Following the sudden resignation of FDA commissioner Lester Crawford, acting commissioner Andrew von Eschenbach has an unequaled
opportunity to set the record straight on the agency's role in drug safety. A good start would be to state clearly that, contrary
to popular belief, it is not the job of FDA to approve drugs that are safe and effective.
Kenneth I. Kaitin
Instead, von Eschenbach should affirm that FDA will vigorously adhere to its mandate to approve drugs for which the overall
expected benefits outweigh the known risks of using those drugs. Most significantly, it's the public that largely determines
where that risk-benefit balance is set.
Why should the acting commissioner be so bold? Because better public understanding of the risk-benefit relationship would
help set realistic expectations surrounding drug safety, and even help curtail the rising cost of new-drug development. It
also would go a long way toward restoring public confidence in FDA.
Critics of FDA and the research-based drug industry claim that there has been an alarming rise in the number of drugs that
have been withdrawn from the market for safety reasons in recent years. They also contend that faster approval times of new
drugs, one of the achievements of the Prescription Drug User Fee Act of 1992, have come at the expense of rigorous standards
for drug approval and have led to more safety problems.
While both assertions seem plausible, they are based on faulty foundations. Recent research by the Tufts Center for the Study
of Drug Development shows that:
- The number of drugs withdrawn for safety reasons since 2000, as a share of total new-drug approvals, has dropped in half since
the 1980s, following a slight increase in the 1990s.
- The rate of drug-safety withdrawals, based on the year of product approval, has not increased since the user-fee era began
- Faster approval times for new prescription drugs do not correlate with increased drug safety withdrawals, when grouped either
by decade of drug approval or by therapeutic class.
Perhaps most importantly, longer approval times do not assure greater drug safety. Average approval time for drugs withdrawn
since 1980 was actually longer than for all drug approvals during that time: 2.14 vs. 2.08 years.
The foregoing is not meant to suggest that drug safety isn't an important public health concern, or that FDA shouldn't address
inefficiencies in the processes of risk assessment and post-marketing surveillance. Much can be done to improve drug safety,
and the following steps will help.
Adequately fund FDA's Office of Drug Safety This three-year-old department needs continued strong support and funding. For example, coordination and communication channels
between the Office of Drug Safety and the rest of FDA (especially the Office of New Drugs) should be enhanced without compromising
its integrity. Calls, by some, for an independent drug-safety review board are misguided and send the wrong message. Drug
safety cannot be looked at in a vacuum; one must consider a drug's risks in relation to its benefits.
Leverage existing clinical data FDA possesses an enormous quantity of preclinical and clinical data on drugs that have been approved or are in various stages
of regulatory review. While current law forbids FDA from revealing proprietary clinical data, the agency could analyze these
data to make general findings, such as which specific enzymes or biomarkers are associated with particular drug toxicities.
This would cut development costs and speed development time by helping drug companies avoid re-discovering lessons learned
from past drug-development programs. It would also likely reduce future safety withdrawals.
Foster exchange of early research findings While it may appear counter-intuitive, companies have a strong incentive to share preclinical and clinical data, to avoid
engaging in dead-end research. Sharing this information will help firms identify unsuccessful drugs early on, lower R&D costs,
and reduce potential withdrawals. FDA, or a third party, could publish findings that provide useful information without identifying