t’s hard not to like the notion of value-based pricing for prescription drugs. It aims to ensure that the prices we pay for drugs reflect the benefits they provide, either in terms of longer life or better quality of life. And what could be more American than letting the market determine the right price through the intersection of demand and supply?
So it is a bit ironic that, when it comes to determining the value of drugs, European health systems have led the charge on value-based pricing. Regulators there have developed sophisticated procedures — formally called health technology assessments — to assess the value of new drugs, devices, and diagnostic tools. These assessments are then routinely used to set prices.
The practice has only recently taken hold in America. The nonprofit Institute for Clinical and Economic Review now routinely conducts health technology assessments on approved drugs, devices, and diagnostics in the later stages of the approval process, often unveiling them with great public fanfare. Medical societies such as the American Society for Clinical Oncology, health care providers such as Memorial Sloan Kettering, insurers, and others are all working on their own frameworks to measure value.
Measuring a therapy’s value turns out to be incredibly difficult. Modest differences in the assumptions that go into these models — such as the size of the treated population, the duration of treatment, and the therapy’s effectiveness, to name just a few — can generate wildly different estimates of how cost-effective a therapy is. For instance, one analysis of omalizumab, a drug used to treat asthma, found a nearly 50 percent difference in the drug’s cost-effectiveness because of differing assumptions.
Using incorrect or misleading assumptions isn’t just an academic error. If ignored, they can distort the cost effectiveness of a therapy and lead to unsound coverage decisions and wasted resources. Here are five issues that complicate the application of health technology assessments:
Choosing the source of data. The effect of a treatment in the real world may differ from its effect in a clinical trial. That’s because real-world patients tend to be sicker than trial participants, who are usually younger and healthier. In addition, clinical trial participants are better at taking their medications and are monitored more closely than those in the real world. While several organizations have advocated using real-world data in health technology assessments, there is no consensus on which should be used. But the choice matters. In one analysis, asthma studies that used real-world data were twice as likely to judge a treatment as cost effective as those using clinical trial data.
Incorrect use of list prices. Most health technology assessments use a drug’s “list price.” This is the publicly listed price. Much like the sticker price of a new car, the list price is typically higher than the actual price customers pay, generally because it doesn’t reflect discounts that insurers negotiate with drug companies. For example, the list price of PCSK9 inhibitors, a new class of injectable cholesterol medications, can be as high as $14,000 a year, yet no insurer actually pays that price. In assessments that compare brand-name drugs to generics, assumptions about the list price can favor generic drugs by making them appear more cost effective than they actually are.
Falling drug prices. The price of a drug usually falls as brand competitors enter the market, generic versions become available, or a combination of the two. That means the cost-effectiveness of a brand-name drug should improve over time (because its price relative to it comparators falls) even if its efficacy does not improve over time. A health technology assessment may or may not capture this trend because there is no consensus on whether to assume a single, constant price over time or a price that falls due to competition from other drugs. To take an extreme example, a brand-name drug judged not to be cost-effective at a price of more than $100,000 per quality-adjusted-life year may become highly cost-effective when the patent expires and one or more generic equivalents enter the market.
Changing evidence base. New studies conducted in larger populations or in specific patient subgroups continually update the evidence on new drugs and devices. Most health technology assessments, however, are slow to incorporate this new evidence in their assumptions. That can be a problem when insurance coverage decisions are based on outdated evidence.
High-cost drugs can still be valuable. Health technology assessments often calculate the impact of a drug on a health system’s budget. Drugs that have a large effect on a budget — either because they are expensive or treat large numbers of patients or both — raise concerns about affordability. But a drug’s budgetary impact is sometimes misconstrued with its value. Drugs that would have a large effect on a budget, for example, are sometimes deemed to be less intrinsically valuable. In reality, though, drugs that are highly effective and treat a large number of people are arguably what society values most — even though these drugs can also cost a lot. For such drugs, like those recently approved to cure hepatitis C, which can cost upward of $50,000 for a course of treatment, it would be tempting for regulators to ask drug makers to set lower prices, which would make these drugs appear more valuable. But this solution sidesteps the real policy challenge of how to ensure that payers can bear the brunt of high short-term costs for drugs that are socially valuable.
Deciding how best to handle these issues is indispensable to determining the reliability — and value — of health technology assessments.
Health technology assessments conducted to measure the value of drugs should recognize that health is a valuable investment, much like owning a home. Both may entail high up-front purchase costs. We’ve solved the high cost of buying a home with home mortgages, not by policies that lower home prices. We should do something similar for drugs, particularly those that cure diseases, which are likely to have up-front costs in the hundreds or thousands, if not millions, of dollars. If a drug creates value for society, we should figure out ways to finance its cost. Annuitized payments by insurers to drug manufacturers, which could be called “drug mortgages,” are one approach. Outcomes-based agreements, which ensure that drug companies are paid for actual — not potential — benefits to patients are another way to better align pricing and coverage of drugs with their value.
In addition, more needs to be done to ensure transparency in health technology assessments. Individuals and organizations conducting them should publicly share their models so the underlying assumptions can be assessed, criticized, and modified. An open-source platform would make it possible for all stakeholders to assess the reliability of models, analyze their sensitivity to assumptions, and promote vigorous debate on the appropriate way to conduct these analyses.
The unfortunate irony is that while data from industry-supported drug trials is now routinely available to the public, the models used to value these drugs remain in the shadows. And that will keep us all in the dark about how to align the prices of drugs with their true value.
Dana Goldman, PhD, is professor of pharmacy, public policy, and economics and director of the Leonard D. Schaeffer Center for Health Policy & Economics at the University of Southern California. He is also a cofounder of Precision Health Economics, a health care consultancy, and owns equity in its parent company. Anupam B. Jena, MD, is an associate professor of health care policy at Harvard Medical School, an internist at Massachusetts General Hospital, and a scientific advisor at Precision Health Economics. He has received consulting fees from Pfizer, Novartis, Bristol-Myers Squibb, Vertex Pharmaceuticals, and Hill Rom Inc.