oversight once research is under way. This chapter thus expands the committee’s responses to Questions 1, 3, and 4 of its charge (see Box 1-1).
The chapter begins with a discussion of relevant features of the complex postmarketing context that bear on the selection of the types of designs to require. It then puts forward two sets of considerations that frame the decision problem more specifically: the scientific and practical criteria for assessing the advantages and disadvantages of alternative research designs and the statutory conditions and ethical preferences that often favor observational designs over randomized controlled trials (RCTs) in the postmarketing setting. It then identifies the conditions under which it is acceptable for FDA to require each. The committee then discusses factors that affect choosing among types of observational studies and RCTs, once it has been decided to require an observational design, an RCT design, or both. That discussion is followed by an analysis of the ethical considerations that should guide the design and conduct of postmarketing research, including issues related to informed consent and safety monitoring, and FDA’s ethical obligations in the postmarketing setting.
Because concerns about a drug’s benefit–risk profile can emerge throughout a drug’s lifecycle, the decision by FDA to require a manufacturer to conduct postmarketing research can occur when the drug is first approved or at any time thereafter.
In the premarketing setting, the RCT is the standard for providing the efficacy data used by FDA to make approval decisions, although occasionally less-well-controlled designs have been accepted if the effect is of sufficient magnitude. In contrast, population-based observational designs, which require drug use in larger clinical populations, play no role in approval decisions unless approval in other countries has provided opportunity for observational study. Over the years, the agency has amassed considerable expertise in the design and interpretation of RCTs and in the ethics of randomizing research participants to receive unapproved, investigational drugs. For example, FDA has provided guidance documents on good clinical practice, institutional review boards, and informed consent (FDA, 2011a), and FDA officials have opined on when it is ethically acceptable to use placebos, rather than active comparators, to evaluate an investigational drug (Ellenberg and Temple, 2000; FDA, 2001; Temple and Ellenberg, 2000).
FDA has less experience in undertaking comprehensive assessments of the ethical and scientific issues arising in the postmarketing context, which differs from the premarketing context in several relevant respects. For example, in the premarketing context, a drug’s manufacturer sponsors all the research with the investigational drug and generally uses randomized controlled designs developed in consultation with FDA (after Phase 1 and early Phase 2 studies); in contrast,