Thursday, January 29, 2015

Jumping Through Quality Hoops

The Washington Post reports, "The Obama administration wants to dramatically change how doctors are paid".

Essentially, they wish to move away from the standard fee-for-service model and towards payments based on "quality".

In theory, this sounds good -- after all, who could possibly be opposed to quality?  And it's true that there are perverse incentives in the current system, where doctors get paid regardless of whether they are doing a good job or not.

But the various "quality" measures and "pay for performance" incentives we've seen so far from Medicare and Medicaid have at best a tenuous relationship with what most patients would see as "quality" care.  And this is an inherent problem in any system where the person receiving the service isn't the person paying for it.

As Dr. Michael Kirsch explained last year, many of these measures have "nothing to do about real medical quality, but... everything do about cost control." 

Patient care can also suffer, as physicians have to choose between following the quality measures vs. doing what's actually right for the patient. These quality measures can introduce their own perverse incentives.

For example, there was a push a few years ago to reward ER physicians for getting patients with pneumonia on antibiotics within a certain time frame.  As an ER radiologist, our job was to call the ER physician ASAP everytime we saw a chest x-ray on a patient that showed possible pneumonia so the treating physician could start the therapy within the allotted time.

Of course, this also took time away from other equally urgent work (for which there was no bonus.)

And this might not have been good for the patients either. Officials at the Centers for Medicare and Medicaid Service admitted that some of these measures could result in "inappropriate delivery of a service to some patients (such as delivery of antibiotics to patients without a confirmed diagnosis of pneumonia), unduly conservative decisions on whether to exclude some patients from the measure denominator, and a focus on meeting the benchmark at the expense of actual improvements in quality or patient outcomes."

As we see more quality measures introduced, we'll also see Goodhart's Law in action: "When a measure becomes a target, it ceases to be a good measure."  People will work to meet the metric, but that will have increasingly less value as a measure of whether actual quality care is being delivered. 

(Think of school teachers who teach "for the test" in order to maximize their students' scores on the standardized tests, and how that corrodes real learning.)

There's much more to be said on this topic, which I'll have to leave for a later time.