Good Calories, Bad Calories

Gary Taubes is not a doctor or a nutritionist. He is a science journalist with a physics degree who spent much of his career investigating errors and fraud inside science itself, including the cold fusion episode. When he decided to investigate why the world was getting fatter, he used the same method: he didn't ask what everyone kept repeating about diet, but what the original studies actually showed.

The book grew out of a 2002 article in Science magazine, which had already caused discomfort by questioning the consensus against fat. Good Calories, Bad Calories is the expanded version of that investigation: more than five hundred pages reviewing decades of nutrition research, the way an investigative reporter audits a case, not the way a professor summarizes a subject.

Taubes is also well known for Why We Get Fat, his next book, which boils the same ideas down to a shorter, more direct version. Good Calories, Bad Calories is the raw material behind that summary: denser, more historical, with the full name of every scientist involved in the dispute.

Ancel Keys was an influential American physiologist, comfortable moving between research and public health policy. In the 1950s, he proposed a chain of cause and effect that seemed logical: saturated fat raises blood cholesterol, high cholesterol clogs arteries, clogged arteries cause heart attacks. It became known as the diet-heart hypothesis.

The proposal arrived at a time when the United States was having a real scare over rising deaths from heart disease, including among middle-aged people with no prior history. Keys had a strong presence on committees and at conferences, the kind of scientist who wins over a room before all the data is in.

Taubes doesn't dispute that Keys was serious. His point is different: the hypothesis gained institutional ground at a speed that didn't match the amount of evidence available at the time, and that would have consequences decades later.

In 1953, Ancel Keys presented a chart of six countries linking fat consumption and death from heart disease in an almost perfect line. Years later, he launched the Seven Countries Study, which followed men in several nations for decades and became one of the pillars of the consensus against fat.

Taubes revisits a criticism made in 1957 by researchers Yerushalmy and Hilleboe: data already existed for about twenty-two countries, and with all of them on the chart the relationship was much weaker. He also questions the choice of countries in the later study. Keys's defenders respond that the Seven Countries Study was a different kind of research, with its own measurements, and that the twenty-two-country criticism applies to the 1953 chart, not to it.

The lesson is about the trap of looking only at the points that confirm the line. A well-chosen correlation is persuasive, but it doesn't prove that one thing causes the other, and that applies to both the defenders and the critics of fat.