Can markets work in medicine?
by Christopher J. Conover, Forbes, August 20, 2012
One of the sharpest dividing lines between conservatives and liberals is whether or not markets can work in medicine. Progressives admit to being "deeply suspicious of the claim that a health care system dominated by powerful vested interests and mystifying in its complexity can be tamed by consumers who are strapped for time, often poor, sometimes uneducated, confused and afraid." But just how valid are these concerns?
First, it is not at all clear that "powerful vested interests" are any larger in health care than in other sectors of the economy, such as restaurants and retail, where we are perfectly content to let market forces discipline the power of large corporations to set prices. Of the 22 U.S. firms with more than $100 billion in revenues in 2012, for example, only three are predominantly health care firms.[1]
"To the average consumer, is health care anymore mystifying in its complexity than, say, computers, cars, or cell-phones?" -Christopher J. Conover
To the average consumer, is health care anymore mystifying in its complexity than, say, computers, cars, or cell-phones? Purchase decisions in these markets may strike some as mind-numbing in their complexity. Yet, we nevertheless rely on consumers to figure out what is best for them, and the magic of the market generally results in steadily lower prices and steadily rising quality over time. Part of the reason for that is that markets do not require every consumer to be well-informed. And indeed, many are not. But all benefit from the diligence of the minority who comparison shops with great vigor. We also get a leg up from the efforts of market facilitators-such as Consumer Reports-which greatly reduce the amount of time and effort needed by consumers to get sufficiently informed to make reasonable purchasing decisions. And these benefits hold true in markets ranging from high-tech to hot dogs.
Health spending (unlike that of gadgets and snack) happens to be highly concentrated, with the top 1 percent of spenders each year accounting for more than one quarter of all spending. Still, the concentration of these expenditures doesn't impede the market's efficacy. How do we know this? Decades ago, the RAND Corporation-nonprofit think tank-conducted a very careful scientific experiment on the very subject.[2]
In 1974, RAND researchers randomly assigned over 7,000 individuals into various types of health insurance policies: ones with completely free care (no cost sharing), ones with a modest deductible (e.g., $200) and 25 percent cost-sharing, and ones with the equivalent of high deductible policies. All the cost-sharing policies had a maximum upper limit on out-of-pocket spending-meaning once a family spent 10 percent of income on health care, the policy paid 100 percent of remaining medical bills for the balance of that year.
The randomized design helped ensure that observed differences in medical care use, spending and health status could properly be attributed to differences in insurance plans rather than an outside factor (e.g., sicker people self-selecting into the plan offering free care).
The observed differences between these different groups were remarkable. Average health spending for individuals receiving free medical care was 32 percent higher than for those who had to pay a quarter of the bill out of pocket. Some of this extra care admittedly was valuable, but fully 93 percent of that spending difference came in the form of "waste" rather than added value to the patient. The kicker was that for the average patient, there was no measureable difference in health status between the two groups.[3]
Remember that these were randomized groups, so they were equally likely to have "heavy hitters" who accounted for the lion's share of spending. Yet, allowing patients to face the discipline of market forces rather than shielding them entirely (as is done in Canada, where cost sharing for health services guaranteed under federal law is prohibited) has the potential to reduce spending by nearly a third.
More remarkable still is that the RAND study was a "demand side" experiment that only enrolled a few thousand families across the country. Thus, there was no incentive for the supply side to change in response to the relatively small number of patients incented to shop more wisely in the medical market. In a world where many more patients were fully responsible for front-end health spending and used health insurance only for catastrophic back-end expenses, we would expect to see the kind of fierce competition (and even price advertising!) we see in markets for Lasik surgery, for example. This makes the potential for savings from a market-oriented system relative to a single-payer "free care" system such as Canada's even higher than those discovered by RAND.[4]
In short, we have substantial solid scientific evidence that markets can work in medicine. Instead of supplanting the judgment of consumers by requiring them to buy the medical equivalent of Hondas instead of Yugos, we should empower patients and harness the extraordinary ability of Americans to find good value for their money. The prescription for our health care reform ills is economic freedom.
Too many reformers seem unwilling to accept that consumers have heterogeneous taste when it comes to health. Given the other ways in which I can spend my money, I might get much more value from spending $500 on food, clothing, or shelter than on government-mandated preventive health services, for example. Thus, giving health care consumers more freedom of choice-which, after all, is exactly what markets are all about-is not only smart, it's the right thing to do.
Footnotes
[1] These include #18 CVS Caremark, #21 Cardinal Health, and #22 UnitedHealth Group. #6 General Electronic admittedly makes medical equipment, but is not predominantly a health care firm.
[2] The experimental nature of the RAND study is one of the elements which sets it apart from others in this field. The vast majority of evidence in health policy is observational in nature. That is, researchers compare people in different groups (e.g., for-profit & non-profit hospitals, people with private health insurance & those with no coverage, etc.) and make observations about them. Such observations yield a wealth of correlations, but it is much harder to pin down causation. For example, researchers have observed that people without health insurance tend to have a higher risk of death than those with private coverage. Yet, it is difficult to be certain that this is because they lack coverage or whether some other factor is at work. Those willing to gamble on being without coverage might also be willing to take more risks in other areas of life such as driving without seatbelts, smoking, or other risky behaviors. So it could be that their greater propensity to take risks rather than lack of coverage per se explains their higher death rates. But the RAND study side-stepped these problems by using the gold standard of scientific evidence-a randomized controlled trial, just like those that the FDA relies on to provide definitive scientific evidence regarding the efficacy and safety of pharmaceuticals.
[3] In contrast, low income patients with free care had somewhat better health status than those in cost-sharing plans: they had better control over their blood pressure and better corrected vision. These modest improvements in health status were not large enough to justify the huge spending difference between the two groups-that is, the health status of the two groups hypothetically could have been equalized by much less expensive interventions targeting hypertension and poor vision. But the study at least signaled that caution was in order when trying to use cost-sharing as a strategy for averting wasted health spending among poor people. This is why cost-sharing under Medicaid is so extremely modest (e.g., $1 for a doctor's visit for which most insured patients might have to pay a $20 or more copay) and why most states instead rely on managed care to hold down use and spending among this population.
[4] Admittedly, a single payer system offers administrative savings, but a) the size of these savings is grossly exaggerated by single payer enthusiasts and b) could never possibly exceed 32 percent.