It's the physician's job (to adapt a recent campaign phrase) to pick the right drug for the right patient at the right time.
It's FDA's job to inform that decision. Clearly, as demonstrated by 2004's controversies over antidepressants and COX-2 inhibitors,
the system is not working optimally.
Jonathan Seltzer, MD, MBA
The problem is rooted in the fundamental nature of the agency: Although FDA has significant expertise in overseeing drugs
in development; it is not structured adequately for the demands of researching a drug after approval. The regulatory culture
and tools used in drug development would be too cumbersome to use in the post-marketing environment. For instance, the gold
standard of development, the double blind placebo-controlled trial, is impossible in the marketplace. A traditional regulatory
approach cannot—in a reliable, prospective fashion—take into account the multitude of ways patients and healthcare systems
To its credit, FDA is aware of these issues, and under Mark McClellan the agency made risk management—defined as the "overall
and continuing process of minimizing risks throughout a product's lifecycle to optimize its benefit/risk balance"— one of
its key initiatives. Theoretically, implementation of a risk management framework would use focused safety interventions instead
of mass-market drug recalls. This would prevent the removal of a compound that, for many, is safe and effective and minimize
the anxiety of patients who are not at risk. But for a risk management framework to succeed, both operational and cultural
initiatives must be implemented:
End perverse regulatory incentives. The finish line for most manufacturers is drug approval; there is little, if any, true incentive for rigorous post-marketing
safety surveillance. Once a drug is approved, companies are free to market it to anyone at any time, and increasingly at any
place—as long as they stay within approved regulatory language. Society's needs would be better filled if FDA gave new products
provisional approval with limited marketing privileges, followed by full approval, based upon robust post-marketing safety
Build trust through transparency. One mutual fund now promotes itself by saying, "The Fund invests in socially responsible, high-quality growth stocks with
no investments in the tobacco, alcohol, gambling, or pharmaceutical companies." This shouldn't be surprising when Merck, Pfizer,
and FDA can't agree on how something works. With regard to the COX-2s, they are all looking at different data in different
ways. In this situation, the scientific truth is ultimately transformed to a legal one. What is needed instead is transparency
of data, methodology, and decision making, supported by open data standards.
Improve legal protection. The goals of transparency and continuous improvement cannot co-exist with the constant threat of tort litigation. Risk itself
connotes a cost and a benefit. Patients and physicians should decide the proper balance for individual conditions. Poor outcomes
cannot be allowed to remain a standard for lawsuits.
Modify Big Pharma company culture. Big Pharma supports FDA's risk management initiative, but does it have the right culture to live with it? Risk management
requires a proactive, innovative culture. Big Pharma, dependent as it is on the regulatory rulings of FDA, tends to be reactive
and ultraconservative. Changing that culture will be a major challenge.
Improve communication and education. Few people understand scientific decision making. Even medical students are not required to study the process of drug development.
This makes it impossible for even physicians to accurately communicate the vagaries of drug therapy to patients and the public.
Establish a special safety agency. It's reasonable to debate whether FDA is the best choice to manage post-market safety surveillance. At a minimum, additional
expertise in economics, psychology, and epidemiology is needed. It's also important to decide whether the authority that declares
a compound safe and effective also ought to monitor its actual performance.