Some Remarks about “Safety”

The term safety is often used in a simplistic and misleading manner. Safety depends on many particulars of the individual case. For example, one of the authors of this Web site (Klein) ingests ten milligrams of a blood thinner called warfarin every day. Warfarin is also available in hardware stores as rat poison. When mixed with something tasty, rats gobble it up and die from internal bleeding. Humans with vascular problems reduce risks of thrombosis by ingesting small amounts of rat poison. Large amounts would be deadly. Dosage and condition are just two of the many factors that affect safety.

Is chemotherapy safe? Medicine is often poison. University of Iowa professor of medicine William B. Bean (1970) explained the point as follows:

[T]he power of a drug [usually] carries with it almost in parallel a potential for dangerous reactions. . . . [E]ach person in the world is biochemically, anthropologically, genetically, and in any other way you wish individual and unique. The postulate that there is a dose, a fixed dose, of a drug which is routine for any person with a specific disease—a standard to be applied by pressing a button in a kind of therapeutic automat—is irrational. Absorption varies. Internal bodily metabolisms are different. At one end of the spectrum there are persons who react idiosyncratically or allergically to a drug. In addition there are sensitivities that exist without harming a person at all until he encounters a biochemical agent to which he may have a violent reaction, even though the majority of people exposed to the same thing do not react in that way. After a while he may learn not to revolt against its administration. . . . [I]n the practice of medicine we should remember that [powerful and potent drugs] are two-edged swords that cut back at you if you are not very careful in the way you wield them. With increased potency we have increased danger. (132–33)

In 1994, it is estimated that at least 106,000 people died from adverse reactions to “safe,” FDA-approved drugs (Lazarou et al. 1998). (This figure includes only hospitalized patients and does not include those people who died because of medical error such as the prescribing of the wrong drug, which also accounts for some 100,000 deaths every year [Kohn, Corrigan, and Donaldson 1999].) It is inevitable that many people, often in weakened and uncertain condition, will suffer and die from unwanted side effects.

Despite the subtleties involved, the FDA is set up to screen out “unsafe“ drugs for all cases of usage, taken in aggregate. But neither the administering of drugs by doctors nor the taking of drugs by patients is done on an aggregate basis. A drug stamped “safe” by the FDA is usually not safe for every particular case or individual, and a drug not so stamped is, nonetheless, safe for many particular cases and individuals. As the Warfarin and chemotherapy examples demonstrate, it might be “unsafe” both to take the drug and not to take the drug. But the relevant benchmark is not a state of perfect health. What really matters is whether taking the drug is safer than not taking the drug. The people intimately concerned in the situation and intimately informed, not bureaucrats in Rockville, Maryland, should make this determination, however.

The safety of a drug depends on myriad particulars about the patient, including age, sex, physical strength, condition, activities, allergies, diet, dosage, medical attention, and drug regimen. Furthermore and importantly, what is “safe” contains an unalterable subjective component (Higgs 1994). Patients faced with the same diseases will make different treatment decisions depending on how they perceive and evaluate the inherently risky trade-offs among health, pain, and disability. The establishment of a single society-wide standard of safety and efficacy does violence to the reality of myriad individuals with different values and experiences. We see the dilemma most clearly when the FDA withdraws an “unsafe” drug from the market, and patients complain. The drug Lotronex, for example, was withdrawn from the market in November 2000 under pressure from the FDA. Lotronex was prescribed for irritable bowel syndrome, a disease that causes abdominal pain and intense bouts of diarrhea. In the ten months that it had been on the market, some three hundred thousand people used the drug without serious problem, but seventy users developed a serious side effect, and three deaths were possibly linked to the drug (Grady 2001). Lotronex is thus not without complications, but it improved the lives of many patients to the extent of allowing them to hold a job and leave their homes without fear of pain or sudden attack of diarrhea. On learning that Lotronex might be withdrawn, many of these patients went to great efforts to stockpile the drug, and when it was indeed withdrawn, they complained vociferously to the FDA. Patient complaints were so extensive that in June of 2002 the FDA allowed the drug back on the market so long as doctors and patients fulfilled a number of conditions ensuring that both were fully informed about the drug’s potential complications.

The FDA’s typical policy is “one size doesn’t fit all,” but the final outcome of the Lotronex example is encouraging. The FDA could better serve all patients if instead of acting paternalistically, by making choices on behalf of patients whom it cannot know or understand, it focused on collecting and disseminating information that people could utilize within the context of their own lives.

It’s important to recognize that the Lotronex case is unusual because patients knew that the FDA had taken away a drug that was beneficial to them. In the usual case, the FDA fails to approve a drug that could have benefited patients, and the patients never learn of the drug’s possible existence. (On the importance of what is seen and unseen in FDA policy, see the glossary item Types of FDA Error and the section FDA Incentives.)