above, establishes a government-wide process, then FDA's process can be a constituent part, and it can focus on the measures over which it has authority. FDA should establish a scientific panel to perform the functions recommended above, even if Congress and other parts of the executive branch fail to do so.
Health risk is proportional to the degree and duration of exposure to tobacco products. One way to reduce health risk is thus to reduce exposure. A history of product modifications that were marketed as health improvements but did not actually reduce exposure—two examples are filter cigarettes that did not reduce tar or nicotine intake and "low-nicotine" cigarettes that merely led smokers to inhale deeper and longer to achieve the same nicotine dose—has raised doubts about harm reduction strategies, as opposed to complete cessation and primary prevention. The advent of new cessation products, however, has turned attention once again to treating tobacco dependence through long-term nicotine maintenance, to reducing tobacco use without eliminating it, and to designing tobacco products that cause less exposure to toxins. The 1994 IOM report noted great uncertainty surrounding harm reduction strategies, and three years later, that uncertainty is not much diminished. The value of long-term use of nicotine products to enhance cessation is, however, the subject of increased study.45 In addition, an international body recently concluded that "whereas total cessation remains the ultimate goal of tobacco control policy, reduction of exposure to tobacco toxins should be added to the existing treatment approaches."46 Many questions can be addressed through empirical research. Can safer tobacco products be developed for those who do not quit? What are their individual and public health impacts? Is it more effective to provide tobacco without nicotine (e.g., nicotine-free cigarettes proposed for marketing in the near future) or nicotine without tobacco (e.g., nicotine gum, patches, nasal sprays, lozenges, and inhalers), or are both useful for different users? How will such products affect nonsmoking teenagers? Would they increase the level of initiation of tobacco use in the short or long term by reducing perceived risk or reduce the level of initiation because such risk is part of the attraction to young people? These and many other questions are closely related to those about long-term regulatory strategies and would rest on some of the same research foundation. That foundation is extremely weak now.
The two principal sources of new public knowledge will be (1) documents made public about studies now available only to private tobacco firms and (2) new research published in the public domain. This implies a need to scrutinize documents that become public in cases such as the Minnesota Medicaid suit to see if information pertinent to harm reduction or other regulatory questions comes to light. More important, the federal government needs to initiate research programs to address the questions that arise in connection with a harm reduction approach.