Doctor Knows Best? If Not, Who Does?

A few weeks ago, as I ate lunch with a few of my coworkers, we came upon the topic of our plans for the future.

“I’m planning on studying medicine,” I announce. That’s always been the plan coming in.

“Even after working here?” my coworker asked, alluding to the…less than flattering view of doctors one can get working in the hospital research industry.

And it’s true: sometimes it seems like doctors are the biggest obstacle to transforming the way we deliver care in this country. They fail to follow care standards. They screw up efforts to reduce supply costs. And as made famous by the landmark Dartmouth Atlas study, physicians’ discretionary decisions are a primary driver of the (presumably) unwarranted health care spending that drives up cost.

In response, the government has taken it upon itself to drive health care decision-making toward higher quality, more rational care. It’s enacted penalties for hospitals with excessive readmissions, designed a value-based purchasing program tying hospital payments to a growing number of quality indicators, and outlined 33 measures for health systems hoping hoping to achieve shared savings by forming Accountable Care Organizations.

It’s not hard to imagine these incentives filtering down to the ultimate decision-maker: the doctor. Indeed, last month the New York Times reported that the country’s largest public health care system, Health and Hospital Corporation, will begin tying their doctors’ raises to performance on quality measures. In fact, Medicare itself has even made plans to apply quality performance measures directly to physicians starting 2015, although it hasn’t quite figured out how it will do that yet.

Given the push to hold physicians more accountable for “doing the right thing”, a 2010 article from Dr. Jerome Groopman struck me as incredibly relevant. Dr. Groopman is a oncologist and staff writer for The New Yorker who has written about the intersection of medicine and behavioral economics. In this article, written when Obamacare was still making its way through a bitterly divided Congress, Dr. Groopman lays out two opposing views of how our government should influence doctors’ decisions:

1. The first, based on concepts in behavioral economics (and supported by Dr. Groopman), argues for creating structures that “nudge” physicians toward making the best clinical decision without forcing them into making one.
2. The second, supported by then OMB director Peter Orszag, believes that to truly change behavior, we need to combine the dissemination of clinical information with “aggressive promulgation of standards and changes in financial and other incentives.”

Three years later, it appears the second viewpoint has won out.

For those of us working in an industry geared toward helping hospitals respond to these incentives, it can be easy to see doctors as the obstacle to implementing widespread adherence to clinical protocols and moving the dial on these performance measures. However, sometimes I wonder if we need to slow down and think about whether it’s really as simple as getting doctors to do the right thing.

Last week, a new study reported (disconcertingly) that excessive hospital readmission rates–one of the measures CMS is already penalizing providers for–isn’t related to mortality at all. It’s not the only study that has questioned whether readmission rates are a good proxy for “good care”, and it makes you wonder how many of the other performance measures have conflicting evidence.

In theory, if we gathered enough clinical evidence and limited ourselves only to measures that have been shown to be strongly correlated with quality outcomes (assuming we can define them first), we could create care standards that we could then promulgate to physicians for adherence. But even for practices for which the clinical evidence is unequivocal on, it still might not be in our best interests to get all doctors to adhere to them. One emerging criticism of randomized controlled trials–the “gold standard” of clinical trials–is that they are based on highly controlled situations with 100% adherence to protocols and a subject sample from which non-ideal participants have been excluded. Real world medicine is rarely as clean. Yet when a health system creates clinical protocols for its physicians, or when CMS lays out performance measures for physicians across the country, we are essentially creating default options without considering the individual circumstances of each case. “Setting the default option that doctors will present to patients requires us to make value judgments,” writes Dr. Groopman. Given the uncertainly of health care, are we really ready to do that for huge swaths of the population?

The answer, I think, should certainly not be a return to the hands-off approach of fee-for-service. There’s enough evidence out there to show that significant care variation is unwarranted, and there are enough success stories of hospitals and health systems that have partnered with their doctors and achieved remarkable improvements in outcomes at reduced costs. But it should serve as a reminder to us that even if it sometimes feel like our health care problems would be solved if only doctors would consistently do the right thing, often times what “the right thing” is isn’t so clear.

Then again, maybe in the future, we’ll just have supercomputers like Watson tell us what to do.