The biggest attempt to create a truly blind trial, in which subjects would switch to a vegetable-oil-based diet without knowing it, was conducted by the National Heart, Lung, and Blood Institute with Jerry Stamler as one of the principal investigators. NHLBI was aware of the ongoing problems with the diet trials. It was clear that only an enormous, well-controlled clinical trial could definitively establish the link between saturated fat and heart disease. Such a trial would need to enroll a hundred thousand Americans in order to get statistically significant results and would need a follow-up period of forty-five years. To see if such a giant undertaking was even possible, NHLBI first conducted a feasibility study in 1962. This in itself was a giant effort involving multistep studies on nearly twelve hundred subjects in five different cities, including Baltimore, Boston, Chicago, the Minnesota Twin Cities, and Oakland, as well as a mental hospital in Minnesota.

Coincidentally, the oversight for these studies fell to those most invested in their outcome: Keys and Stamler. Stamler remembers walking the streets in New York City “all night long” with Keys, debating how they might set up the study so people could be “blinded” to the food they ate. Eventually, they came up with a solution that satisfied them:

the food company Swift & Co. would make custom margarines with varying levels of fatty acids that both groups would eat; butter would therefore be a non-issue. Even so, the undertaking remained daunting, because other special foods also had to be made for all diet groups, in order to assure that the taste, texture, and cooking experience for all participants would be the same. Hamburger patties and hot dogs were therefore made in two versions: one high in vegetable oil and one made with tallow or lard. Milk and cheese for the intervention group came “filled” with soybean oil. (No one could figure out how to make a simulacrum of an egg, however, so everyone just got two normal eggs a week.) “A housewife would order once a week from a special store that had been set up for the study and was sent the proper foods assigned to her group,” says Stamler. Neither the participants nor the study administrators knew who was getting which diet, in an attempt to “double-blind” the study, which was a milestone in diet-heart research. No one had ever managed to do this before, and according to various confirmation tests performed by the investigators, their methods were largely successful: “No one noticed who was getting which types of food! It was all done so well,” Stamler asserted.

In retrospect, it is perplexing why scientists did not question the assumption that entirely newfangled foodstuffs could restore a population to good health. How could it be that a healthy diet would depend upon these just-invented foods, such as milk “filled” with soybean oil?

It’s true that vegetable oils had been shown to lower total cholesterol successfully, and this effect held great appeal to a research community obsessed with cholesterol. Yet cholesterol-lowering was just one of the many effects of these oils on biological processes, not all of which seemed to be so beneficial. In fact, no human population had been documented surviving long-term on oils as a major source of fat until 1976, when researchers studied the Israelis, who at the time consumed “the highest reported” quantity of vegetable oils in the world. Their rates of heart disease turned out to be relatively high, however, contradicting the belief that vegetable oils were protective.

When I asked Stamler about the novelty of vegetable oils he said that he and Keys had been concerned about the absence of any historical record for human consumption of these oils, but that ultimately it wasn’t considered an impediment to promoting a “prudent” diet.