Study May Refute Theory That Physician Supply Boosts Healthcare Spending

May 14, 2010

May 13, 2010 — When the Obama administration set out to reform healthcare, it based its game plan partly on the work of researchers at Dartmouth Medical School who say that huge geographical variations in per-capita healthcare spending represent a huge opportunity to cut costs.

The investigators in the so-called Dartmouth Atlas Project have reported that per-capita Medicare expenditures in some regions are more than double those in other regions, but that patients in the higher-spending regions do not necessarily enjoy better care. They contend that these geographical differences hinge in large measure on how many physicians and hospital beds a region possesses. More healthcare resources help trigger higher per-capita spending.

As Peter Orszag, director of the White House's Office of Management and Budget, once put it, "supply appears to generate its own demand."

An article published online May 12 in the New England Journal of Medicine, however, puts a dent in this supply-side theory. The authors found that "differences in the supply of medical resources are neither significant nor quantitatively important."

"We understand that our conclusion is contrary to conventional wisdom," lead author Stephen Zuckerman, PhD, told Medscape Medical News. "Our study basically says those supply differences aren't driving the spending."

Much Still to Be Learned About Geographic Variation

Dr. Zuckerman, a senior fellow in health policy at the Urban Institute, and his coauthors analyzed Medicare spending — considered a proxy for all of healthcare — from 2000 through 2002. Following in the footsteps of the Dartmouth Atlas Project, they grouped geographic regions into quintiles based on their per-capita Medicare spending. When Dartmouth investigators did this, they found spending per beneficiary was 52% higher in regions in the top quintile compared with those in the lowest quintile, after adjusting for baseline differences in health status, according to an article published in 2003 in the Annals of Internal Medicine.

To find out what might account for the 52% geographic difference, Dr. Zuckerman and colleagues looked at variables consisting of baseline measures of health such as body mass index, changes in health, demographic characteristics, income, and supplementary insurance. Another variable was the area-level supply of healthcare resources, measured by the number of hospital beds and physicians per 1000 elderly patients, the percentage of physicians in primary care, the number of resident physicians per hospital bed, and whether the nearest hospital with 100 or more beds was a teaching hospital.

After crunching all the numbers, Dr. Zuckerman's group found that adjusting for demographic characteristics and measures of baseline health and changes in health reduced the difference between the highest and lowest Medicare-spending quintile from 52% to 33%. Most of the reduction resulted from measures of baseline health and changes in health. What did not narrow the gap at all were measures of healthcare resources.

The authors noted that more than 60% of geographic variation remains unexplained. Accordingly, policy makers bent on cost-control need to proceed cautiously until they have a better idea of what accounts for the gap. The authors suggest several possibilities ripe for research, such as differences in how medical practices are organized, the profit-seeking behavior of providers, and cultural or social preferences of Medicare beneficiaries.

Supply-Side Theory Led to Caps on Residency Slots, Says Dartmouth Critic

One critic of the Dartmouth Atlas Project views the NEJM study as a corrective to health policy that has limited the number of physicians during the last 13 years.

Hematologist Richard Cooper, MD, a professor at the Leonard Davis Institute of Health Economics at the University of Pennsylvania, Philadelphia, points to the Balanced Budget Act of 1997, which froze the number of Medicare-funded residency training slots at 1996 levels.

"They cut back with the notion that there was a surplus of physicians," Dr. Cooper told Medscape Medical News. Congress was acting on the assumption that physician supply helped drive demand for healthcare. Narrowing the residency pipeline, it was thought, would lower demand, and therefore healthcare spending.

Dr. Cooper advocates creating more Medicare-funded residency training positions to increase the physician workforce. He contends that geographical variation in Medicare spending ultimately reflects patient income, not the supply of medical resources.

"If Dartmouth is wrong, then the caps are wrong, and if the caps are wrong, we need to expand graduate medical education," he said.

An editorial by Arnold M. Epstein, MD, is somewhat more measured in its response to the study by Dr. Zuckerman and colleagues. "The article...both supports and modifies the Dartmouth gospel," said Dr. Epstein, from the Department of Health Policy and Management, Harvard School of Public Health, and the Division of General Medicine, Section on Health Services and Policy Research, Brigham and Woman's Hospital, Harvard Medical School, Boston. Massachusetts.

He pointed out that limitations of Dr. Zuckerman's study are the lack of data about a variety of conditions as well as information about disease severity. These likely would have reduced geographic variation even more, Dr. Epstein said.

Where a Patient Lives Influences Number of Tests and Diagnoses

Investigators from the Dartmouth Atlas Project did not respond to a request for an interview by press time, but they did have something to say about geographic variations in the May 13 issue of the NEJM. Their article asserted that patients in areas of more intense clinical practice, as measured by Medicare spending during a patient's last 6 months of life, are tested more, and diagnosed with more illnesses.

The researchers grouped geographic regions in 5 quintiles based on the intensity of physician and hospital services. They tracked roughly 255,000 Medicare patients during a 5-year period who moved either from a lower-intensity quintile to a higher-intensity quintile, or vice versa.

Due to aging, these patients naturally experienced a rise in the number of diagnosed illnesses. However, the increase in diagnosed illnesses — as well as laboratory testing and imaging — was greater for patients who moved from a lower-intensity quintile to a higher-intensity quintile than for those who stayed within their quintile, or moved to a lower-intensity one.

This pattern suggests that patients diagnosed with various illnesses in regions of more intense medical practice are, on average, less sick than typical patients with these illnesses, said one of the study's authors, Elliott Fisher, MD, MPH, in a press release. Dr. Fisher is a co-principal investigator at the Dartmouth Atlas Project.

N Engl J Med. Published online May 12, 2010.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: