« iepriekšējāTurpināt »
Reprinted with permission from
The Decline in Effective Patent Life of New Drugs
Martin M. Eisman and William M. Wardell
The effective patent life for new chemical entity drugs has fallen sharply in recent years as a result of an increase in the clinical testing periodh later starting of clinical testing after the patent application, and quicker issue of patents.
In a recent statement of concern about the state of domestic industrial innovation, the President recommended strengthening the patent system (1). That statement implied that the historical role of patent protection as a major stimulus for innovation had weakened. To determine the extent to which the problem affects pharmaceuticals, this paper examines the state of patent protection afforded new drugs.
The Patent Act of 1836 was adopted because of a perceived need to encourage innovation by eliminating the reluctance to disclose an invention. As incentive for disclosure, the Patent Act granted the inventor a 17-year exclusive right to his inven. tion. As the innovative process became uncertain, lengthy, and expensive, patent protection acquired even greater importance.
In the research-based prescription pharmaceutical industry, patents play an important role. Approximately one out of 10,000 compounds initially examined survives the intense scrutiny and demonstrates the potential to justify marketing. (The Pharmaceutical Manufacturers' Association surveyed its member companies in 1962, 1967, and 1970 asking for "an estimate of the number of chemicals, compounds, mixtures, filtrates, or other substances obtained, prepared, extracted or isolated for a medical research purpose, and subjected to biological tests or screens." This included material obtained from outside the company. The estimates were 144,559 for 1962, 175,760 for 1967 and 126,060 for 1970, averaging 148,793 items tested per year. (Our studies showed that an average of 15.3 New Chemical Entities (NCEs) were introduced annually from 1962 to 1978. Using these averages, the ratio
of chemicals tested per year to NCEs introduced per year is 9725:1.)
Bringing that single drug to market has been estimated to cost $54 million in 1976 dollars (2). Because of this uncertainty and high cost, patent protection is a necessary incentive for the infusion of capital to stimulate research and development. Since drugs are technically easy to copy, the patent provides the primary protection against imitation and competition.
Another form of protection against competition - one probably not intended by Congress is afforded by the regulatory system of the Food and Drug Administration. The expense involved in see. ing a new drug through the demanding system of regulatory review to demonstrate safety and efficacy creates a substantial barrier to entry into the industry.
However, while certain aspects of the regulatory process may offer some protection against competition, other aspects reduce the duration of patent protection that is of commercial value to the original patent holder. Most drug patents are filed when biological activity is first observed (3,4). Since this occurs long before the drug receives regulatory approval for marketing, the "effective" patent life will be reduced considerably from its nominal period of 17 years. We will now examine the extent of this reduction, and its change with time. Time Trend in Effective Patent Life (EPL)
Effective Patent Life (EPL) is defined as the period of patent protection remaining for a drug at the time of U.S. NDA approval (i.e., the time from NDA approval to expiration of the patent). Recent studies (3,5,6) show that EPL has declined substantially over the past 15 to 20 years. This trend is generally attributed to the concomitant increase in the time required for human investigation and NDA approval (3,5). To examine this hypothesis, we need to analyze the time trends in both EPL and the period from the start of clinical investigation to U.S. NDA approval.
Dr. Eisman is an Associate in the Department of Pharmacology and Toxicology. University of Rochester School of Medicine and Dentistry. Dr. Wardell is an Associate Professor of Pharmacology. Toxicology and of Medicine, and Director of the Center for the Study of Drug Development, at the University of Rochester School of Medicine and Dentistry. He is also Chair. man of the Committee on Government Allairs of the American Society for Clinical Pharmacology and Therapeutics.
Methods – The analysis is based on all patented new chemical entities (NCES) receiving NDA approval from 1966 through 1979 (a). The information needed to determine EPL included dates of the start of clinical testing in the U.S., NDA approval, and patent application and issue (b).
Data were available for nearly all variables from 1966 through 1979 (c).
Sources for the patent data included the patent consultant Louis Leaman. SmithKline Corporation, direct surveys of individual pharmaceutical companies, and varoius reference sources, including Chemical Abstracts and Official Gazette of the U.S. Patent Office. For multi-source drugs (i.e., the same drug marketed under different brand names by dif. ferent companies) only the drug of the original patent holder was included in the averages. Of all 191 NCEs approved from 1966 through 1979, 168 had patents. The data from those 168 drugs were used to calculate EPL.
of the three types of drug patents (new compound, medical use, and chemical process), a patent on the new compound provides the most reliable protection. To calculate EPL, we used the earliest compound patent listed for a drug Ino compound patenl existed weised the earliest patent. erardless of type
Data are grouped according to year of NDA approval. For each variable (e.g., time from start of clinical testing to NDA approval), the time difference was calculated for each drug, and those differences averaged for all drugs approved during that year. The averages were plotted and the raw plots smoothed (Figures 1 and 3) according to the "moving median of three" technique of Tukey (7)
Drugs tested before 1963: Length of clinical investigation phase – The IND filing dates assigned retrospectively to drugs in clinical trial before August 1962 do not represent the start of clinical testing in the U.S. (d).
Thus, the true period of clinical investigation for pre-1963 drugs began earlier
than the date represented by retrospective IND filings. Of the 168 patented NCEs approved from 1966 through 1979, 43 had been assigned retrospective IND filing dates. We were able to obtain the date of first U.S. clinical testing in man in the U.S. for 21 of the 43 retrospective filing dates. From this information, we have derived a standard value of 24 months to apply as a correction to the remaining 22 drugs for which this information was unobtainable (e).
Effective Patent Life - Figure 1 displays the relationship between the patent and drug develop ment processes, showing the times of NDA ap proval and the start of clinical testing in relation to the time of patent issue. The data are plotted according to year of NDA approval. EPL, the time from NDA approval to patent expiration, can be read directly from the right-hand ordinate. As shown in the Figure, EPL for pharmaceuticals was
Figure IINDA approval (averaged 0: smoothed and start of clinical testing (averaged O; smoothed corrected for retrospective IND filings, are plotted in relotion to patent issue. Smoothing was done by Tukey's
"moving median of three" technique (7).
✓ considerably less than 17 years, even at the beginn
ing of the 14-year study period. It declined from 13.6 years in 1966 to 9.5 years in 1979, a decrease of 4.1 years.
Time from start of U.S. clinical investigation to NDA approval – Figure 1 also shows the pattern (after smoothing (7)) of the period from the start of clinical testing to NDA approval during the 14 years from 1966 to 1979. During the 12-year period from 1968 to 1979, EPL dropped by 4.0 years, from 13.5 years to 9.5 years . The time from the start of U.S. clinical testing to NDA approval increased by 24 years li... from 5.9 to 83 years) from 1968 to 1979. accounting for 60% of the decrease in EPL Ligh
Thus the increase in the period from the start of clinical testing to NDA approval accounted for only slightly more than half of the decline in EPL. Therefore, we need to examine the components of EPL in more detail to determine where the re mainder of its decline occurred. Effective Patent Life and the Drug Development Process
From our data (presented later in this paper) we know that the sequence of events in the process of drug development is generally as shown in Figure 2. The sequence begins with the filing of a patent application during the preclinical phase, and continues
Figure 21 Effective Patent Life (EPL) is a function of the timing of the patent application, the pendency period, and the duration of the clinical and regulatory period, as well as the 17-year period of patent protection. The pendency period is the time from patent application to patent issue. with the start of clinical testing, patent issue, NDA approval, and finally patent expiration.
From this pattern and Figure 2, we see that EPL (i.e., the period from NDA approval to patent ex. piration) is a function of the timing of the patent ap plication, the pendency period, and the duration of the clinical and regulatory periods, as well as the 17-year period of patent protection.
Thus, in addition to its dependence on the duration of the clinical and regulatory periods, EPL depends on two other important factors. It decreases if clinical testing is begun later in relation to the patent application, and conversely will increase if the patent pendency period increases. The final EPL depends on the algebraic sum of the · changes in the components.
The changes that occurred in the two additional components of EPL are shown in Figure 3. For the years 1968 and 1979, the two years most representative of the general trend during the study period, the time from patent application to the start of U.S. clinical testing increased 0.5 years (accounting for 13% of the decrease in EPL). The time from earliest batent application to patent isque decreased 1.1 years (accounting for 27% of the decrease in EPL) (h). Coupled with the 2.4 year increase in the period from the start of clinical testing to NDA approval, these changes account for the entire 4.0 year decrease in EPL from 1968-1979. (i) Discussion/Conclusions
EPL was 13.6 years at the beginning of our study period, 1966. This is considerably less than the 17.year nominal period of patent protection. As time progressed, EPL fell further. This trend is similar to that reported by other investigators (3,5,6). The decrease over time has generally been attributed entirely to an increase in the time between the beginning of clinical testing and NDA approval (3,5), although Statman suggests that this may be responsible for only part of the decrease (6).
Dur Analysis shows that in the specific sample of NCEs analyzed, almost hall of the decline in EPL was caused by two additional factors: An increase in the time between patent filing and clinical testing,
Figure 3/Averaged and smoothed values for NDA ap proval start of clinical testing, and patent application are plotted in relation to patent issue. The symbols and smoothing are defined as in Figure 1, with the addition of earliest patent filing laveraged 0 ; smoothed -:- and start of clinical testing, uncorrected for retrospective IND
filings f... and a reduction in the pendency period. It should be noted, as seen in the Figures, that the relative contribution of each of the three components depends to some extent on the years compared.
For the 12-year period from 1968 to 1979, the declining EPL can be explained by two trends. The clinicalregulatory period increased (with all of the increase being in the clinical period), and more of the clinical/regulatory period fell within the period of patent protection (i.e., after the date of patent issue). This latter trend was caused by quicker issue of the patent by the Patent Office (thereby starting the patent clock sooner in the drug development process), and by later starting of the clinical testing.
It should be clearly understood that the start of clinical testing" being described in this analysis is clinical testing in the U.S. only. Although approx. imately half of the drugs approved in the U.S. originate abroad (10), and a significant fraction of U.S. originated NCEs are now also first tested clinically abroad (8,9), this study is limited to the U.S. component of the drug development process.
Although a decrease in the pendency period results in earlier issue of patents, it contributes to the erosion of EPL by placing a greater proportion of the clinical regulatory process within the period of patent protection.
It is not clear why U.S. clinical testing is starting
The value of 24 months was obtained by calculating the mean of the available values after eliminating two outlier drugs. The general trends over the study period are better represented by comparing 1979 with 1968 rather than with 1966. This is shown more clearly in Figure 3. This period is made up of two components, the INI phase and the NDA phase, which we have examined in detail in other publications (8,9). For the specific set of drugs used in this paper, the mean value of the period from NDA submission to approval was 2.4 years from 1966 to 1972, and 2.2 years from 1973 to 1979. The period of clinical testing increased from a mean of 3.3 years in 1966-1972, to a mean of 4.8 years in 1973-1979. We used the date of earliest patent filing (including date of foreign claims priority) as an indicator of the company's initial active interest in the NCE. The dotted line in Figure 3 represents the start of clinical testing. uncorrected for retrospective IND filings. Failing to correct for the retrospective IND filings would substantially underestimate the
period of clinical testing and regulatory review (by more than one year from 1966 to 1970). Thus, the
uncorrected estimate of the increase in the clinical/regulatory period would be artifactually high by that amount. This could account for the apparent agreement previous authors observed be tween the decline in EPL and the increase in clinicalregulatory time for the period 1966 to 1976 (3).
References 1. Presidential press conference, 31 October 1979. 2. R. Hansen, *The pharmaceutical development process:
Estimate of developmental costs and times and the effects of proposed regulatory changes," in Issues in Phar maceutical Economics, R. Chien, ed. Washington, D.C.:
D.C. Heath and Company, 1979, 3. D. Schwartzman, "The life of drug patents," in Innova
tion in the Pharmaceutical Industry. Baltimore: The
University of Chicago Center for Policy Study, 1973. 5. G.F. Roll, "of politics and drug regulation," Medical
Marketing and Media, April 1977.
later in the drug development process realtive to the date of patent application, although one possible reason is the increase in preclinical data re quirements prior to first human testing. Related factors, such as compliance with the Good Laboratory Practice (GLP) regulations, could also require more time. Another possibility is that more prolonged in.. itial clinical testing is being done overseas – either by U.S. firms, or because a greater proportion of foriegn-originated drugs are getting U.S. INDs now than previously, either by licensing to U.S. firms, or through foreign-owned sponsoring firms. Further refinement of the data into subsets for selforiginated and licensed drugs of U.S. and foreignowned firms will enable us to examine the latter possibility.
Thus it is clear that the decline in EPL is a result of factors in both the drug development and patent processes. Taking the preclinical and clinical components together, a possible 73% (2.9 years) of the decline in EPL between 1968 and 1979 was accounted for by an increase in components influenced by the IND-NDA regulations, with the remainder of the decline influenced by the Patent Office Acknowledgement
This material is based in part upon work supported by the National Science Foundation under grant IDAR79--17602. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation, Footnotes (a) In this study we define NCEs as compounds of
molecular structure not previously marketed in the U.S., excluding new salts or esters, vaccines, antigens, antisera, immunoglobins, surgical products,
and diagnostic agents. (b) For NCEs with INDs filed after 1963, we used the
date of IND filing as the start of clinical testing in the U.S. The 30-day waiting period required since August 1970 has a conservative influence on our testing of the hypothesis. As described later, for NCEs that preceded the 1963 IND requirement, we used the actual date of first human administration
in the U.S., where available, (c) All data are complete for NCEs approved from 1966
to 1979, except for the following. Data on start of clinical testing are based on 81% (13 of 16) of patented NCEs for 1977, and 69% (11 of 16) for 1978. Two drugs were excluded from the pendency averages because their pendencies were excessive compared to all other drugs approved during the
game years (ie., 1978 and 1979). (d) The final IND regulations (Procedural and Inter
pretive Regulations, New Drugs for Investigational Use) printed in the Federal Register of January 8, 1963 required all drug sponsors to submit completed INDs by June 9, 1963 for all drugs in clinical trials as of August 10, 1962. Approximately 1100 drugs were assigned 1963 \i e., retrospective) IND filing dates during the initial period.
D.C.: American Enterprise Institute, November 1979.
Publishing Company, 1977.
Therapeutics, In Press.
origin as a measure of innovative output: The national origin of new chemical entities marketed in the U.S..." Report for the National Science Foundation grant number 75 19066. Also see Clinical Pharmacology & Therapeutics Vol. 19. January 1976. p. 108.
The Journal of LAW & ECONOMICS
ESTIMATING THE EFFECTS OF REGULATION
COMPARATIVE ANALYSIS OF THE
HENRY G. GRABOWSKI, JOHN M. VERNON,
Duke University INNOVATION in the U.S. ethical drug industry in recent years has been characterized by a number of adverse developments. In particular, there has been a sharp decline in the rate of new product introductions and the incentive for engaging in research and development (R & D) activity has been negatively infuenced by rapid increases in the costs and risks of developing new products. While there is little debate about the exist nce of these adverse trends, there is considerable controversy about the factors producing them.
Briefly, we list below five hypotheses that have been discussed as explanations for the declining rate of innovation. (1) Tighter regulation of the industry by the Food and Drug Administra
tion (FDA) has been largely responsible for the declining rate of inno
vation. (2) The decline is illusory—while there has been a decline in the total
number of new drugs being introduced, the number of “important"
new drugs introduced annually has not declined. (3) There has been a "depletion of research opportunities” brought about
by the rapid rate of new drug development in the 1950s. (4) The tragic thalidomide episode in the early 1960s made drug firms and
physicians much more cautious in their decisions concerning the mar
keting and prescribing of new drugs. (5) Advances in pharmacological science have led to increased safety test
ing and, therefore, higher costs of developing new drugs. In this paper, we present some new evidence on these hypotheses. Our
We are grateful for the comments we received on a preliminary version of this paper presented at the Third American University Seminar on Pharmaceutical Public Policy Issues. In addition, we received helpful comments from Sam Peluman, Dudley Wallace, and Oliver Williamson. The research was supported by the National Science Foundation, Division of Policy Research and Analysis.