Preclinical Testing

The preclinical testing phase is composed of laboratory and animal studies designed to show biologic activity against the target disease and to evaluate the short-term safety of compounds in animals. Long-term animal studies are initiated to detect possible mutagenicity, carcinogenicity, and teratogenicity. These often continue for several years, concurrent with early human trials. The FDA has laboratory practice regulations that dictate the way in which animal studies must be conducted if they are to be considered for product approval, but these probably add very little to the length and cost of the overall drug development process.

Beyond this is the much more fundamental issue of how much and what kind of animal studies need to be done before an investigator can begin testing in humans and how much animal work needs to be done to receive FDA approval. Some 12 to 15 years ago a citizen advocacy group urged the FDA to require that all animal toxicity studies be completed before any drug be given to any human, including full chronic carcinogenicity studies in two species. Fortunately, the FDA did not adopt this overly stringent strategy.

The issue of animal testing in the preclinical stage remains important. Until 1982, the FDA had no rule requiring a full carcinogenicity bioassay in two species prior to NDA approval. It requires it now. This rule came into effect because Congress said, in no uncertain terms, “either you adopt the rule or we will criticize you publicly.” The FDA adopted the rule. Now, any drug to be used chronically in humans must have full carcinogenicity bioassays in two species to be considered for FDA approval.

I mention this example to make us more mindful of the potential impact of regulations regarding preclinical testing. Grabowski's calculations show that 1 year added to the NDA approval process costs a lot of money. If a manufacturer had to add 1 to 3 years for comprehensive animal toxicity studies before initiating Phase I human testing, the financial impact would be tremendous.

Clinical Investigation: The IND Phase

Let me now turn to the clinical investigation or IND stage. After completing preclinical testing, a manufacturer files an IND with the FDA to receive permission to begin testing in humans. The IND contains all the information known about the compound, including its chemical structure, proposed mechanism of action, stability and manufacturing information, the methods and results of preclinical laboratory and animal studies, and the proposed plans, methods, and investigators for clinical trials. The IND must also be reviewed and approved by the Institutional Review Board where the proposed clinical studies will be conducted.

There is no statutory requirement that there be an IND, and there is no statutory requirement that the FDA be notified about clinical investigations with new compounds. The elaborate protocol that the FDA has constructed is based upon what it, not Congress, regards as good public policy. As a result of the thalidomide incident, the FDA promulgated regulations requiring that a sponsor, commercial or noncommercial, submit an IND plan to the FDA and wait 30 days for the agency to review and approve it.

Technically, if you do not hear from the FDA by the end of 30 days, you can start your clinical study. This almost never occurs. Who would take the risk that the FDA would later disapprove an IND? Informal methods have been developed to deal with this problem. If you do not hear from the FDA, you call them; within the next 30 to 90 days, you will get an answer. Sometimes the FDA will respond with, “We have not reviewed it yet; would you mind waiting another two months?” Sometimes it will call and say, “We have some real questions about your IND,” and you say, “Is this a clinical hold?” They reply, “No, we are not telling you not to start the study, it is just that we think it is a problem.” Sometimes they will call and say, “We not only have questions, we have deep concerns.” The meaning of “deep concerns” can vary widely, depending on the individual reviewer making the comments.

In 10 percent of the cases, there will be a formal clinical hold on an IND. What happens in the other 90 percent of cases? National Cancer Institute (NCI) representatives, before the National Committee to Review Current Procedures for Approval of New Drugs for Cancer and AIDS (the so-called Lasagna Committee), testified that they would not consider going ahead with a clinical study if they did not have absolute FDA approval. The pharmaceutical industry behaves similarly. Thus, months can go by where nothing occurs because of concern about the FDA assessment of the IND.

Once an IND is approved, technically you can tell the FDA you are changing the protocol and proceed at will. There is no requirement to seek its views or approval of the study changes. Not many investigators do that. Researchers at the NCI said they would never go ahead with changes without getting FDA agreement, and many in the pharmaceutical industry do the same thing.

When an investigator goes from a Phase I to a Phase II clinical trial, it may take 6 months to a year to get agreement from the FDA on the protocol. Yet academics and industry researchers alike are unwilling to continue with the next phases of trials without informal approval of the design from FDA officials. If they disagree with the protocol 3 years later, you do not get approval of the NDA.

This clearly puts the FDA in a difficult position. If it does not give the investigator feedback on the study and it turns out to be the wrong type of study, the FDA must turn down the drug and face the criticism of the manufacturer and other advocates. If it does tell the investigator what is wrong with the study, it is accused of “micromanagement,” “fussing over minor details,” or “telling researchers how to run good science.” It is a catch-22 for the agency.

The industry itself is ambivalent about the involvement of the FDA in overseeing research protocols. The industry wants FDA feedback, and the resulting implied approval, as long as the agency agrees with its trial design. It does not want oversight if the feedback is critical, and it does not want FDA micromanaging a project the manufacturer feels it knows more about. NCI says the same thing.

Two proposals for reform have been suggested to help resolve the conflict between feedback and micromanagement. One is to use the advisory committees within the FDA to resolve some of these protocol disputes. The advisory committees would be able to give quick scientific advice and possibly provide an appeals mechanism to adjudicate disputes between companies and the FDA. Reformers suggest that this would remove delays in the process, be more objective, and diminish the amount of micromanagement.

Another, and more radical, proposal is to deregulate at least Phase I of clinical testing. ² The proposal would do for drugs what is done for many medical devices. An investigator would obtain approval of an IND by an Institutional Review Board (IRB) ³ and not go to the FDA until a later point in drug testing. The rationale is that Phase I involves basic pharmacology and toxicology research that poses little risk. Much of it is done in universities. The IRB can do as good a job as the FDA in evaluating the protocols. In addition, since 30 percent of drugs in Phase I testing never go on to Phase II, unburdening the FDA from regulation of this early step would lead to a substantial decrease in its administrative workload.

What are the safety ramifications of such a proposal? Louis Lasagna, chairman of the National Committee to Review Current Procedures for Approval of New Drugs for Cancer and AIDS, says that no human being was ever harmed in a Phase I study. The issue is the extent to which we need a federal regulatory structure to monitor the earliest phase of the drug development process. Could we do without it?

One other factor should be considered in terms of economic impact. The FDA has divergent requirements for investigational drugs and devices. Manufacturers cannot charge the patient for investigational drugs, but they can charge for investigational devices. It seems a clearcut rule. However, in cases involving drug-device combinations there is much confusion. Often, the FDA is unclear whether it will regulate such technologies as drugs or devices. The distinction is not trivial.

There are similar inconsistencies within the class of INDs. The recent treatment IND regulations stipulate that a manufacturer can charge for a product if it has been approved as a treatment IND but cannot charge if it was approved as a compassionate use IND. Nonpayment for certain INDs may be a barrier to entry for small pharmaceutical firms, particularly biotechnology firms, which need a cash flow to continue investigating potentially important new drugs.

The overall time required to do the clinical studies necessary to complete an NDA is highly variable, somewhere between 1 and 8 years. The large range is partly explained by the discretionary nature of the clinical development stage. For example, there are no uniform requirements on the number of Phase III trials, which are the largest, most expensive, and most time consuming. ⁴ If the FDA requires two or three Phase III studies, it is going to take a long time for your NDA. If only one Phase II study is required, as has been true in a few instances of expedited review, ⁵ it will take as little as 1 year.

Criteria for Phase III requirements are not mandated in the statute. The law is broad and nonspecific. It says there must be substantial evidence, supported by adequate and well-controlled clinical trials, sufficient to convince qualified experts that a drug is safe and effective. The statute also says safety must be proved by all tests reasonably applicable to show safety. This extraordinarily broad statutory requirement is nothing more than a slogan. It tells the FDA to approve good drugs and not to approve bad drugs.

FDA Approval: The NDA Phase

When considering the FDA official assessment of new drugs, the first question is how to set priorities for regulatory review. Do you evaluate drugs in the chronological order in which you receive them—first in, first out—or do you set up some kind of priority system? The choice has serious economic implications. The FDA has established a priority system based upon its assessment of a drug's chemical novelty and potential treatment benefit over existing therapies. ⁶ Internal FDA review of pending NDAs results in an inventory that lists drugs from most important to least important.

The FDA priority system does not coincide with the economic needs of companies or the economic needs of the innovation process. The system currently give highest ranks to new chemical entities (NCEs) that are unusual and that will treat diseases that are otherwise untreatable. It might, for example, rate as the most important drugs compounds for treatment of orphan diseases, which might have as few as 100 or 200 patients in each category. Because the FDA criteria place emphasis on breakthrough drugs, second-and third-generation drugs for cancer and cardiovascular disease may be lower on the list.

If a manufacturer has two drugs, an orphan drug of no economic consequence and another of major economic importance, the orphan drug may well be reviewed first. The second could be held up 3 or 4 years while all the other orphan drugs go through. Do we as a society want that kind of priority system? This is a key issue that we must address explicitly. At present, the FDA has total discretion. It could easily reverse the system if it wanted to.

Another question is how much data are needed to show safety and effectiveness. How many adequate and well-controlled clinical trials should there be, with how many patients, under what protocols, and involving how many subpopulations (such as the elderly, children, or any other group)? The statute leaves these issues to FDA discretion. De facto criteria vary from drug category to drug category and among individual reviewers. Such variations can result in differences in regulatory approval time of 2 or 3 years for a drug.

Finally, there is the impact of social and political pressure on the process. The degree of public pressure on the National Institutes of Health (NIH) to develop new drugs to combat acquired immune deficiency syndrone (AIDS) and on the FDA to expeditiously approve such innovations is unprecedented. No group of heart patients or cancer patients ever marched on the FDA or even succeeded in getting a congressional hearing to object to the way that FDA was doing its job. It is reasonable that the FDA concluded from the absence of public protest that it was doing its job correctly. Perhaps it has been. If that is true, we should not pay any attention to what the academic community, the economic community, the industry, and patient advocacy groups are complaining about.

It is important to understand that the regulatory process is most crucially driven by individuals and not by policies. Thus, it is the decision-making style of the people who review and approve NDAs that drives what happens. Every NDA for AIDS drugs has gone through in record time. Every biotechnology drug, whether or not it is a breakthrough, has gone through in record time. The reason is that some people in the agency wanted to prove that the FDA and the United States were going to be the leaders in biotechnology. They sought to show this by making biologic agents a high priority. This has not been true for cancer drugs, which have been slower to receive NDA approval. The same has been true for cardiovascular drugs. Surprisingly, there has been no vocal public constituency pushing for the development of cancer and cardiovascular drugs.

Overall, the FDA approval system is like a balloon. If you squeeze it in one place, it bulges out in another. A choice to approve one drug expeditiously means that another will wait longer. Grabowski has shown that the average time to NDA approval has not changed dramatically in 20 years. The example of AZT, which got approved in 6 months, only meant that some other drug that would have gotten through in 2 years is going to get through in 3.5 years.

Surprisingly, the work of Cook and colleagues has shown that it is not these global issues of safety and effectiveness that hold up most NDAs. A study done a decade ago showed 60 percent of issues delaying approval were related to manufacturing, chemistry, and quality control data, and only 20 percent were related to safety and effectiveness issues—the big issues that we all tend to associate with delays. Furthermore, almost all drugs that enter Phase III trials get approved. The only exceptions occur when manufacturers give up on a drug after years of testing because the market has changed or the compound is not worth marketing for some other reason. In fact, a new phenomenon has been emerging recently: drug companies receive FDA approval but decide not to market the compound because it has become obsolete while awaiting official clearance.

How could we change this system? There are two major strategies. We could approve all NDAs at the end of Phase II and eliminate Phase III entirely. We could get rare adverse reaction information in another way. This proposal would have a truly major impact. Richard Crout used to tell me, when he was Director of the FDA Bureau of Drugs, that he knew at the end of Phase II whether 90 percent of drugs were safe and effective. Approving drugs at the end of Phase II would cut out an average of 3 or 4 years from the approval process.

One other way of reforming the system is to shift the burden of balancing risk and benefit to the doctor-patient relationship. Advocates envision the pharmaceutical companies giving information about new drugs to physicians and having them openly discuss the risks and benefits of the new therapies with patients, rather than the current FDA approval process. I think this approach goes too far. Some form of FDA approval is necessary. It is clear from the history of Western drug development that the lack of regulatory standards means there will be more failures, more people hurt, and more drug tragedies, as well as more drug triumphs.

Other things that have been suggested I think are also unrealistic. For example, it has been argued for 15 years that if we had more Phase IV studies—post-marketing studies to monitor side effects—we could approve drugs earlier. Everyone I have talked to who has worked in the process believes that we would have the same drug approval time with additional Phase IV studies on top of that. It might be a more complete system, but it would not shorten approval times.

People have suggested making better use of FDA advisory committees and critics have called for institutional changes whereby FDA reviewers would have a chance to do research in a better working environment. These suggestions are good ones. My analysis, however, is that these improvements probably would not decrease the time to approval.

Post-Marketing Requirements

Let me now consider the fourth phase of the drug approval process: post-marketing requirements. This phase is composed of Phase IV testing ⁷ and post-marketing surveillance. Many fewer formal post-marketing studies are done than might be expected. Most information on adverse drug reactions comes from the current system of informal post-marketing surveillance and adverse reaction reporting, and the FDA believes that it works quite well. As a result, expensive formal studies are not needed.

One area worth considering in terms of economic impact on the pharmaceutical innovation process is the use of supplemental NDAs to gain approval for new uses of approved drugs, changes in the drug label, changes in manufacturing procedures, or other changes in the original NDA. It is important to understand that the original NDA is a contract. Once an NDA is approved, no one can vary one iota from it.

It is a system that makes enforcement easy. From an innovation perspective, however, such a rigid system makes pharmaceutical manufacturing and development very difficult. Supplemental NDA are the lowest priorities for the FDA to review. Thousands of supplemental NDAs go to the bottom of the bottom of the reviewers' pile. They can sit there for months and years. This inefficiency in the system is significant because approving the new clinical indication often would allow the drug to be reimbursed for a previously unapproved use. ⁸ Here is something that is ripe for reform.

Generic Drug Competition

The fifth and final phase of the process involves generic drugs and generic competition. The major impact of post-marketing approval of generic competition began in 1984 with the passage of the Drug Price Competition and Patent Term Restoration act. I think it is fair to say that this has had a bigger impact than any of us who worked on it at the time anticipated. Basically, the act is a trade-off of patent term extension for generic competition as soon as the patent runs out. Because of the way in which generic competition has shortened the effective commercial life of a drug there is a smaller window of time within which all the profit has to be made to recoup the investment on that drug and to set up a reserve for research for future drugs.

That window used to be much longer—10, 20, even 30 years. We are, therefore, seeing increased prices. Congress passed a statute that gave the pharmaceutical industry two alternatives: get rid of research or raise prices to finance research. Generic competition has caused drug companies to raise prices while they still have marketing exclusivity, and these higher prices are what finance drug research.

On the other hand, the American public and Congress are concerned about these price increases. The economic return on investment in research and development is a major public policy issue that needs to be addressed.

The Overall Impact of Regulation on Innovation

It is clear that the total impact of the regulatory system on pharmaceutical innovation is very large. There are barriers and impediments at almost every stage in the process, many of which could be lower. The conglomeration of barriers has, among other things, resulted in an overall barrier to entry. How many new small pharmaceutical companies are there? There are a few in the biotechnology field, but how many are surviving the consolidation process, and how many are being swallowed up by mergers or outright acquisition? How many of those will survive another decade because of the enormous cost that we as a society have placed upon the drug development process?

The Big Players: Public and Private Payers

Two basic categories of major players affect the reimbursement of new medicines. One is the public sector—federal and state governments with Medicare and Medicaid—and the other is the private insurance industry. There are even more rules for reimbursement than for getting new drugs approved. They vary widely from company to company and from city to city in the same company. There is no nationwide reimbursement rule for the Health Care Financing Administration (HCFA), Medicaid, the private insurance industry, or even individual insurers.

Inconsistencies in Payment Policies

There is no consistency, no predictability, and very little fairness in the system. One explanation of these inconsistencies seems understandable. Insurance programs are based on insurance principles. They are not public health programs. The job of the insurance companies is to conserve funds, not to conserve or improve public health. Payers usually disagree vigorously with such a characterization, so let me give three scenarios: use of an unapproved drug in a clinical investigation, use of an unapproved drug for treatment outside a clinical investigation (i.e., compassionate use of an investigational drug), and use of an approved drug for an unapproved purpose. Let us see how the insurance system handles these cases.

Assessing the Overall Impact of Payment Policies

Having said all of this, though, I believe the impact of payment policies on pharmaceutical innovation is relatively small. There is a potential impact in two places. One is on clinical trials. Lack of reimbursement for all the attendant hospital costs and other services certainly serves to discourage some clinical trials. I am told, however, that as a practical matter a practitioner usually can find a way to say that the patient would have been hospitalized anyway and thus guarantee payment. The second impact is in the area where there is, say, no diagnosis-related group that covers a very expensive new drug that will not be paid for. The lack of proper reimbursement may reduce the frequency of the use of drug and limit the companies' ability recoup their research investment.

In conclusion, I do not think that reimbursement policies retard the innovation process the way the FDA approval system does. However, the reimbursement system is likely to be reformed, so its impact may be changing.