicians to administer antibiotics within four hours of arrival to a patient suspected of having pneumonia, before it was known if the patient was actually infected.
“It made sense, and studies said that on average those pneumonia patients who get antibiotics more quickly do better,” Wachter says. “But no one ever thought about what would actually happen: The clinician looks at the clock and says ‘I still don’t know if this patient has pneumonia, but in three minutes he or she will turn into a metaphorical pumpkin if I don’t give an antibiotic to this patient.’”