Impact of Physician and Patient Preferences on Spending

Regional variation in health spending exists.  Explaining it has created a small industry.  The latest paper, published by the National Bureau of Economic Research, focuses on the extent to which either patient demand for certain services or physician preference to supply those services may account for some of the variation.   (NBER Paper)   This paper looks only at end-of-life care and at Medicare beneficiaries, and therefore may have limited utility, since other research has shown that there are differences in variation for Medicare and other payers.  The researchers used survey data which they then linked to regional utilization measures.  Cardiologists and primary care physicians were presented with four vignettes on elderly patients with chronic illnesses and asked how they would manage those patients.  Patients were also surveyed regarding their preferences for care in various potential situations.  A further attempt was made to understand why physicians held the care preference beliefs that they did–malpractice concerns, patient pressure, peer pressure, training, believes about the efficacy of treatments, etc.  About 1400 beneficiaries were surveyed and about 500 cardiologists and 840 primary care doctors.  The research analysis suggests that patient preferences about care has only a very modest effect on regional end-of-life expenditures.  On the other hand, physician beliefs regarding treatment options appear to have a much stronger relationship with spending variation, and in particular, physician perspectives on the effectiveness of various treatment options seems to have the strongest effect on Medicare end-of-life utilization and spending.  These beliefs are often in conflict with accepted care guidelines and other evidence.  The authors say that if all care followed guidelines, Medicare end-of-life spending would be 36% lower.  The paper provides a better understanding than before about why physicians have varying treatment patterns and how much their practice styles may be based on often-erroneous beliefs about the efficacy of particular treatments.