Assessing the Effects of the Drug Lag

We still do not know if the 1962 Amendments have generated more good than harm: Jondrow says they have, fadlow and Peltzman contend they have not, but all three analyses suffer from serious limitations or errors. Wardell argues that the therapeutic costs of the Amendments are very large, yet he does not consider the therapeutic gains or any of the economic effects.

While benefit/cost analyses of the 1962 Amendments are useful, Wardell has suggested that much of the therapeutic cost of regulation can be mitigated through administrative practices and interpretation. The implication is that much of the benefit of the Amendments, and regulations in general, can be maintained and much of the cost reduced by relatively small changes in the existing regulations rather than by wholesale repeal or major reconstruction. Accordingly, benefit/cost analysis with a large scope (drug regulation as a whole or of major portions of it), or a small scope (pertaining to revisions, perhaps minor, of existing regulatory policy), may well provide useful guides as to both the direction and form of change that will serve society best. None of the studies to date, as we have seen, has provided a satisfactory frame of reference for such analyses. However, they collectively offer most of the necessary components. This study now turns toward an integration of these components to provide guidelines for benefit/ cost analyses of existing and proposed drug regulatory policies.

The appropriate benefits and costs to include in such an analysis are presented in Table 11. The table shows those benefits and costs applicable to the drug lag, and those additional benefits and costs necessary for an assessment of the regulations as a whole.

The quantification of these benefits and costs is more readily accomplished by following Wardell's approach rather than Peltzman's, thus using actual experience of the drugs in question rather than hypothetical or generalized historical results. Such an approach relies heavily on subjective medical judgment, but it must suffice in the absence of other ways of measuring health care outcomes.

Carefully selected foreign experience, perhaps that of Canada, or Great Britain, or several countries who lead the United States in the rate and timing of innovation, provides the basis for quantifying most of the benefits and costs in Table 11. Benefit I and Cost I, which probably are the largest of the components comprising total

benefits and costs, can be estimated by following the general steps: (1) Identify the specific new drugs "lagged" in their introduction into the United States.

(2) Estimate the incidence ratios both of adverse consequences (whose aversion is a benefit) and of outcomes more favorable than offerred by the best alternative therapy available in the United States at the time (these foregone superior outcomes are costs of the lag to U.S. patients).

(3) Apply these ratios of adverse and superior outcomes to the respective American populations-at-risk to determine the absolute frequencies of both in the United States that would have occurred but for the lag.

(4) Estimate increased or reduced disability and work, subtracting periods that result from these adverse or superior outcomes. (5) Calculate the dollar values associated with the averted or incurred medical treatments and productivity gains and losses. The product of (3) and (5), with appropriate discounting, represents the value of the benefit or cost in question.

Benefit II can be calculated by looking at foreign drugs that arc removed from the market after appearing only in countries outside of the United States. Their usage rates and prices, adjusted for dif ferences between the United States and foreign nations, estimates the savings that patients in the United States would have obtained from drugs never entering the American market.

Benefit III is the gain to customers from the increased knowledge about all drugs resulting from increased testing. We can calculate this benefit by employing inter-temporal U.S. comparisons, international comparisons, or some combination of both that show shifts or differences in the ratio of unfavorable to favorable outcomes associated with drug use attributable to the additional information generated by the heightened requirements. Complex factors are involved, including the effects of learning by experience, international transfers of knowledge, improved testing methodology, and legal liability. However, multivariate analysis may enable us to sort out the relative influences of each such factor.

Costs II and III, the increase in real costs both for lagged and other drugs, can draw on the many studies cited in this report that deal with changing R&D outlays. Special attention here needs to be given to the use of appropriate definitions of R&D activity and accurate indexes of changing R&D input prices. Cost IV, administrative costs of the regulations, can be quite easily derived from FDA budgetary and activity reports.

The above guidelines admittedly are very general and unqualified. All the nuances of benefit/cost analysis must eventually be utilized; but as a general overview, they offer initial directions that can produce useful retrospective studies of the effects of the lag and of the regulations creating the lag. Once this approach is worked out, it can tell us much about the marginal benefits and costs resulting from changes in the requirements that apply to the quantity or timing of drug development in general, or certain drugs or groups of drugs in particular. While the prospective benefit/cost analysis of major regulatory changes may be the most problematical task of all, this framework at least provides some useful methodology for both identifying and quantifying those anticipated effects.

New drug development during and after a period of regulatory change: Clinical research activity of major United States pharmaceutical firms, 1958 to 1979

The 1962 drug amendments fundamentally changed the way in which U.S. pharmaceutical firms
could test new drugs in man and receive New Drug Application (NDA) approval. Although it is
well known that the amendments and associated events caused a profound decline in the annual
number of new drugs receiving NDA approval, the amendments' effects on clinical research into
new chemical entities (NCEs) have not been investigated because data were unavailable. To
study this we requested drug development information dating back to 1958 from most major
United States-owned pharmaceutical firms and obtained complete responses from nine. The
results showed that the introduction rate of NCEs into human testing dropped sharply in the
early 1960s and declined substantially thereafter. The number of NCEs entering human testing
fell from a mean of 89 a year in 1958-1962. to 35 a year in 1963-1972 (a reduction of 61%).
and to 17 a year in the last 5 years of the survey, 1975-1979—an overall reduction of 81%.
The number of NDA approvals received by these firms fell sharply by 19% in the early 1960s
and more slowly for 10 years thereafter, from the mid-1960s to the mid-1970s. In the case of
self-originated NCEs, the size of this later fail was 71%. Causes of these changes in NCE flow
include the amendments and the events that prompted them; changes in scientific philosophy.
standards, and state of the art; and economic factors.

Maureen S. May, Ph.D., William M. Wardell, M.D., Ph.D., and Louis Lasagna, M.D.
Rochester. N.Y.

Center for the Study of Drug Development. Department of Pharmacology. University of
Rochester Medical Center

Based on research supported by the National Science Foundation under graat 79-17602. Any opinions, Andings, and conclusions or recommandations are those of the authors and do not necessarily reflect the views of the National Science Foundation. Received for publication Nov. 27, 1982.

Accepted for publication Nov. 29, 1982.

Reprint requests to: Dr. William Wardell, Department of Pharmacology. University of Rochester Medical Center, Rochester, NY 14542