In 2009, during the “Obamacare” debate that was dominating the news, Atul Gawande wrote an article in the New Yorker that was widely praised and cited, including by president Obama himself. The article is a thought-provoking discussion of why some communities in the US have much higher health care costs than other regions. I took two main conclusions from the article.

The first is the success of the Mayo model – organizing care as a team approach. The idea here is to pool optimal expertise in the care of each patient. Greater expertise leads to “more thinking and less testing,” as Gawande puts it. I agree with this. It takes expertise to be comfortable not doing a test. Often testing is ordered because a physician does not feel secure in their diagnostic assessment.

The second main conclusion was the McAllen model, a town in Texas that has double the average Medicare costs per capita in the country. Gawande concluded that these increased costs are likely due to the culture of medical practice in the region, leading to greater unnecessary care and procedures. He wrote:

The Medicare payment data provided the most detail. Between 2001 and 2005, critically ill Medicare patients received almost fifty per cent more specialist visits in McAllen than in El Paso, and were two-thirds more likely to see ten or more specialists in a six-month period. In 2005 and 2006, patients in McAllen received twenty per cent more abdominal ultrasounds, thirty per cent more bone-density studies, sixty per cent more stress tests with echocardiography, two hundred per cent more nerve-conduction studies to diagnose carpal-tunnel syndrome, and five hundred and fifty per cent more urine-flow studies to diagnose prostate troubles. They received one-fifth to two-thirds more gallbladder operations, knee replacements, breast biopsies, and bladder scopes. They also received two to three times as many pacemakers, implantable defibrillators, cardiac-bypass operations, carotid endarterectomies, and coronary-artery stents. And Medicare paid for five times as many home-nurse visits. The primary cause of McAllen’s extreme costs was, very simply, the across-the-board overuse of medicine.

Is that, however, a necessary conclusion from that data? The data support the conclusion that McAllen (the highest cost region) uses many more medical procedures than El Paso (the lowest cost region), but does that necessarily equate to “overuse” of medicine? Evidence does not support the conclusion that the population in McAllen is sicker than El Paso, but it is also possible that El Paso simply underdelivers care.

This is a very important question as we grapple with the rising costs of health care. There are many possible sources for this rising cost, including: increased availability of expensive technology, the aging population, defensive medicine and tort costs, public health problems such as obesity, overuse of medical procedures, poor access to care leading to delayed treatment or inefficient treatment, and the use of ineffective or worthless treatments (i.e. just about everything considered CAM). To make a significant difference and stem the rising tide of health care costs we probably have to address all of these issues.

However, it is estimated that about half or more of the rise in health care costs are due to advances in medical technology, and that this technology offers reasonable value. The core problem of health care costs may simply be that we have the technology to deliver more care than we can afford to.  Combined with an aging population and a culture of individualism that often demands the most aggressive care for every person, that explains a large piece of the rising cost. It is not clear what the potential for savings is from increased efficiency and avoiding unnecessary care (although obviously we need to work towards greater efficiency).

Let’s get back to the question begged by Gawande – is the increased use of procedures in McAllen due to overuse? A recently published review of the literature does not support this conclusion. Keyhani et al performed a systematic review of the literature and came up with only five relevant articles comparing the overuse (not just use) of medical procedures in different regions. Here are the results:

Five papers examined the relationship between geographic variations and overuse of healthcare services. One study in 2008 compared the appropriateness of coronary angiography (CA) for acute myocardial infarction in high-cost areas versus low cost areas in the Medicare population and found largely similar rates of inappropriateness (12.2% vs. 16.2%). A study in 2000 using national data concluded that overuse of CA explained little of the geographic variations in the use of this procedure in the Medicare program. An older study of Medicare patients found similar rates of inappropriate use of CA (15% to 17% vs. 18%), endoscopy (15% vs. 18% 19%), and carotid endarterectomy (29% vs. 30%) in low-use and high-use regions. A small area reanalysis of data from this study of 3 procedures found no evidence of a relationship between inappropriate use of procedures and volume in 23 adjacent counties of California. Another 2008 study found that inappropriate chemotherapy for stage I cancer was less common in low-cost areas compared with high-cost areas (3.1% vs. 6.3%).

They make two conclusions from these results. One is that there is insufficient data to really answer this question definitively – only five studies. We therefore need to conduct further research into this important question. But, what evidence we do have does not support the conclusion that regional differences in utilization of medical procedures are due to overuse. Differences in overuse were slight and simply did not explain the disparity in health care use and costs in the regions compared. Their review suggests that Gawande’s assumption may not be warranted. It suggests, at the least, that the story is more complicated.


Like any complex issue the answer is probably “all of the above.” In other words – the problem of rising health care costs has many causes all of which need to be addressed if we are going to have a significant impact. We need to be realistic about the potential of addressing each contribution, however. It is tempting to blame the culture of medicine and overuse of procedures, because that is entirely fixable. I also completely agree with Gawande that optimizing health care to take maximal advantage of  expertise and specialists is a good way to go. Specialty care is generally more cost effective.

We probably, however, will have to face the uncomfortable truth that a major contribution to rising health care costs is that we simply have the technology to deliver more care, and people want this care. This leads to the further conclusion that one primary mechanism to constrain health care costs is to ration care – something no one wants to do, and for which there is little political incentive to address. For example, one obvious way to ration expensive health care is to limit futile end-of-life care. We currently lack an efficient mechanism to enforce rational decisions about the utility of very expensive care for individuals who are very likely at the end of their life due to a catastrophic illness or event. Most of the time the physicians and families get together and do make reasonable decisions, but sometimes this does not happen, and millions of dollars of futile health care can result. Any possibility of addressing this issue, however, was killed during the “Obamacare” debate by the dismissing of this issue as “death panels.” This very important issue then became politically untouchable.

Another potential way to address the driving of health care costs by advancing technology is to adjust our research priorities. It is now reasonable to ask whether or not we should fund research that has the potential to make small incremental improvements in outcomes with expensive treatments or procedures. Perhaps we should give priority to research that has the potential of replacing an expensive treatment with a cheaper alternative, or prevent the need for the expensive intervention. Cost effectiveness is already a factor taken into consideration, but it can be given higher priority. Adjusting research, however, will take decades to have a significant impact.

One conclusion, I think, is clear. Whatever the ultimate solutions turn out to be, our best chance of getting there is to follow the science and evidence. This not only applies to medical practice, but the meta questions about the strategic organization and deliver of health care itself.


Posted by Steven Novella

Founder and currently Executive Editor of Science-Based Medicine Steven Novella, MD is an academic clinical neurologist at the Yale University School of Medicine. He is also the host and producer of the popular weekly science podcast, The Skeptics’ Guide to the Universe, and the author of the NeuroLogicaBlog, a daily blog that covers news and issues in neuroscience, but also general science, scientific skepticism, philosophy of science, critical thinking, and the intersection of science with the media and society. Dr. Novella also has produced two courses with The Great Courses, and published a book on critical thinking - also called The Skeptics Guide to the Universe.