IV. SUMMARY AND CONCLUSIONS

Innovation in the pharmaceutical industry has been subject to a number of adverse structural developments in recent years. There has been a sharp decline in the annual number of introductions of new chemical entities and rapid increases in costs and risks. We have reviewed these developments and listed five hypotheses that have been used to explain them: (1) increased regulation of the industry associated with the 1962 amendments to the Federal Food, Drug, and Cosmetic Act is the cause; (2) the decline is illusory since only ineffective NCEs have declined; (3) a depletion of research opportunities has taken place; (4) the thalidomide incident has made firms and physicians more cautious; and (5) costs have risen as a result of advances in the technology of safety testing.

In order to separate the effects of regulation from these other confounding factors, we developed an international comparative analysis of R & D productivity changes in the United States and the United Kingdom.

A principal finding that emerges from this international comparative analysis is that U.S. “productivity”—defined as the number of new chemical entities discovered and introduced in the United States per dollar of R & D expenditure-declined by about sixfold between 1960-61 and 1966-70. The corresponding decrease in the United Kingdom was about threefold. Clearly, some worldwide phenomenon, which might be labelled a "depletion of research opportunities”—but which probably also includes the effects of other factors such as the thalidomide incident and higher costs due to new developments in safety testing-seems to hold for pharmaceutical R & D. However, there is also strong support for the hypothesis that an additional factor has been at work in the U.S. industry.

We conclude that this additional factor, which has lowered U.S. productivity at a significantly more rapid rate, is the increased regulation resulting from the 1962 amendments. On the basis of the regression analysis presented in Section III, we estimate that the 1962 amendments have probably, at a minimum, doubled the cost of a new entity.

Our analysis also suggests that nonregulatory factors have an important aggregative effect on innovation, but does not allow us to say which factors in particular have been most important in this respect. Further research on this question would seem warranted.

APPENDIX

This appendix presents in summary form the sources and methods of computation for statistics used in the paper.

NCE INTRODUCTIONS AND DISCOVERIES

Data on new chemical entities and their years of introduction for both the United States and the United Kingdom were obtained from the publications of Paul de Haen. In a very few cases, information on British introductory dates was supplemented by the work of William Wardell." Biologicals and diagnostics were here deleted from data lists and analysis due to problems of data availability and reliability prior to 1966.

Information as to which of these NCEs were also discoveries by industry research laboratories was obtained for the United States from Paul de Haen," for the United Kingdom in 1960-1970 from the National Economic Development Office," and for the United Kingdom in 1970-1974 from, again, Paul de Haen." An NCE was regarded as discovered in a particular country if the research laboratory producing the entity was located in that country, irrespective of the nationality of laboratory ownership. Thus the discoveries of Pfizer in the United Kingdom are credited to Britain while those of Hoffmann-La Roche in the United States are considered as American. It should be recognized that the discoveries of NCEs are denoted by year of introduction in either the United States or the United Kingdom (depending on origin) rather than first year of introduction on a worldwide basis (should these dates differ).

R & D EXPENDITURES

Expenditures for research and development are here considered as those domestic outlays by the pharmaceutical industry for discovery of humanly usable ethical drugs. In the United States, data were obtained from publications of the Pharmaceutical Manufacturers Association (PMA)” for worldwide human R & D expenditures, 1948-1974, of member firms. However, the breakdown of domestic versus foreign

54 Paul de Haen, Compilation of New Drugs, 33 Am. Professional Pharmacist 25-62 (Nov. 1967); id., 7 New Drug Analysis USA, 1966-1970 (1971); id., 10 New Drug Analysis USA 1969-1973 (1974); id., New Products Parade (20th ed., mimeographed, Feb. 1975); id., New Single Drugs Marketed in England, France, Germany, and Italy 1960 to 1965 (mimeographed, Feb. 1973); id., New Single Drugs Marketed in England, France, Germany, and Italy 1966, (mimeographed, Oct. 1973); id., 1 New Drug Analysis Europe, 1967-1971 (1972); id., 4 New Drug Analysis Europe, 1970-1974 (1975).

55 W. M. Wardell, Introduction of New Therapeutic Drugs in the United States and Great Britain: An International Comparison, 14 Clinical Pharmacology & Therapeutics. 773-90 (1973).

56 Paul de Haen, Compilation of New Drugs, 33 Am. Professional Pharmacist 25-62 (Nov. 1967); id., 7 New Drug Analysis USA, 1966-1970 (1971); id., 10 New Drug Analysis USA, 1969-1973, (1974); id., New Products Parade (20th ed., mimeographed, Feb. 1975).

57 National Economic Development Office, A List of 466 Pharmaceutical Compounds and Country of Discover (mimeographed, 1971) (prepared for NEDO by the Centre for the Study of Industrial Organization as part of the study, Innovative Activity in the Pharmaceutical Industry).

5 Paul de Haen, 1 New Drug Analysis Europe, 1967-1971 (1972); id., 4 New Drug Analysis Europe, 1970-1974 (1975).

" Pharmaceutical Manufacturers Association, Annual Survey Report (various years); id., Office of Econ. Research, Prescription Drug Industry Factbook (1967).

expenditures in this total was available only for 1960-1974, from the same sources. By fitting an exponential trend for foreign R & D expenditures of PMA member firms against time, 1960-1974, estimates of this parameter were obtained for earlier years. Subtraction of these estimates from the worldwide total gave the data used in the

text.

R & D data for the United Kingdom for 1954-1966 and 1973 were taken from releases of the Association of the British Pharmaceutical Industry.60 For 1954 to 1965, the data aggregated human and veterinary research expenditures. These statistics were multiplied by 86.1 per cent (the 1966 value) to obtain estimates of expenditures for purely human research. For the years 1966 to 1974 an exponential trend on time was fitted to obtain R & D estimates for intervening years.

R & D estimates for both industries were deflated by the gross national product deflator to constant (1958) dollars for the United States and to constant (1958) pounds for the United Kingdom.62 Statistics for deflated expenditures on R & D as well as introductions and discoveries of NCEs are plotted in Figures I and II of the

text.

PHARMACEUTICAL SALES

Data on U.S. sales of ethical drugs were obtained from the publications of a marketing research firm, Intercontinental Medical Statistics." These data were based on a projection from a 1,000 drug store sample to the population of all U.S. drug stores, and on a sample of about 10 per cent of total hospital beds. Sales directly to other institutions, such as to the U.S. government are here excluded, but they account for less than 20 per cent of U.S. ethical drug sales.

FDA STRINGENCY

Estimates of the mean time in months to FDA approval of NCEs introduced in the United States were taken from an unpublished dissertation of Joseph M. Jadlow." Jadlow obtained his estimates through private communication with the FDA. The figures used in the text extrapolate from Jadiow's and are as follows:

Association of the British Pharmaceutical Industry, Annual Report 1973-1974, (1974); id., Pharmaceutical Research and Development Survey (mimeographed, Jan. 17, 1975). "Economic Report of the President, Together with the Annual Report of the Council of Economic Advisors (1975).

Central Statistics Office, Annual Abstract of Statistics (London, various years). "Intercontinental Medical Statistics, Pharmaceutical Market-Hospitals (various years); id., Pharmaceutical Market-Drugstores (various years).

"J. M. Jadlow, Jr., The Economic Effects of the 1962 Drug Amendments 174 (1970) (unpublished Ph.D. dissertation, University of Virginia).

These values are defined as the variable S, the logarithm of which is used in equations (3.2) and (3.4) of Table 3.

LAGS FOR EFFECTIVE R & D EXPENDITURES

Estimates of development times for NCEs were interpolated from figures offered by Dr. Lewis Sarett.65 Addition to these development times of the regulatory approval times given above yields the following estimates of total lag times, from first expenditure to introduction:

R & D expenditures in a given year become effective over a three-year period centered around the (mean) total development period. For example, expenditures in 1967 are seen as effective in 1973, 1974, and 1975 at the rate of one-third of original 1967 expenditures. Total effective expenditures are obtained by summing over all expenditure portions which become effective in the given year and are defined as the variable V in Table 3. While admittedly stylized, this lag system appears to capture the essence of the process at issue. Further, alternative lag structures based on the above mean lag estimates, as well as minor alterations of the mean lag estimates themselves, yielded qualitatively similar results in all cases.

It should also be noted that in estimating the U.K. trend for the restriction in the increasing lag case, an increasing development period ranging from two to five years was assumed.

MECHANICS OF ESTIMATION

The specification assumed for equation (2) in the text can be written as:

where (1) as is restricted to equal U.K. trend

(2) is 1 in 1954, 2 in 1955, .

(3) X = 0 from 1954 to 1959 and unity thereafter.

Hence, the variable Tpreso in equation (2) is the multiplier of a, above and Teo is the multiplier of a,. The reason for the rather complex definitions of these two time trend variables is to ensure that the two time trend segments join properly in 1960. Thus, a, is the rate of decline of N/E from 1954 to 1960 and a, is the rate of decline thereafter. Similarly, the specification of the log-log version of the above equation, equation (3.1) in Table 3, can be written in terms of and X as follows:

log(N/E) = b2 + b2D + b2[(1 − X) log + Xlog 7] +b3(Xlog - Xlog 7),

where by restricted to equal U.K. trend.

Thus, as above, the variable LT preso in Table 3 is the multiplier of b, above and LT is the multiplier of b1.

65 L. H. Sarett, supra note 5.

HARVARD JOURNAL

of

LAW AND PUBLIC POLICY,

LESSONS FROM THE DRUG LAG:
A RETROSPECTIVE ANALYSIS OF THE
1962 DRUG REGULATIONS

by

Leonard G. Schierin*

INTRODUCTION

This paper examines the medical and economic literature concerning the effects of the 1962 Drug Amendments' on drug innovation in the United States. The effects represent different facets of what has come to be called the “drug lag," and have been discussed and debated in a wide variety of studies over many years. Among these studies have been periodic overviews of the literature that have weighed the sum total of the existing evidence on the magnitude, causes, and impacts of the drug lag.

This study follows the overview approach, but extends its perspective both within and beyond the drug industry. While continuing to survey the literature as a whole in order to test hypotheses about the characteristics of the drug lag, it also examines the methods by which the impacts of a lag may accurately be assessed and the processes by which regulation generates or contributes to such a lag. By understanding the strengths and weaknesses of available assessments and the dynamic causation linkages between regulation and innovation, we can move closer to accomplishing what quantitative estimates alone of the lag cannot provide. First, we may be able to render more accurate evaluations of the effects of existing drug regulations and of proposed changes in those regulations. Second, knowledge of the successes and failures in assessing the full societal impacts of drug regulation and of the

• Professor Schifrin received his B.A. and M.A. degrees from the University of Texas at Austin, and his Ph.D. degree from the University of Michigan. He taught at Michigan and Yale before coming to William and Mary in 1965. His main research area is in the economics of health care, particularly prescription drugs, and the application of cost/ benefit analysis to health care decision making.

This article was produced in part from work funded by the Office of Technology Assessment (OTA) of the United States Congress for use in its study "Technological Innovation and Health, Safety and Environmental Regulations." The views expressed do not necessarily represent those of the OTA.

1 U.S. Code 1976, Title 21, § 321 et seq., P.L. 87-781, 76 Stat. 780 (October 10, 1962).