About a month into her first job at the Food and Drug Administration, Frances Kelsey was handed what her supervisors called an easy file: a sedative named Kevadon, already sold across Europe for about three years, with nothing in it likely to cause trouble. Under the law then in force, a new drug became approved automatically once sixty days passed without an objection from the agency. All she had to do, as far as anyone expected, was let the clock run.
She did not let it run. Of the roughly thirty physicians' reports the manufacturer, Richardson-Merrell, had submitted as proof of safety, most read as testimonials rather than controlled studies, and the claims outran what they actually showed. Kelsey, the medical officer on the three-person review team, wrote back that the application was incomplete. Under the rule, that reset the sixty days. Merrell resubmitted. She wrote back that it was still incomplete.
In late January or early February 1961, a British Medical Journal issue delayed by a mail strike finally reached her desk, describing painful, sometimes irreversible nerve damage in European patients who had taken the drug for months. She asked Merrell's liaison, a bacteriologist named Joseph Murray, whether the company had already known. She came to believe it had, and had not said so; on May 25 she wrote telling him as much. The nerve reports raised a further question nobody at the company had thought to ask: what nine months of exposure might do to a fetus. Merrell had tested the drug in pregnancy only briefly, late in term, for the mother's comfort. Nothing covered the length of exposure that mattered.
For about a year the company's representatives phoned, visited, and finally, in September 1961, convened a conference of its own investigators to argue the FDA was obstructing a good drug. Pressed there on whether it was safe in pregnancy, they found they had nothing to say. Two months later, on November 30, 1961, Merrell told the FDA that the German manufacturer was pulling the drug after reports linking it to birth defects. The nerve damage had already given Kelsey real evidence of harm, not merely an absence of proof; what November added was the far graver finding that the same drug could also deform a fetus. By then more than a thousand American physicians had been supplied with it to give to patients who, in many cases, had not been told it was still unapproved.
Investigators eventually traced ten documented cases of the same limb deformities to patients in the United States, and at least seven more to Americans who had received the drug overseas: at least seventeen children harmed by a medicine that, on paper, had never been approved for sale. In the dozens of countries where it had gone on the market outright, thousands of children were born with the same deformities.
Kelsey found no defect herself. She refused, for eighteen months, to treat an unproven claim as a proven one, against a company that kept insisting the absence of bad news was itself the evidence. The Kefauver-Harris Amendments of 1962 wrote her habit into federal law: proof that a drug worked as well as proof it was safe, and, with narrow exceptions for cases an investigator judged infeasible, a requirement that patients be told a drug was still being tested and asked for their consent. A clock built to approve a drug by default can only be beaten by someone willing to keep saying no.