Mike is right that the failure to distinguish “price” from “cost,” is the central conceptual failure in the current health-policy discourse. (See the post immediately below.) He might have added that the myriad ways in which our current health care finance system more or less randomizes the relationship between the two constitutes the central practical problem we face in redesigning that system.
But the problems are even more complicated than Mike allows.
If we’re going to insure people against the risk of getting sick and needing expensive care, then it’s going to be true by definition that the prices they face aren’t the same as the costs they incur. That’s what “insurance” means. That in turn implies that the decisions by patients who aren’t paying the full cost, advised by physicians whose own revenue stream is a big part of that cost, won’t in general be socially optimal, suggesting that someone or something needs to be empowered to audit those decisions. No conceivable system could maximize at the same time choice, risk spreading, and cost containment, so any actual system will be an uncomfortable compromise.
Even if insurance weren’t a problem, “cost” in the healthcare setting isn’t at all a straightforward concept, because for many of the goods and services involved the average cost is way above the marginal cost (i.e., the cost of the next unit). Much of the cost of the healthcare system is the cost of availability of services which, once made available, can be used at small or no additional cost.
A physician’s time is expensive partly because it reflects the cost of medical education, including both college and medical school tuition and foregone earnings during education and post-graduate training. For physicians whose calendars are crowded, one patient’s use of that resource conflicts with another’s; consumption is “rival.” But for physicians who aren’t as busy as they’d like to be, including those forced into early retirement by the high fixed costs the malpractice-insurance system puts on being in business at all, the actual cost of seeing one more patient or doing one more procedure is actually fairly low.
An MRI machine is expensive to have. So are some of the other pieces of fancy diagnostic hardware that go into a modern hospital. So is the clinical laboratory. If those facilities are being used to capacity, so doing a scan of X means not being able to to it for Y — “rival” consumption, again — then the cost of that test clearly includes its pro-rata share of the capital cost of the machine and the fixed cost of the staff that constitutes its infrastructure. But when those facilities have slack capacity, which is usually the case because they’re “lumpy” goods (if a hospital’s patient load only keeps a CAT scanner busy half the time, it can’t compensate by buying half a CAT scanner), then consumption is non-rival, and the social cost of running one more patient through the mill is the (much smaller) marginal cost of the time of the technicians and the physician who interprets the results. The extreme case is in the laboratory, where ordering an extra test sometimes means doing more actual chemistry and sometimes means simply asking the machine to print out the results of a test that was run as part of a routine automated battery, whether “ordered” or not. So it makes sense to ask how many MRI machines a city needs, but (unless some hospital is in the position where it has to consider buying a second unit) much less sense to ask whether a given MRI was necessary once the machine is in place.
The extreme case, of course, is prescription drugs. Some of the biologicals are actually expensive to make. But in general the cost of pressing an extra pill is measured in pennies. The cost was incurred in making that drug available: inventing it, putting through its tests, getting it through the FDC, spreading the word about it to the physicians who might want to prescribe it, and covering the costs of all the “dry holes” in the form of seemingly-promising candidate drugs that failed to meet the grade after consuming millions of dollars in R&D expense. So it makes sense to worry about the side-effects of drugs, including the side-effects on other people of building disease-resistant pathogens by spraying antibiotics all over the place. But (except for the biologicals) it doesn’t make sense to worry about the cost of the drugs, because the marginal cost is close to zero.
That leads to two sets of irrationalities, one on the demand side and one on the supply side.
When the system extracts the full average-cost price for a low-marginal cost good or service, medical decisions are distorted. (From the patient’s viewpoint, they’re distorted in the opposite direction from the distortion caused by risk-spreading, so to some extent the errors offset. But the insurers have the same incentive to eliminate “unnecessary” high-price but low-marginal-cost activities, such as prescribing the latest anti-psychotic instead of an older one that’s out of patent, as they do to eliminate truly costly interventions.)
On the flip side, the owner of a high-price, low-marginal-cost resource — a drug company with a patented drug for which it can charge a monopoly price or a physician with a fancy diagnostic machine in his office — has a strong incentive to encourage the use of that resource, regardless of its actual medical value or danger.
All this creates complexity. We will have better health-care delivery to the extent that we can make this complexity digestible by the political system, along with four key points, three of them hinted at but not stressed in Mike’s analysis, and the third omitted entirely, as it usually is in this discussion:
1. Health is a superior good. The richer we get, the higher the fraction of our total income we want to spend on it. In particular, it’s not nearly as subject as most ordinary consumer goods to competitive consumption (except for aesthetic treatments) or to the addictive process by which people become habituated to consumption patterns, converting pleasures (enjoyed when present) into comforts (noticed only when absent). So we shouldn’t either expect or desire to reduce the share of health care expenditures in GDP, except when we can do so by true efficiencies: getting the same value while incurring less cost.
2. Health care is only one means of producing health, and at the current margin probably not the one with the highest payoff per additional dollar. Improving the functioning of the markets in which people buy advice and services that help them maintain their own health, from gym memberships to healthy foods (including restaurant menus) and books on nutrition, may have better prospects for extending life and reducing disease and discomfort as tinkering with health-care finance.
3. On the other hand, the current health-care finance system spreads economic misery and economic insecurity, and supports lots of other distortions in decision-making. (There would be less need for a tort system to sort out who is liable for paying the health-care costs due to an auto accident or a bad medical outcome if those costs were borne centrally. The labor markets would function more smoothly, and self-employment and entrepreneurship would be more attractive, if people didn’t have to choose their jobs in part depending on their anticipated health-care needs.) So making health insurance and health care cheaper and more predictable for individuals and households is an important objective, even if reducing the share of health care in GDP isn’t.
4. One source of inefficiency is outright theft: true health-care fraud, as opposed to the more nebulous waste and abuse. (Billing for services never provided, for example, or treating non-existent conditions.) Malcolm Sparrow, who has studied the problem closely guesstimates that something like 10% of the $1.7 trillion we spend on health care each year is flat-out stolen; the “durable medical equipment” trade is especially corrupt. And, as Sparrow points out, auditing systems aren’t the same as fraud-control efforts; the current arrangement has too much auditing and not enough anti-fraud enforcement. The neglect of the fraud issue in the health-care finance debate is the counterpart of the tendency to overstress fraud in thinking about how to shrink the costs of government. This is a case in which rabble-rousing populism could actually do a lot of good.