Regional Practice Patterns and Health Spending

I admit it, I am a bit of a nerd when it comes to research.  I especially enjoy innovative and elegant research techniques that help us uncover deeper explanations for certain phenomenon and behavior.  Geographic spending variation has been intensely debated since the initial Dartmouth Atlas work was published.  A new paper from the National Bureau of Economic Research explores the contribution of physician practice patterns.  (NBER Paper)   As with any health spending issue, the geographic variation problem can be decomposed into two primary components–unit price and quantity of utilization.  In Medicare, unit price is basically identical nationally and is set by fiat, so practice variation in Medicare is largely due to different levels of service use.  There are a large number of analyses demonstrating this regional variation in service use, both in Medicare and commercial populations.  Utilization could differ because of real health needs of patients, patient treatment preferences or physician treatment preferences.  While health status and other socioeconomic and sociodemographic factors are associated with some utilization variation, those factors don’t explain more than a small fraction of the regional differences.  The authors of this paper focus on physician-driven utilization differences.  Even in the same geographic area, there can be signification differences in how individual physicians treat patients.  They used 15 years of Medicare data to examine heart attack treatment patterns for cardiologists who moved from one area to another.

Two extreme hypotheses were tested.  One is whether a particular physician’s treatment pattern is so ingrained that even when he or she moves, the same pattern persists.  The second is whether physicians are so influenced by the standards of practice in the geographic area that even if they had a very different practice style, once they move they quickly conform to the standards in their new location.  About 20,000 cardiologists were included, 15% of whom moved during the study period.  Heart attacks can be treated in a so-called aggressive manner using catheterization, or more “conservatively” using drugs.  Relative use of these approaches varies greatly across geographic areas.  The primary result was that cardiologists who move into a region with a different standard of practice than the doctor’s prior location change their treatment behavior very rapidly to conform with the standard in the new area.  The change happens quickly and completely–there is little further change beyond the first few months after the move.  For example, if a doctor formerly in a conservative treatment area moves to an aggressive one, that doctor’s propensity to treat aggressively rises by two-thirds.  And it does not appear that physicians selectively move to regions that comport with their prior practice standard.  Physicians who move later in their careers are as likely to change practice pattern as those who changed early.   It does appear, however, that doctors who move from a more intensive practice area to a lower one are less likely to change their practice pattern than those moving from low intensity to high intensity regions.  That particular finding simply has to be a matter of financial incentives.

The most interesting aspect of the study is the confirmation of what intensely social creatures we are; with much of our behavior shaped and reinforced by interactions with other humans.  There is no reason to expect doctors to be different.  And if financial incentives coincide with environmental behavioral influences; a change in behavior is even more likely.  One way to encourage doctors to make more individualized decisions is to train them in medical school about the impact of practice environment on their treatment decisions, so that they can more explicitly recognize these influences and then hopefully discard them and make judgments solely on the basis of what is best for a patient.  There are numerous follow-up questions that would be extremely helpful to understand–how exactly does a regional practice approach come about–are there a few key “opinion” leaders?  What encourages other clinicians to follow the leaders?  How is a practice pattern sustained if the leaders depart or retire?  How can a pattern be changed?  We seem to be homing in on a better understanding of all the factors that might influence a practice pattern.  Ultimately we should want to remove all of them other than what truly provides optimal patient outcomes.