Outside the operating room—economic, regulatory, and legal challenges
An economic value perspective: Setting limits on health care can be ethical
By Peter A. Ubel, MD
I am a fan of innovation: my patients benefit from it every day. But I am also concerned about the cost of health care. In the Veterans Affairs health system, I see patients who cannot afford their medications and who cannot afford to get private insurance; such problems are largely due to the high cost of health care.
As an example, consider a new pharmaceutical innovation, bevacizumab (Avastin), which costs approximately $106,000 per year when used to treat lung cancer.12 On average, the treatment leads to a 2-month increase in survival, making the cost of this intervention more than $600,000 per quality-adjusted life-year. Or consider the use of a left ventricular assist device rather than medical management for patients with congestive heart failure who are not eligible for transplantation. The estimated cost is approximately $900,000 per quality-adjusted life-year.
These examples illustrate that some benefits to patients can come at a very high cost. For this reason, I believe that we need to set limits in (ie, ration) health care. I will outline here why we need to do so and why third-party payors—both government and private insurance companies—need to consider the cost-effectiveness of health care interventions in deciding whether to pay for them. In the process, I will discuss common thresholds for defining the price of life and explore whether special moral considerations are required for life-saving treatments—ie, whether the price of life should be higher for severely ill patients.
WHY IS IT TIME TO RATION MEDICAL CARE?
Spending on health care in the United States has risen steadily in the last few decades both in real dollars and as a percentage of the gross domestic product. One important reason for setting limits on health care spending is that we have other things to spend our money on. Medicare budgets compete with tax cuts, education, military spending, homeland security, and many other national interests. Economics teaches us that we have to make difficult choices: when we spend more on health care, we have less money to spend on other things.
Cost-effectiveness analysis provides insight on why it is important to set limits. When I trained at the Mayo Clinic, we used to send patients home with six fecal occult blood test cards to screen for colon cancer. (Patients smear stool on a card and mail it to the laboratory, where it is tested for blood; if blood is present, the patient needs a colonoscopy. The six card samples are taken and mailed at periodic intervals to maximize sensitivity.) What is the cost-effectiveness of the sixth card? The answer is surprising: although the cards cost only a couple of dollars, the cost per life saved is an estimated $26 million, which most would agree is more than we can afford to spend to save a life from cancer.
Why is the sixth card so expensive? If any of the first five cards shows blood, the sixth card is worthless, as it provides no new information. On the other hand, if none of the first five cards shows blood, the chance is minuscule that the sixth card will show blood that actually comes from a precancerous lesion that can be removed and save a person’s life.
This example illustrates that cost-effectiveness does not apply only to expensive new therapies like Avastin; it also applies to really inexpensive items like fecal occult blood test cards.
WHAT IS A YEAR OF LIFE WORTH?
If our own child were sick, we would say that a year’s life is worth an infinite amount of money; we would do anything we could to save our child’s life. But the job of the cost-effectiveness community is to address this question from a societal perspective, and they have a different answer. The most commonly cited view among experts in cost-effectiveness analysis is about $50,000 per quality-adjusted life-year, although it typically ranges up to $100,000.13
This figure has not risen with inflation, and it probably should not. If enough new technologies were developed at the threshold of $50,000 per quality-adjusted life-year, the entire budget of the country would quickly be used up.14 Making payment decisions based on a certain cost-effectiveness threshold sets no real limit on health care spending. The threshold is not meant to be a realistic number but should illustrate the kind of thinking required about how much we want to spend on health care relative to other things. The aim is to help us decide how much “bang for the buck” we should expect from our dollars spent on health care.
WHAT DO PEOPLE VALUE WHEN SETTING LIMITS?
In light of the above, how do we set limits when trying to decide what the price of life is? Might our limit-setting be changed if we are facing a desperately ill patient? Examination of questions like these reveals that people value other factors beyond just economic efficiency, as can be illustrated with a couple of theoretical policy dilemmas.
Dilemma 1: Cost-effectiveness vs fairness
Imagine that the Medicaid program decides to screen for colon cancer. They have enough money either to offer an inexpensive test (“Test 1”) to everyone and save 1,000 lives or to offer a more expensive test (“Test 2”) to half the population (selected randomly) and save 1,100 lives.
If the decision were made according to rational cost-effectiveness principles, the choice would be to go with Test 2 in half the population, as it saves 10% more lives and thus maximizes the average health of the population. However, a survey found that the option of offering Test 1 to everyone was favored by 55% of the general US public, as well as by 55% of medical ethicists and even by 45% of cost-effectiveness experts, all of whom were willing to give up some cost-effectiveness for fairness.15
This tendency to favor fairness suggests that moral considerations affect health policy decisions in important ways. Yet further analysis raises questions about the extent to which these considerations are based truly on moral values as opposed to psychological quirks.
For instance, my colleagues and I presented this same choice of colon cancer testing scenarios to a separate survey sample, and again a highly similar rate of respondents—56%—favored offering Test 1 to the full population as opposed to offering Test 2 to half the population. However, to test whether this preference for equity over efficiency persists when neither test can be offered to the entire population, we changed the scenarios for a separate group of randomly selected participants. In one version of the scenario, we told participants that only 90% of the population could receive Test 1 and only 40% could receive Test 2. (As in the original scenario, we indicated that Test 1 saves 1,000 lives, whereas Test 2 saves 1,100 lives.) With just this small variation in test availability, the proportion of respondents favoring Test 1 plummeted to 27%. Similarly, we randomly selected another group of participants to receive a third version of the scenario, in which 50% of the population could receive Test 1 and 25% could receive Test 2, saving 1,000 and 1,100 lives, respectively. Once again, the proportion of the respondents favoring Test 1 remained low (28%).16
These results suggest that people’s preference for equity versus efficiency depends, in large part, on whether the more equitable option can be offered to everyone in a population. But people’s preferences are actually not nearly that coherent. Consider a follow-up study in which we repeated the scenario again for each respondent, but with a twist.
In one group, we began with our original scenario: 100% of the population can receive Test 1, saving 1,000 lives, or 50% can receive Test 2, saving 1,100 lives. As expected, 60% of participants chose Test
1. But then we told this same group of participants that the number of people qualifying for Medicaid had doubled, so that the tests could be offered to only 50% and 25% of the population, respectively (still saving 1,000 and 1,100 lives, of course, since the population was now twice as large). Remember that when people were initially presented with this 50% versus 25% option (without any other scenario being presented first), the preference for Test 1 plummeted. In this case, however, almost no one changed their mind: the majority (60%) still favored Test 1.17
People’s preferences for how to allocate scarce health care resources—the moral values that they believe should guide our health system choices—are often disturbingly arbitrary.18
Dilemma 2: Targeting severe vs moderate illness
Now imagine a new scenario. A treatment is available that will help patients with an illness that causes severe health problems, but it provides only modest benefit. Another treatment helps patients with an illness that causes moderate health problems, and it provides considerable benefit. The cost of the two treatments is the same. How should funding be allocated?
Although a majority (60%) of survey respondents say that most funding should go toward treating the moderate illness where considerable benefit is expected, a sizeable share of people (40%) favor devoting most funding to the severe illness despite the more modest benefit.19 This is another instance where moral values seem to come into play, as a large minority will favor helping the severely ill even at the expense of efficiency.
A variation of this dilemma illustrates another salient point—that people like “easy outs.” When we present people with an additional option—“How about spending money equally between the two treatments?”—the vast majority (75%) choose that “compromise” option over the option of devoting most funds to either of the individual illnesses.19 The lesson is that we hate making difficult decisions, both as individuals and as a society.
COST-EFFECTIVENESS IS THE MOST RATIONAL AND ETHICAL WAY TO SET LIMITS
These surveys make clear that many of the moral values that people express are fragile at best or even psychological quirks. I have heard no compelling moral arguments to support treatments that cost more than $500,000 per quality-adjusted life-year, which leads me to conclude that many new medical interventions are unaffordable. The resistance to limiting such treatments is psychological and political, but it is not ethical.
The appropriate response is for third-party payors, such as Medicare and insurance companies, to let industry know that cost-effectiveness matters. If a treatment is not cost-effective, it should be limited to people who pay out of pocket or for experimental purposes. To make this happen, we need cost-effectiveness analyses of new technologies. Because such studies are expensive and time-consuming, we should develop new incentives to motivate companies to conduct such studies of their products, perhaps by extending patent protection for products that are shown to be cost-effective. We need to work with industry on how to implement such a plan. But continuing to ignore the cost-effectiveness of interventions when they come to market is harming patients who can no longer afford insurance, which has real consequences on people’s health and well-being.
