[This post is an unfinished working draft, but I'm presenting it for the benefit of Complex Systems listeners / readers who want to read a bit more about the FDA history we discussed on the pod.]

If you spend any time at all talking in public about the potential benefits of reforming the Food and Drug Administration to speed the approval of new drugs, some interlocutor will pop up and argue that the story of thalidomide proves that such a reform is too risky to contemplate. (Ask me how I know...) However, this argument too often ignores key parts of the story, and even more importantly misunderstands the subsequent FDA reforms that reacted to thalidomide and led to significant changes in the system of drug regulation we see today.

The Pure Food and Drug Act

"There is a probability of increased interest during the next few years in methods of testing the genuineness of all articles of food, beverages, and medicines," reported the editors of the magazine Scientific American in January 1882. "Comprehensive laws of this kind have been passed within a year or two in Wisconsin and New York. [...] Indiana required analysis of all fertilizers in market[...]; Maine, of vinegar; Massachusetts, Nevada, and New Jersey, of milk; Ohio, of milk, butter, cheese, and meat, and of fertilizers."

The magazine's prediction proved prescient, as in 1902 the Bureau of Chemistry's Chief Chemist, Dr. Harvey Washington Wiley, began a highly-publicized investigation into toxic food preservatives. His "poison squad" trials recruited volunteers for $5 a month and free food to be fed increasing doses of preservatives including borax and formaldehyde. Multiple times, trials had to be stopped when participants became too sick to continue.

But it was the 1905 publication of Upton Sinclair's The Jungle which most obviously prompted a breakthrough in Dr. Wiley's lobbying for a federal law banning specific additives and requiring other ingredients to be accurately labeled. "Dr. Wiley's Law", as it would later be dubbed by the press, was passed in 1906, officially as the Pure Food and Drug Act. By 1907, Wiley's much-expanded Bureau of Chemistry added formaldehyde as the first entry on the list of banned additives.

In 1909, as his crusade against dangerous chemicals continued, Wiley would order the seizure and inspection of forty barrels and twenty kegs of Coca-Cola. His Bureau brought a federal suit aiming to force the company to remove caffeine from its product, on Wiley's theory that caffeine was a toxic and dangerous additive. The resulting legal affair gave us the gloriously-named federal case United States v. Forty Barrels and Twenty Kegs of Coca-Cola, which would be argued before the Supreme Court in 1916. In the end, the court did not determine whether caffeine was toxic or non-toxic, Wiley was fired from his post, and Coca-Cola voluntarily agreed to reduce the amount of caffeine in their drinks.

Later expansions of the PFDA would require manufacturers of beverages or medicines to label their products with the dosage of drugs such as caffeine, alcohol, cannabis, cocaine, heroin, or morphine, bringing transparency to the contents of then-popular "patent medicines". While specific additives would remained banned, the focus was placed mainly on accurate labeling.

Elixir sulfanilamide

On this scene, in 1937, the S. E. Massengill Company introduced and sold 633 bottles of "elixir sulfanilamide" which was (pure and unadulterated) sulfanilamide and raspberry flavor dissolved in diethylene glycol (DEG).

Sulfanilamide is a reasonable antibiotic, and raspberries can be delicious, but diethylene glycol is an industrial solvent and antifreeze that's a deadly poison to humans, rats, and basically anything else with a kidney. More than 100 deaths were recorded (which is certainly an under-estimate), with the first reports reaching the American Medical Association and the FDA about one month after the drug's commercial release. DEG's toxicity and specific role in the elixir sulfanilamide poisoning was confirmed by FDA consultant Prof. Eugene Geiling and his 23-year-old graduate student Frances Oldham, who we'll meet again later.

According to an FDA Consumer article from 1981, "Practically the entire field force of 239 FDA inspectors and chemists was assigned" to recall and retrieve the drug. The article includes this charming 1930s anecdote:

One East St. Louis woman told an inspector she had destroyed the drug. The inspector persisted in his questions however. What did she mean "destroy"? Had she poured it down the sink? Had she buried it? No, the woman said, she had thrown the bottle out the window into an alley. The inspector found the bottle still unbroken, still containing enough elixir to kill any child intrigued enough to swallow its sweet, raspberry-flavored contents.

Samuel Massengill would later testify in court:

My chemists and I deeply regret the fatal results, but there was no error in the manufacture of the product. We have been supplying a legitimate professional demand and not once could have foreseen the unlooked-for results. I do not feel that there was any responsibility on our part.

Massengill Co.'s chief chemist didn't quite agree with his boss, though; he committed suicide while awaiting trial.

In the end, Massengill Co. was fined either $16,800 or $26,100 (reports differ) under the 1906 Pure Food and Drugs Act for mislabeling the drug "elixir" when it did not contain the required ethanol. Civil suits recovered at least an additional $150,000 in damages, or $3.3 million in 2023 dollars, from the company.

A legislative update to the PFDA regulations had been in the works for a while:

By the 1930s it was widely recognized that the Food and Drugs Act of 1906 was obsolete, but bitter disagreement arose as to what should replace it. By 1937 most of the arguments had been resolved but Congressional action was stalled.

The Massengill affair naturally made the legislation a top political priority, and shortly after Massengill Co. paid their mislabeling fine, Congress passed the 1938 Federal Food, Drug, and Cosmetic Act (FFDCA).

Interlude: What's in the FFDCA?

For the first time, the FFDCA required drugs and food additives to be tested for safety, not just purity. The FDA's chief of the Investigational Drug Branch described the act in 1965:

The revised version required that before a new drug could be distributed in interstate commerce the sponsor of the drug, must submit information to the Food and Drug Administration to establish the safety of the drug for the proposed clinical use or uses. This information, submitted as a New Drug Application or NDA, consisted of details concerning the quality control of the compound, animal studies undertaken to establish its acute and chronic toxicity and pharmacological effects, and clinical studies designed to demonstrate its safety.

Under the implementing regulations, a new drug could however be distributed for investigational purposes prior to the submission of a NDA provided that the preparation was labeled clearly as an investigational drug, that the investigators signed a statement to the effect that they had facilities to test the drug and that the drug would be tested under their supervision only, and that accurate records of the distribution of the investigational drug would be kept by the firm. There was no requirement that the government be notified that the drug was under investigation or that an investigation was discontinued when this occurred.

Once an NDA was submitted, the FDA had 60 days in which they could notify the company that the information was "incomplete", requiring the company to revise, resubmit, and restart the 60-day clock.

Contergan and Kevadon

In November 1956, the German drugmaker Chemie Grünenthal released a proprietary drug combination for treating the symptoms of respiratory infections that mixed a new sedative called thalidomide, quinine, vitamin C, and aspirin. In 1957, they released thalidomide as a standalone sedative and nausea reducer under the name "Contergan". The drug would become infamous for its use to promote sleep among pregnant women with morning sickness (where both the sedative and anti-nausea effects would be helpful), but it was also more generally promoted for insomnia, coughs, colds, and headaches. Grünenthal licensed the drug to Cincinnati-based drug manufacturer Richardson-Merrell, who distributed investigational samples and submitted an FDA NDA in September 1960 to sell thalidomide in the US as "Kevadon".

Dr. Frances Kelsey (née Oldham), who had worked on elixir sulfanilamide twenty-three years prior, rejected Richardson-Merrell's application. Later, in a retrospective article in the American Journal of Public Health, she explained her reasoning as:

In particular, although this drug appeared to be remarkably nontoxic in animals and human beings, little or no information was available concerning its absorption, distribution in the body, or its excretion. Since the possibility existed that the low toxicity of the drug in certain species might be related to poor absorption in those species, and that under certain conditions the absorption in other species might be increased, further work was requested relative to the metabolism of the drug.

In modern terms, we'd call this "pharmacokinetics", though the term doesn't appear in her AJPH article.

In fact, Kelsey's academic work at the University of Chicago in 1943 had included an investigation of a candidate antimalarial drug that was metabolized less rapidly in pregnant and fetal rabbits than in nonpregnant adults (leading to buildup of circulating dose). However, in a forty-year retrospective, she would claim that the assignment was a coincidence and not tied to her previous academic work: "They gave it to me because they thought it would be an easy one to start on. As it turned out, it wasn't all that easy."

Kelsey rejected Richardson-Merrell's resubmission ("The clinical reports were more on the nature of testimonials, rather than the results of well-designed, well-executed studies."), and in December 1960 during their following resubmission attempt, the British Medical Journal published a letter with case studies of peripheral neuropathy in (non-pregnant) chronic thalidomide users. Kelsey would reject the next NDA in May 1961 on this basis, writing after the fact that "In the absence of more comprehensive data, we felt the new drug application for thalidomide should not be approved unless it was made clear in the physician's brochure that the safety of the drug in pregnancy had not been established."

The second half of 1961 saw the British investigations into the adult neuropathy and German investigation into a "sudden alarming increase of congenital limb deformities" both reach conclusions blamning the drug. "[B]y the end of November, 1961," wrote Kelsey, "word was received from Germany of the possible association of this drug with [birth defects] and the application became inactive." The company withdrew the application permanently in March 1962.

The Kefauver–Harris Amendment

[Section to be expanded - Sen. Estes Kefauver had been pushing for greater regulation of pharma companies (he claimed) keeping prices high by re-releasing drugs on new patents that didn't improve their efficacy. He had been working on the agenda for many years when the thalidomide scandal broke, and he would be successful in passing an amendment to the FFDCA in 1962, just months after the news broke. The amendment itself, however, does nothing to require the kinds of pharmacokinetic testing that Frances Kelsey insisted on. Rather, it follows Kefauver's agenda by requiring pharma companies to demonstrate the efficacy of their products to the satisfaction of the FDA, basically creating the modern Phase-3 trial (and the Phase-2 trial, which is mostly there to derisk the expense of the Phase 3). Pre-1960 safety review matured into today's Phase-1 trials, by adopting Kelsey's approach as well as other improvements, mainly as a series of bureaucratic decisions within the FDA, rather than being driven by statute. Though a modern Phase-1 trial can cost upwards of $10mln if performed in the US and sparing no expense, I'd argue that those requirements are just clearly good for society.]

Discussion

The received wisdom is that thalidomide was an unmitigated disaster in the United States, and that we need a strong FDA to prevent another such disaster. As a historical lesson, though, this is clearly misplaced. An FDA reviewer, doing her job in the pre-Kefauver–Harris regime, rejected thalidomide on the grounds of safety concerns, and in the face of significant commercial lobbying. She was right, and thalidomide wasn't sold in the US.

But there are right lessons to learn from this history.

The economist Milton Friedman, speaking about his own field, once stated that "Only a crisis - actual or perceived - produces real change." So it is in drug regulation; we have a regulatory system built by crises. The Jungle prompted the Pure Food and Drug Act. The elixir sulfanilamide poisonings directly drove the FFDCA. Outcry against the European regulatory failure of thalidomide spurred passage of the Kefauver–Harris Amendment.

But Friedman's quote continues: "When that crisis occurs, the actions that are taken depend on the ideas that are lying around." Again, we see just that in the FDA. Precursors to the PFDA were being discussed in the 1880s, but in 1905 The Jungle catalyzed a leglislative response within the year. The PFDA was recognized as inadequate in the 1920s and ’30s, but the 1937 elixir sulfanilamide affair drove the FFDCA, within about a year. Estes Kefauver had started grilling big-pharma execs about excessive products and unjustified efficacy claims in 1959, but his legislative push was floundering until 1961's thalidomide scandal drove the 1962 legislation that would bear his name.

So the system we have in 2024 has a great deal of path-dependency on the ideas that were lying around in 1905, and 1937, and 1961. Why does the Kefauver–Harris Amendment include a requirement that pharma companies have their drugs approved for efficacy before they can sell them? Not because that had anything to do with the thalidomide affair, but because Senator Kefauver had for years been pushing for a crackdown on over-hyped, expensive big-pharma drugs.

Conclusion

In certain circles, a robust, two-sided debate is simmering about whether and how to reform the FDA. Many argue that slower, more expensive processes cost lives -- who pass unnoticed into an "invisible graveyard". I'm generally sympathetic, and I think that great social gains could be unlocked by speeding the adoption of drugs that reach patients too slowly -- or never at all.

But focusing on finding the single best policy is chasing the wrong goal. Government policies on a hyper-sensitive issue like medical regulation are unstable, lurching from reaction to reaction as the public eye turns towards them in the immediate wake of some new crisis. The question that would-be reformers should ask is not what policy would save the most lives, but instead what policies can create a stable system that saves lives while weathering the inevitable political shocks. Getting this wrong can lead to policies that overreact to what should be narrow issues: Frances Kelsey's heroic work on thalidomide should have inspired the FDA to spread her insights and precision to all reviewers' work, but instead the leglislative response centered an efficacy-testing mandate that wouldn't have helped keep Kevadon from American patients.

Even among the reformist camp, few would argue that we should abandon modern pharmacokinetic testing. Perhaps if it had been standard in the 1950s, the FDA would have been recognized as an example of regulation done right, rather than a system in need of reform. Reformers should look for policies that even a scandal-inflamed public can see as reliable

It's easy enough to agree that the FDA's drug-approval procedures aren't perfect, and to debate what should be changed. Perhaps patient advocates should have a greater role in the process. Perhaps approval should have more the nature of a cost-benefit analysis. Perhaps patients should have the right to try experimental drugs that have been proven safe but don't have conclusive advocacy data, so long as all the facts are disclosed. Perhaps.

But focusing on finding the single best policy is chasing the wrong goal. The government is ultimately an instrument of the public will, and on sensitive issues like medical regulation, no policy will be stable for long if it can't weather public outcry in a crisis. And crises and scandals will occur every few decades; they will either slip through the cracks of the last incomplete response, they'll arise from regulation gone too far, or they'll happen in another country entirely but Americans will fear our near miss all the same. When they do, the system we have will be examined and found wanting in some way. Reforms will fly through the legislative process, but they won't be written from scratch -- they'll be based on "the ideas that are lying around".

If you want a reform -- like giving patients the chance to try drugs that are proven safe before their efficacy is proven -- your reform will need to be able to weather the next round of rapid-reform-using-whatever-is-lying-around.

On the other side of the coin is that whatever-is-lying-around is the way that improvements become changes to policy. To crib a quote from the Complex Systems podcast (specifically, guest Dave Guarino):

[P]olicy has a very odd shape: nothing’s possible for 20 years, everything’s possible for six months, and then nothing’s possible again for another 20 years. It’s like healthcare policy—people spent 20 years building up to the Affordable Care Act, then there was six to twelve months of horse-trading, and suddenly every idea and white paper was relevant. Then the window closed, and we wait another 20 years. (...)

If there's something that you think should be different about the FDA, it's not enough to be right or win a debate. No, the winning strategy seems to be what worked for Dr. Wiley and Senator Kefauver -- do what you can while your prospects for immediate success are limited, so that your ideas are closest to hand when something needs to happen fast. I feel like I don't see enough of this approach from the compatriots and fellow-travelers of mine who care deeply about how the FDA regulates drug development, but that's an argument for another day. (And another topic for another day is how the FDA's regulatory stance was loosened by two punctuated crises, the global pandemic and then the global pandemic.)