Congress Responds to the IOM Drug Safety Report—In Full

The US drug safety system developed fitfully through a series of historical accretions, typically in the wake of major drug safety scandals. In 1962, largely precipitated by the thalidomide tragedy, the Kefauver-Harris amendments to the Federal Food, Drug, and Cosmetic Act required the sponsor to provide “substantial evidence” of both safety and efficacy during a premarket evaluation. The resulting multistage approval process improved the evidence base for premarket regulatory decision making about drugs, but it also created, as an unintended consequence, a postapproval environment in which sponsors were largely free from requirements for additional evaluations. At approval, the sponsor's scientific teams responsible for meeting US Food and Drug Administration (FDA) requirements were generally replaced by marketing teams, and the manufacturer's efforts, energies, and studies were generally directed toward marketing its drugs.^{1 - 2}

The formulaic preapproval evaluation process, even though usually of high quality, slowed drug approval. The 1992 Prescription Drug User Fee Act (PDUFA) provided new funds from user fees, limited their use to speeding preapproval evaluations, set timelines for reviews, and thus brought drugs to market more quickly. As the United States became increasingly the country of first launch for many new drugs (from 2% in 1980 to 60% in 1998),³ the US postmarketing safety system accordingly shared a larger responsibility for the early detection of safety signals. For the first 10 years following enactment of PDUFA, however, Congress prohibited the FDA from applying user fees to drug safety and provided no additional funds for monitoring marketed drugs. Not until the PDUFA reauthorization of 2002 was the FDA permitted to apply some user fees for selected safety issues.⁴ In effect, Congress largely entrusted the detection, evaluation, and timely reporting of postmarketing safety issues to the pharmaceutical industry.^{5 - 6}

The withdrawal of rofecoxib and other drug safety problems⁷ have again focused attention on the limitations of the US postmarketing system. In response to an FDA request, the Institute of Medicine (IOM) undertook a thorough assessment and issued a comprehensive set of recommendations for reform.⁴ The major theme of the report was the need for an ongoing effort to acquire, integrate, and communicate information about a drug's risks and benefits during its entire market life. This “lifecycle” approach requires the translation of premarket safety signals into high-quality postmarketing studies, sophisticated active surveillance to identify new serious safety signals, and timely assessments of the drug's risk-benefit profile. While the FDA has already responded in part to some recommendations,^{8 - 9} others dealing with the agency's resources and authorities required congressional action.

In crafting the FDA Amendments Act (FDAAA) of 2007 (Pub L No. 110-85), Congress used many of the IOM committee recommendations as a blueprint (Table). The FDAAA provides for new resources for drug safety, new authorities and enforcement tools, an active postmarketing surveillance system, expanded clinical trials registration, new requirements for public disclosure of trial results, and improvements in transparency and communications with patients and physicians.

The FDAAA gives the FDA the authority to require postmarketing studies to identify or assess potential serious risks. In response to new safety information about marketed drugs, the FDA can also initiate timely label changes or new postmarketing studies. To ensure that drug benefits outweigh risks, the agency may use Risk Evaluation and Mitigation Strategies (REMS), which include various restrictions on distribution or use. Failure to complete a postmarketing study, a label change, or REMS may result in a determination of misbranding or a civil penalty of $250 000 for each violation up to $10 million for an ongoing violation.

The FDAAA requires the FDA to establish an active postmarketing risk identification system that provides access to federal and private electronic medical records data. The system is to include 25 million patients by July 1, 2010, and 100 million by July 1, 2012. Working with a committee of experts, the FDA is to develop validated methods for the timely identification of adverse events and potential drug safety signals and to establish collaborations to provide for advanced analyses of drug safety data to improve risk-benefit analyses. Complementary methods of postmarketing drug surveillance are permitted. Congress allocated $25 million per year in 2008-2012.

The 1997 reauthorization of PDUFA had only required registration of clinical trials involving serious and life-threatening conditions. Now, FDAAA requires that within 21 days of the enrollment of the first patient, all phase 2, 3, and 4 drug trials must be registered in a database publicly available online. For approved drugs, FDAAA also creates a clinical trials results database that, when fully implemented within 3 years, will include a set of Internet links to key FDA documents; summary tables of primary and secondary outcomes; expanded results to include nonpromotional summaries; and information about frequent or serious adverse events. Clinical trial results must be submitted within 1 year of trial completion or within 30 days of drug approval. The penalty for failure to submit data is $10 000; for violations not corrected within 30 days, the fine is $10 000 per day until corrected. The FDAAA also requires the FDA to decrease the agency-wide proportion of conflict-of-interest waivers granted to advisory committee participants by 5% per year.

The FDAAA establishes a new Advisory Committee on Risk Communication and provides for improved transparency and access to drug safety information. For new molecular entities, a summary review, to be published within 48 hours of approval, is to include conclusions from all reviewing disciplines, information about critical issues or disagreements with the applicant or within the review team and their resolution, final recommendations, and an explanation for any division-level nonconcurrence with review conclusions. The full package of all documents generated by the FDA review is to be posted within 30 days of approval.

The FDAAA creates a new office of the Chief Scientist, who will oversee intramural regulatory science and develop rigorous safety performance measures (IOM 4.7 and 3.5). FDAAA includes other significant provisions: a scientific review is the reviewer's work and may not be altered by management; FDA employees' scientific articles, prepared for journals, must be reviewed by agency management in a timely fashion; and an annual review of postmarketing commitments is designed to clear out the backlog,¹⁰ many of which have been “pending” for years.

Although the IOM committee expressed a strong preference for increased resources from general revenues, Congress retained the user-fee approach and provided the agency with additional fee revenues for drug safety, increasing from $25 million in 2008 to $65 million in 2012.

Congress embraced and passed many of the science-based recommendations from the IOM committee. The House vote, 405 to 7 on HR3580 (the FDAAA), bespeaks an important bipartisan consensus about drug safety. The new authorities and resources and the mandate to create an active surveillance system will enable the FDA, for the first time, to implement an ongoing evaluation of the risks and benefits of new drugs during their entire market life. This difficult task needs to be done well.

Perhaps the most fruitful approach will involve a concerted effort to identify potential safety signals during the premarket evaluation. These signals need to be pursued aggressively, sometimes with postmarketing commitments conducted by sponsors, sometimes by active surveillance of a large cohort with electronic medical records. Although not explicitly mentioned in the FDAAA, the FDA will on occasion need to insist on large randomized trials to convert surrogate end points such as glycemic control into valid and reliable information about actual health benefits.^{7 ,11}

As the FDA acquires access to large databases, attention to quality may be more important than an emphasis on size. Large error-prone data sets provide precise but biased answers.¹² Before FDA scientists undertake exploratory analyses with large data sets, they need to have a plan for evaluating findings. The possibility of false-positive safety findings is large, and failure to recognize them can undermine the credibility of the surveillance system.

The FDAAA represents the most extensive revisions of the Food, Drug, and Cosmetic Act since 1962. In time, evolutionary adaptations by an industry that has extensive resources, devoted scientists, and a duty to shareholders is likely to occasion the need for future legislation. In the meantime, ongoing congressional oversight will be essential.¹³

Despite high-quality preapproval evaluations, some drugs with unrecognized toxicity will reach the market. The goal of a well-functioning postmarketing system is to identify risks and confirm health benefits of drugs in a timely fashion so that patients and physicians can make informed decisions. An optimal system will require cooperative interactions among the pharmaceutical industry, academic medicine, and the FDA; yet the credibility of the system demands that the agency orchestrate and oversee the system, foster improvements, and respond promptly and effectively to new safety signals. Congress has provided FDA with the resources and the tools to fulfill its mission to protect and promote the health of the public. The measure of the FDA leadership in coming years will be the degree to which the newly deployed postmarketing system becomes a model for other drug regulatory authorities.