Frances Oldham Kelsey

Frances Oldham Kelsey did not prove that thalidomide caused birth defects before blocking it. She noticed that its manufacturer had not proved the drug safe. In September 1960, one incomplete application landed between the United States and a disaster already developing across Europe.

Kelsey was a Canadian-American pharmacologist and physician, 1914–2015. She joined the US Food and Drug Administration on August 1, 1960; thalidomide was the second application assigned to her.

The 60-day clock

Under the 1938 law, an application became effective after 60 days unless the FDA formally found it incomplete. The Kevadon application occupied four volumes and was reviewed by Kelsey, chemist Lee Geismar, and pharmacologist Jiro Oyama. All three found deficiencies.

The clinical reports read more like endorsements than controlled studies. A December 31, 1960 letter by Leslie Florence in the British Medical Journal then described severe peripheral neuropathy among long-term thalidomide users. Kelsey asked why a supposedly mild sleeping pill could damage adult nerves and what prolonged exposure might do during pregnancy.

Each incomplete notice bought another 60 days. The company wanted the drug available for the Christmas sedative market; Kelsey kept asking for case histories, animal evidence, and an accounting of neurological toxicity. On November 30, 1961, the manufacturer told the FDA that thalidomide was being withdrawn in Germany after reports of congenital malformations. The US application was formally withdrawn on March 8, 1962.

She had watched missing tests kill before

In 1937, Kelsey was a graduate student in Eugene Geiling’s University of Chicago laboratory when the FDA asked the group to investigate Elixir Sulfanilamide. Its manufacturer had dissolved the drug in diethylene glycol without toxicity testing.

Kelsey monitored rats while the laboratory separated the ingredients into experimental groups. The animals receiving diethylene glycol developed kidney failure. That episode helped drive the 1938 law under which she later reviewed thalidomide.

Her decision in 1960 was therefore not a flash of intuition. It was concept first principles applied to a familiar failure shape: separate the claim from the evidence, locate what has not been tested, and refuse to treat absence of reported harm as proof of safety.

What changed

On August 7, 1962, John F. Kennedy presented Kelsey with the President’s Award for Distinguished Federal Civilian Service. She was the second woman to receive it.

The larger artifact arrived on October 10: the Kefauver-Harris Drug Amendments. Manufacturers now had to establish effectiveness as well as safety, clinical investigators needed informed consent, and adverse reactions had to be reported. Kelsey later led the FDA division that inspected clinical investigators and the evidence behind their submissions.

One reviewer delayed one application. Congress converted that delay into a rule.

What the hero story misses

Kelsey deserves the credit, but the lone-gatekeeper version is too tidy. Geismar and Oyama found their own deficiencies, FDA supervisors supported the repeated incomplete notices, and clinicians Leslie Florence, Widukind Lenz, and William McBride supplied signals from three countries.

The United States also escaped mass marketing, not all exposure. Thalidomide samples had already reached American doctors and patients under the loose “investigational” rules of 1960. The episode is evidence for institutional scrutiny, not proof that institutions always work.

The disputed question is how much authority a reviewer should have when harm is uncertain. Too little authority turns a deadline into approval by inertia; too much can delay a useful medicine. That is a concept bayesian inference problem with unequal error costs: a false approval and a false rejection do not injure the same number of people in the same way.

A cross-realm test

Kelsey and person feynman used different objects to expose weak institutional reasoning. Feynman placed Challenger O-ring material in ice water in 1986; Kelsey placed a four-volume drug application against a 60-day clock in 1960. Both made concept falsifiability operational by asking what evidence would make an official claim fail.

An open question

How do I design a review system that preserves Kelsey’s right to say “the evidence is not here” without requiring the next Kelsey to be unusually stubborn?

Key Sources

Frances Oldham Kelsey, Autobiographical Reflections (FDA History Office, compiled from interviews conducted in 1974, 1991, and 1992): her account of the application, the review team, and the 60-day procedure.
A. Leslie Florence, “Is Thalidomide to Blame?” British Medical Journal 2 (1960), 1954, DOI: 10.1136/bmj.2.5217.1954: the early report of peripheral neuropathy.
W. G. McBride, “Thalidomide and Congenital Abnormalities,” The Lancet 2 (1961), 1358: the clinical report connecting thalidomide with congenital malformations.
Cheryl Krasnick Warsh, Frances Oldham Kelsey, the FDA, and the Battle Against Thalidomide (Oxford University Press, 2024): a book-length account of Kelsey’s career and the regulatory fight.
US Food and Drug Administration, “Frances Oldham Kelsey: Medical Reviewer Famous for Averting a Public Health Tragedy”: the agency’s historical record of her work.

Abhishek's take

Kelsey’s useful move was not predicting the precise birth defect. She refused to let a 60-day clock convert missing evidence into permission. I keep returning to that distinction when I build review gates: uncertainty should trigger a named question, not disappear inside a deadline.

Tags: #drug-safety #regulatory-science #thalidomide #evidence #institutional-design

Abhishek