by Wyoming Liberty Staff

It has now been five years since President Obama signed the Affordable Care Act (ACA) into law. Notoriously known as "Obamacare", this highly debated reform - probably the most complicated piece of legislation in U.S. history - is still dispensing unintended consequences. With yet another case related to the ACA being heard by the Supreme Court, one is inclined to wonder if this piece of legislation is producing more new lawyer-paid hours than hours of health care.

More important, though, is the effect that the ACA is having on the American health care system. There is no longer any doubt that the reform has made access to health care more costly, especially through health insurance. There is no shortage of examples of cost drivers: in 2009 Aaron Yelowitz of the Cato Institute rightly predicted that the ACA would be "a bad deal for young adults"; in 2011 Cato Institute senior fellow Jagadeesh Gokhale pointed to the ACA-created pending cost increases in Medicaid; in another 2011 Cato Institute report, senior fellow Michael Tanner identified massive cost excesses already a year after the ACA went into effect; in 2012 a consulting report requested by the Association of Health Insurance Providers pointed to long-term cost increases from the reform; in 2013 the Department of Health and Human Services acknwoledged hidden cost drivers, which were well explained by Congressman Paul Ryan.

These examples do not come close to a full picture of the impact of the ACA on the cost of health insurance in the United States. They do, however, point to an overwhelming trend where regulations, fees, taxes and politically created, artificial incentives wedge themselves in between health care consumers and health care providers.

For understandable reasons, much attention has been given to the effects that the ACA has on health insurance. After all, in 2013 approximately two thirds of working-age Americans still relied on private insurance plans for their access to health care, access that undoubtedly and drastically is changed by prohibitively costly insurance premiums. Unless the drivers of rising insurance costs are removed, there is a significant risk that Americans in large numbers will find themselves priced out of health care access.

To reinforce the seriousness of this point, there is some indication that American health care itself could be under negative pressure from the rising costs of health insurance.

Media reports sound the alarm about pending shortage of health care providers; a recent example is from the Forbes magazine. So far, though, it does not appear to be the case that Americans in general have fewer physicians and surgeons to choose from. But this does not mean that health care is not negatively affected by Obamacare. On the contrary.

According to data provided by the Bureau of Labor Statistics under its Occupational Employment Statistics (OES) program, since the ACA was passed there has been no significant change in the density of physicians and surgeons in the nation as a whole:

• In 2009 there were 572,000 physicians and surgeons working in the 50 states, a density of 1.87 per 1,000 residents;
• In 2013 there were 602,000 physicians and surgeons working, a density of 1.95.

In other words, if anything, the national supply of MD-level health care services seems to have increased marginally with the ACA.*

However, a disaggregation of the data down to the state level reveals a potentially troubling pattern: Physician Assistants replacing Medical Doctors in physician-level health care provision.

Medical doctors seem to be flocking to some states: Utah, Washington, Rhode Island, North Dakota and Georgia have all experienced sharp increases in doctor density. Others, primarily Minnesota, Kansas and Oregon, have seen the relative supply of physicians and surgeons plummet.

What is more interesting is the tendency to shift health care provision from MDs toward Physician Assistants. Figure 1 above shows the change in core density of medical providers from 2009 to 2013. Core density is the number of MDs per 1,000 residents (as opposed to expanded density which includes PAs). State-level changes in the core density are reported by the red trend line.

The figure also reports changes in the share of physician providers, i.e., MDs plus PAs, that are PAs. For example, in 2009 PAs constituted 13.3 percent of all physician providers in Kansas; in 2013 their share was 27.6 percent. This is an increase in the PA share by 208 percent. By contrast, in Rhode Island and Hawaii the PA share fell by about half.

As Figure 1 indicates, states where the core density of physician providers has decreased have also seen a rise in the PA share. Similarly, states where the core density has increased have experienced a decline in the PA share. This points to a cost-conscious shift: in some states less-paid physician assistants replace medical doctors, while in others it appears to be cost-advantageous to provide medicine by means of MDs.

The data reported here does not give any clear-cut clue as to what states would experience a cost-and-quality downshift toward PA provision of health care. It is also too early to definitively conclude that the cost-shift pattern is directly linked to the ACA. However, there are reasons to believe that there is a causality behind the correlation: with sharply rising health insurance premiums health care providers do not want to lose patients. To counter the drive upward in insurance costs they would logically want to shift provision "downward" in the skills hierarchy.

At least one recent commentator suggests that the ACA has a built-in incentive for health care providers to downgrade health care provision. If this is indeed the case, it means that Americans will be forced to buy less quality health care in the future - by government dictate. The risk for this is significant enough to warrant further research into this matter.

-----

*) There are minor reporting discrepancies in the OES reports from year to year, which marginally affect the accuracy of the data. To compensate for those discrepancies, the state-level data reported here does not include Arizona and Montana. Overall, the raw data provided by the OES is relatively difficult to use for the purposes of time series analysis, increasing the risk for errors in the final reporting data.