I am going to suggest that in order for the U.S. health care system to be more efficient, administrative costs have to increase. The reason is that, just as in the days of Paracelsus, it is a challenge to align compensation and incentives in medicine.

Three Types of Administrative Costs

Conceptually, the administrative overhead of the health care system can be put into three categories:

-- Adverse Selection Costs

-- Claims Processing Costs

-- Supplier Management Costs

Adverse selection costs reflect the insurance industry's efforts to choose healthy patients and avoid sick ones. Some economists fret that adverse selection costs are a major factor in insurance company overhead, but there is little empirical evidence that insurance companies devote a significant share of their budgets to such efforts.

Claims processing costs are the much-maligned paperwork of the health insurance industry. In my opinion, these are too high, in part because of unnecessary interfaces but mostly because we have the wrong concept of health insurance. Just as it is self-defeating for consumers to try to rely on mail-in rebates from CompUSA, it is self-defeating for consumers to prefer health insurance plans that reimburse every form of medical expenditure.

Supplier management is the feedback loop between health care service providers and the individuals and institutions that pay the bills. I believe that supplier management needs to be strengthened, which probably means that this form of overhead will require more resources. It also makes doctors uncomfortable, so that we can expect that they will put up resistance.

How to Reward Doctors

Paracelsus was told that he would be rewarded on the basis of outcome. However, the canon reneged and decided instead to reward him on the basis of effort. In all, there are four ways to reward health care suppliers.

-- Capitation -- pay a fixed amount per year per patient

-- Outcome -- pay on the basis of the health of patients

-- Effort -- pay on the basis of labor and capital costs of providing services

-- Process -- pay on the basis of adherence to guidelines

We can assume that every health care supplier would prefer a higher ratio of reward to effort. Then no compensation scheme will work perfectly.

The problem with paying on the basis of outcome is that outcomes often are outside the control of the physician. The ailment's severity can have a subjective component, such as pain, which is difficult to measure. Even if outcomes are observable, they may depend on genetic and other factors that do not reflect the doctor's skill or efforts.

Physicians will likely adapt to outcomes-based compensation by attempting to see only the easiest-to-treat patients. They will avoid complex cases or severe illnesses, because such cases will offer little reward and high cost.

The incentive to treat expensive or complex illness is similarly lacking under capitation. Thus, it is not surprising that capitation-based systems, such as HMO's, produce friction when patients have severe needs. Under capitation-based systems, as long as you are reasonably healthy, you can get the care that you need. However, once you develop an expensive illness, you run into conflicts with the budget constraints of the health care supplier. The recent Canadian Supreme Court decision giving Quebec citizens the right to obtain private health care shows what happens when the supply of health care is based on a fixed budget with rationing. The court found in favor of a plaintiff who had to wait a year for hip replacement surgery, saying that he should have had the right to seek care elsewhere.

The opposite problem exists in the United States, where effort is the primary basis for compensation. Under our system, a patient who requires a lot of procedures represents a cash cow, not a cash drain, to medical suppliers. We pay for procedures, and so we get lots of procedures. The question is whether the procedures that we get are necessary and cost-justified.

How can suppliers be incented to balance costs and benefits in choosing treatment plans? One solution is to study the results of different treatment approaches, in order to indentify the best practices, based on criteria of costs and benefits. This information can be used to construct guidelines, and doctors can be compensated according to how well they adhere to guidelines. That is what I call process-based compensation.

The challenge with implementing process-based compensation is that it requires overhead. Someone needs to invest in gathering data on how treatment practices affect outcomes. Someone needs to audit doctors to determine how well they adhere to best practices.

The impetus for process-based compensation has to come from whoever pays for health care. If consumers were paying directly for health care, then they would have to demand data from their doctors, who in turn could look to research centers to provide information. If private insurance pays for most health care, then insurance companies would be the natural parties to fund the research into medical guidelines and to monitor doctors' compliance with those guidelines. If government pays for most health care, then government will need to develop guidelines and monitor compliance.

Compatible Insurance

I am sure that many doctors do not want to see anything like a medical guidelines commission. From the perspective of a proud, trained, experienced professional, formal guidelines threaten one's independence and professional judgment.

When insurance companies attempt to influence the medical process, doctors chafe. It is understandable that doctors would prefer to have complete autonomy, free of the "administrative overhead" of having to justify their decisions to insurance companies. But eventually, in order to align costs and benefits in health care, we are going to have to implement process-based compensation.

The insurance companies could be given a less powerful role under a system of "event-based" health insurance. See my essay or Susan Feigenbaum's similar proposal for indemnity insurance. She writes, "Before seeking medical care, a subscriber's insurance company would conduct claims 'appraisals' based on diagnostic and treatment cost information generated in-house or supplied by outside diagnosticians."

Under this sort of health insurance, consumers would be paid by insurance companies, just as some auto insurance companies write you a check today if your vehicle is in an accident. As Feigenbaum puts it, "by paying insureds directly, diagnosis is divorced from the ensuing medical care, thereby reducing moral hazard on the part of providers and allowing subscribers to benefit from acting as prudent purchasers. Thus, policyholders and insurers become partners in their efforts to hold down medical costs through comparison shopping for price and quality and conserve health care dollars."

Under an indemnity system, the insurance company would use its guidelines to determine the size of the check to write to the consumer. However, the actual choice of treatment plan, and the price, would be determined by the doctor and the consumer. Thus, this form of insurance is compatible with physician autonomy, and yet the consumer is incented to make choices that balance benefits and costs.

Doctors Beware

Some doctors see universal government health insurance as a perfect solution. It would get them out from under the hated private insurance companies. It would ensure that they get paid. It would keep them from having to raise the delicate issue of cost with patients.

However, doctors should be careful of what they wish for. If government pays for health care, the reduction in overhead, if any, will be short-lived. Sooner or later government is going to have to pay attention to cost. And that means that government will want to enforce guidelines for medical practice. At that point, doctors who value their autonomy may realize that single-payer health care is not their friend.

Modern American medicine is characterized by an abundance of technology, specialization, and possible treatments. In this environment, minimizing overhead may be a misguided objective. Instead, making the right choices is what is important. That may in fact require more "overhead" in the system. More information is needed in order to make cost-effective decisions in health care, and more monitoring is needed in order to align compensation with quality.