Understanding the root causes of excessive health spending and geographic variation in spending continues to be a difficult task. Many factors have been suggested. Because physicians are responsible for ordering most health care, their behavior naturally comes under scrutiny. A study from Harvard researchers follows up earlier work, examining the role of both patient preferences regarding care intensity and physician practice styles and beliefs. (Harvard Study) The study used survey responses from Medicare beneficiaries and physicians as well as hospital referral region-level end-of-life and overall spending data from Medicare claims. The patient survey data asked questions designed to elicit preferences on end-of-life or serious care. The physician survey sought reactions to care vignettes from cardiologists and primary care doctors. It also asked about frequency of follow-up visits after hospitalization for a serious cardiac problem. Note that the survey sample sizes were small on a per HRR basis, which can be a problem since there is substantial intra-HRR variation.
On the basis of the survey responses physicians were categorized as cowboys, who almost always used aggressive care, or comforters, who rarely did. About 25% of the cardiologists were designated as cowboys, as were 22% of primary care doctors. 27% of cardiologists and 47% of primary care physicians were categorized as comforters. HRRs were then categorized according to the survey responses from the doctors in the HRR. Then spending was compared across the regions. Patient care preferences accounted for very little of the variation in end-of-life or general spending. End-of-life spending is positively-related to a preponderance of cowboys and to an increasing fraction of doctors who do follow-up visits more frequently than guidelines recommend, and negatively-related to a preponderance of comforters. In fact, 36% of variance in end-of-life spending appeared to be accounted for by the cowboy/comforter ratio. The researchers conclude that 35% of end-of-life Medicare expenditures and 12% of overall Medicare spending could be eliminated if all doctors were comforters and met follow-up visit guidelines.
Demographic and other factors did not appear related to physicians having a cowboy practice style. Being in solo or 2 person practices was associated with a much greater likelihood of being a cowboy, but few other examined structural or other variables were. So why do some doctors hold the practice beliefs that they do? That is a significant unanswered question, although the researchers suggest that the cowboys believe in the effectiveness of the treatment they are ordering. And it could well be that since guideline-driven medicine does not fit all patients and may not lead to the best outcomes for all patients, that at least in some cases the cowboys are actually doing what is best for their patients.