Untangling the Economics of Drug Pricing
Examining the Pharmacoeconomics of U.S. AIDS Drug Access, Washington, DC, April 6, 2004
I am honored to be here, but I need to issue a disclaimer: I am a health economist, but my main focus is on reimbursement issues -- that is, pharmacy benefits management. I have also done some work on prescription to over-the-counter switches but I have not done much work on AIDS drugs, at least not in specific detail, nor am I an expert on drug pricing, per se. However, that being said, reimbursement and drug pricing are certainly related, as we will see shortly.
The theme throughout my talk will be that patient access to prescription drugs is a function of three items, the first being "availability" of the drug. That is, it needs to be approved or go through the regulatory approval process, and implemented by the US Food and Drug Administration (FDA), and FDA counterparts abroad. Second, and very important, is "coverage." Christine [Lubinski] touched on that topic. In fact, she gave a detailed talk on coverage, and I will supplement some of what she said. Finally, although this is more self-explanatory than the others, the third item will be "personal income," because some fortunate people may not even need to be covered, may not need insurance. They may simply have income sufficient to pay for their prescription drugs. Most people, of course, are not this fortunate.
I will cover the main FDA regulatory initiatives to speed up and improve the approval process, and then I will turn my attention to coverage issues. I am going to give a brief historical overview, going back about 15 years. Finally, I will give a brief, quick, and dirty case study that I designed last week, examining the placement of AIDS drugs on several US formularies, as well as the Dutch National Formulary. Of course, many of you may wonder why I chose Holland. Well, I have spent half my life in Holland, I have advanced research there, and I have studied there, so it was a natural reference point for me.
So again, access is a function of availability, coverage, and personal income.
Availability depends on regulatory mechanisms and the research and development (R&D) environment. Before I look at availability and coverage separately, let me briefly touch upon a controversial topic, the relationship between the R&D environment and pricing. Suffice it to say, without a functional and profitable R&D environment, drugs would not be developed and made available to the general public in a timely and efficient manner. The R&D environment depends to a degree on pricing. Some say, controversially, that a free market for pharmaceuticals, with no price controls, is essential for long-term growth and innovation; others contend that this is not the case. As we will see, a completely free market for pharmaceuticals does not exist even in this country.
Let me briefly state my take on the pricing debate. We need to recognize the importance of patent protection and a pharmaceutical company's ability to market a drug at a price that matches supply and demand. This being said, I think we must all acknowledge that when making decisions pertaining to drug pricing, we enter the realm of ethics and the need to balance economics with our moral duty to people in need, especially when we are dealing with life-saving treatments, such as AIDS drugs. Needless to say, drug discovery and development take time, money, and they are labor intensive. Basic research can take decades, discovery can be like finding a needle in a haystack, and development requires numerous, lengthy clinical trials. Only one in five investigational drugs actually makes it to the pharmacy shelves.
Availability, of course, depends on many things, not least of which is R&D for new drug development. It also depends on the FDA approval process, and issues related to the marketing of drugs, such as patents. Let me give you a brief historical overview of regulatory initiatives that were intended to improve this process, starting with the Hatch-Waxman Act of 1984. Hatch-Waxman attempted to balance competing objectives -- innovation, price competition and prescription drug affordability. Not an easy balancing act. Hatch-Waxman sought to shorten the time it takes for generics to reach the market by creating the Abbreviated New Drug Application (ANDA) process, which eliminated lengthy and expensive clinical trials for generic products.
Since Hatch-Waxman, there has been continued public concern about the length of time for FDA review, limited access to drugs during clinical trial testing, and the lack of existing therapies for certain disease states, including HIV and AIDS. As a result, there have been more regulatory initiatives, including the expedited approval programs starting in 1987; the Prescription Drug User Fee Act (PDUFA) in 1992, which charged pharmaceutical companies user fees for review of New Drug Applications (NDAs); and the Food and Drug Administration Modernization Act (FDAMA) of 1997, which formally introduced the Fast Track Initiative.
From 1987 through 1992, the FDA developed and implemented several programs designed to expedite patients' access to emerging therapies, either by allowing patients access to unapproved therapies, or by accelerating the drug development and approval process. With Treatment Investigational New Drugs (treatment INDs), desperately ill patients gain access to a drug while the clinical development and FDA review continue. For treatment INDs, drug sponsors may not commercialize an investigational drug by charging a price higher than that necessary to recover costs of manufacture, research, and development. Thirty-nine of these have been granted, among which eleven were for AIDS drugs. Then there are the New Drug Submissions (NDSs) approved under the Center for Drug Evaluation and Research (CDER) Accelerated Approval Programs, Sub-Part H. Thirteen out of 38 of drugs approved through CDER have been AIDS drugs.
The FDA fast-tracked azidothymidine (AZT) in 1987. This first of now 19 AIDS drugs passed through the FDA approval process in four months. Another example is the fast-track approval of saquinavir (SQV), which was approved just three months after the filing of its NDA; and, in 1996, indinavir (IDV) was approved in a little over one month. In all, the FDA's expedited development and approval programs have particularly focused their attention on AIDS drugs, as well as cancer drugs (Slide 1). That is the good part. However, the share of AIDS fast-track designations, for a variety of reasons, have fallen to less than 20 percent in 2001 and less than 10 percent in 2003 (Slide 2). If we summarize the ups and downs of AIDS drug R&D, many drugs have been approved. There is no question about that. And many of these drugs are clearly life-saving drugs, and are very important to the patients and their care providers. There has been improved efficacy, improved compliance, and improved dosing, but it certainly has not met the growing worldwide need. We need to focus our attention not just on the United States, but worldwide, where millions need the medications and do not have access to them.
Now, we will move to the second access factor: coverage. Let's talk about coverage. In this country, and I stress, in this country, a person's degree of drug coverage depends on insurance status, and insurance status is co-determined by socioeconomic status, factors such as income, employer, age, where one lives; and last (and unfortunately unique to the United States) but certainly not least, preexisting conditions. This is certainly a unique American phenomenon.
In Slide 3, for the sake of convenience, I make the ideal assumption that the needed drugs are "available." If all the drugs are out there, the flowchart on this slide shows the relationship between degrees of access and coverage. We can define access in terms of coverage and earned income. The first question is: "Does a person have coverage?" If the person has coverage, we have to look at restrictions in the formulary, possible co-pays, and a high premium burden as ways of unfortunately limiting their degree of access to drugs that are available. Having coverage is not the whole story. This is fairly obvious to most of us, but what I have done here is try to summarize the main points of the relationship between coverage and access. Now if you do not have coverage, it is not necessarily a bad thing if you are really well off. Of course US$50,000 may not sound like a lot to some in the United States, but suppose it is US$100,000 or US$200,000 or at whatever level we establish the cut-off. For those of us who are well off, perhaps coverage is not an important issue, but for most of us coverage is important and insurance status is important.
Then there is the fact that in this country, coverage is a voluntary choice on the part of the patient, the enrollee and his or her employer, for instance. Again, while this may be the case in the United States it is not the same in some other countries. Coverage in the United States is completely voluntary. There is no mandatory mechanism mandating employers to provide coverage. Many do, many do not. The same thing applies to the individual patient. There is no one saying that you have to have coverage, such as the case in Europe, where it is a mandatory mechanism, providing a way of spreading and pooling risks across the population. In the United States we do not have that mechanism, even for those in Medicaid or who are Medicaid-eligible. The earlier talk included the fact that many HIV-positive patients, up to 50 percent, are indeed in Medicaid, but up to 50 percent who are Medicaid-eligible may not even be enrolled in the program at any given time. So many of those people also do not have coverage, despite the fact that Medicaid, compared to private plans, is a fairly generous program across the 50 US states. Of course there are differences, there are nominal co-pays, for example, but many of those who are Medicaid-eligible are not enrolled in the program.
Now contrary to conventional wisdom -- and here is where I deviate slightly from the previous talk -- coverage of pharmaceuticals has actually increased dramatically throughout the 1990s, and ironically, I think that is what has been part of the problem of perception. We see that we have perhaps reached a turning point in 2000, where our own out-of-pocket costs are now going up again. But, certainly throughout the 1990s, as part of the "managed care" revolution, and I know we have been prone to bash managed care, but as part of the managed care revolution, prescription drug coverage actually increased dramatically. Now it has become an expensive proposition. Providing prescription drug coverage became something of a burden to the managed care companies, to government and to Medicaid. Why? Because spending was increasing at rates of up to 19.5 percent back in 1999. It has come a bit more down to earth recently. As shown in Slide 4, the most recent Milliman USA Health Cost Index Survey estimate was that it increased 8.5 percent last year at this time. I do not know if that is a one-off blip, but clearly prescription drug spending has been increasing at a much higher rate than the other components in healthcare (Slide 5): inpatient, outpatient, and physician services.
While our out-of-pocket costs have gone down -- and I am not speaking now to HIV-positive patients, but speaking generally -- prescription drug spending overall has increased. This makes sense, because if you provide coverage, as the Rand experiment 20 or 30 years ago proved, and basically subsidize pharmaceutical care, there will be more utilization, thus more spending. I think what all of us fail to see, though occasionally we do see well thought-out pieces in the New York Times or Wall Street Journal which point out the fact that prescription drugs are still only 9 or 10 percent of the overall healthcare spending pie, is that maybe there is some short-sightedness on the part of certain insurers and Medicaid and the government. We just heard Patrick [Clay] talk about cost-benefit and cost-effectiveness analyses; perhaps prescription drugs actually are a somewhat cost-effective component. It is a rising component, but it is not really a big part of the healthcare spending pie and it is perhaps reducing the growth rate in inpatient and outpatient, and maybe even physician services. There is no proof of this, but we can certainly see that while prescription drug spending was outstripping all the other components by a margin of 3 percent, inpatient costs were growing at 2, 3, 4, and 5 percent per year instead of 15 percent per year. Maybe there is a trade-off in growth rates.
What has been the response to the increase in drug spending? In Western Europe, Japan, Canada, and other countries as well, where healthcare is largely government-run, insurers have responded by cutting drug budgets, raising co-pays, and a few have imposed cost-effectiveness thresholds prior to admitting new drugs to the formulary. Australia is the best example of a country that has introduced these thresholds. It is quite controversial, and they especially apply to breakthrough medications. In the United States, most people who have insurance have private insurance, and private insurers have responded to increased drug costs by caps on drug spending and formulary restrictions. They have also, of course, introduced the multi-tiered co-pay arrangement in a big way because five or 10 years ago, you did not have this three-tiered approach, and now I think from 60 to 70 percent of insurers have a tiered approach to co-pays.
Let me briefly explain other ways to reduce drug spending, which have included price controls. Many countries have imposed direct price controls. The United States generally does not have direct price controls other than the federal ceiling price, which is a Veterans Administration (VA) price. Still other countries have gone down a different path. It is called "reference pricing," based on a web, a nexus of cash information and also drug flows between the government intermediaries, patients, and providers (Slide 6).
Let's start out looking at price controls with the patient who gets his prescription and goes to the local CVS/pharmacy or Rite Aid pharmacy to get a prescription filled. The pharmacy has ordered the drug from a distributor and the distributor got it from a manufacturer. When insurers enter the game, particularly when pharmacy benefit managers (PBMs) enter the game, it gets far more complicated, with payments, rebates, and reimbursements. The pharmacy collects payments from patients (the co-pay) and health plans, and pays distributors for products. Retailers and distributors each take a percentage from what is called average wholesale price (AWP), which is a list price. It is like the price on a car. It is not the actual price paid. It is the sticker price.
On the insurers' end, pharmaceutical companies also provide rebates to PBMs for moving market share and credit distributors. Pharmacy benefit managers negotiate on behalf of insurers; they take their cut of rebates but do pass on a portion to the health plan. Pharmacy benefit managers can thus exert some downward pressure on drug prices; estimates are between 10 and 30 percent. Unfortunately, PBMs are relatively secretive about their rebate deals, and they do not reveal to researchers or others the exact percentages that they are getting off of AWP. We have to estimate these percentages, and when you say 10 to 30 percent, it is clearly a broad range. Again, a common feature of formulary design is the three-tiered co-pay approach through which enrollees pay lower co-pays for generics than for brand names, higher prices for brand names off the formulary than on the formulary. The pharmacy's incentive to participate in this nexus is based on its virtual guarantee of access to an enrollee base of health plans. Again, you are talking about basically 200 million people, or it may be more because almost everyone now is in managed care of some sort. Pharmacy benefit managers have between 200 and 220 million covered lives, and in exchange for this access, the pharmacies agree to a reimbursement formula established by the PBMs, which is expressed as a discount off of AWP. They do get a dispensing fee per prescription.
Clearly there is controversy and confusion surrounding the complex web of cash information to drug flows between the pharmaceutical companies, wholesalers, government, PBMs, and end users, especially concerning drug prices. If the retail price at the pharmacy for a particular dosage and quantity of a brand-name pharmaceutical was US$100, then on average an uninsured individual is paying the highest amount. Christine [Lubinski] in her presentation and José [M. Zuniga] in his introduction, both touched upon the fact that we have something like 40 or 45 million uninsured individuals in the United States; they are paying the highest price, the full retail price for that drug. Health maintenance organizations (HMOs) and insurance companies would pay between US$65 and US$80, Medicaid pays around US$70 (though it could be between US$60 and US$80), and the VA pays between US$50 and US$60. We have a confusing array of drug prices. We have the retail price, the highest in the chain of distribution, and really none of us probably know what the retail price is, because when we go to the pharmacy, at least those of us of who are insured, we are paying a co-pay. We do not really know what the price is and for the most part, physicians and providers do not know either. That is part of the moral hazard.
The fact is, once there is coverage, since no one really knows what the price is, sometimes you can have what is called "physician-induced demand" for certain products. People will just say, "Oh look, you are covered. You can take this or that pill." It does not apply as much to AIDS drugs because these are still life-saving treatments, but it certainly applies to other drugs that we know about from direct-to-consumer advertising. We have then the AWP, which is a list price. It is referred to as a sticker price. Most estimates are that discounts for HMOs and large purchasers are more than 20 percent off AWP. And then there is the average manufacturer's price (AMP), which is the price paid to a pharmaceutical manufacturer by wholesalers for drugs distributed to retail pharmacies. Finally, there is the only direct price control in the United States, which is the federal ceiling price, or maximum price manufacturers can charge for the Federal Supply Schedule.
Now to get back to what I briefly mentioned before, reference pricing. In an effort to contain drug spending, Germany, Australia, New Zealand, and The Netherlands allow reference pricing. Let me take the Dutch as an example. The Dutch formulary reimburses at the level of the cheapest product in a therapeutic "cluster." The government also imposes price ceilings at the same time, and these are based on the average retail price of drugs in four surrounding countries. In Holland, it is important to note that non-clusterable products, which are about 15 percent of the products, and which tend to be life-saving drugs for diseases such as HIV, are put on a separate list and are always reimbursed in full. In 2005, however, all new non-clusterable products will have to pass a cost-effectiveness threshold, a test prior to being reimbursed in full. It will be similar to the situation that currently exists in Australia.
Reference pricing is an approach to reimbursement for pharmaceuticals you will be hearing more about in the coming months in the United States as well. Some health plans are experimenting with a version of reference pricing, and reimbursement based on functional equivalence is closely akin to reference pricing. In reference pricing schemes, products are clustered into groups, as I said, which are based on therapeutic effects. In contrast to generic referencing, therapeutic referencing treats compounds with different active ingredients as equivalent, despite possible differences in efficacy and/or side effects. They cluster on-patent with off-patent compounds. This is often considered quite controversial, even in The Netherlands. The payer -- the insurer -- sets a reference price for each cluster based on a relatively low-priced product in the cluster. Say you are looking at second-generation antihistamines. Suppose that Allegra® was the cheapest. That would be the product that would be given the reference price. The reference price is the maximum reimbursement for all products in the group. Suppose you are a patient in Holland and you want Claritin®. Fine, you can get Claritin®, but you pay the surcharge, whatever the difference is between the reference price and the price for Claritin®.
Now the AIDS drugs and other life-saving treatments are put on the non-clusterable list, what is called 1B in The Netherlands, and I believe it is the same in Australia, Germany, and New Zealand as well. Though I am not sure about this, I am assuming that it is the case, so 15 percent of drugs are considered "life-saving," and therefore they are fully reimbursed.
In Slides 7-10 you can see that Kaiser Permanente and the Blue Cross/Blue Shield plans, for instance, include AIDS drugs on their formularies. The dollar signs simply refer to the tiering of that product. If it has few dollar signs, it is a low-priced product; if it has many dollar signs, it is a high-priced product. We do not know exactly how much people are paying in co-pays, but clearly Kaiser Permanente and the Blue Cross/Blue Shield plans are putting these products on their formularies. Aetna and Harvard Pilgrim coverage may have changed slightly since I created these slides, but I believe that coverage is for the most part the same. Kaiser Permanente has numerous formularies, not just one. We picked a representative formulary. Some of the products are not on certain formularies; other products are on all formularies.
If we compare it to The Netherlands' national formulary, we can see that these drugs have been put on a non-clusterable list and they have been given full coverage, other than the products that are labeled as "N/A," which are not approved by the local FDA counterpart. You have to look at this very carefully. I mentioned availability as being an important component of access. Certain drugs, Baycol® is perhaps the best example, never made it through the approval process in The Netherlands, in fact never made it through the approval process in several European countries. They made it in the United States and approval later had to be withdrawn. That does not mean that The Netherlands or European FDA counterparts are more or less cautious than we are, but it certainly tells us a story about globalization. We are not yet there in terms of harmony of our drug approval processes and I do not think we ever will be, because there is a culture that is different at the FDA from its FDA counterparts such as the European Agency for the Evaluation of Medicinal Products (EMEA). So, certain drugs are just not available. But, the ones that are available are given full coverage.
To recap, availability depends on the regulatory initiatives that are in place to increase availability, but they do require fine-tuning to meet patient demand for newer medications. For instance with AIDS drugs, only one new, novel drug has been approved over the past few years, and certainly something needs to be done about that. Moreover, we have seen that third-party payers are increasingly restricting coverage. Thank you.
Joshua P. Cohen is a Senior Research Fellow at Tufts Center for the Study of Drug Development.
Back to the May 2004 Supplement issue of IAPAC Monthly.