10 The Journalist Gary Taubes 10: Anomaly Hunter 4
Gary Taubes has selected a rather trivial paper with which to discredit Ancel Keys. To set it up, he wrote, “Once having adopted firm convictions about the dangers of dietary fat based on his own limited research among small populations around the world, Keys repeatedly preached against the temptation to adopt any firm contrary convictions based on the many other studies of small populations that seemed to repudiate his hypothesis.” Repudiate his hypothesis? This is not a very good sentence but his point is that Keys based his ideas entirely on just a few studies of select communities here and there. There were other little papers about other communities that he didn’t like so he ignored them, but Taubes thinks those should have mattered, too. Taubes apparently believes that all lipid science at this time was based on anomaly hunting by epidemiologists. Actually, Keys drew on a wealth of research, including many metabolic ward studies, and those weren’t all focused on fat and he wasn’t alone. He was one researcher among many. He just happened to be the best. The anomaly hunter here pretends that a paper about a small town called Roseto should have caused Keys to disregard all that other research and change his mind about the etiology of heart disease. Instead, Keys said Roseto didn’t change much.
This is the Roseto paper Taubes referenced and its abstract from 1964. There were no deaths in this little town from myocardial infarction for anyone under the age of 47 during this study. Yes, I said 47. This isn’t exactly the China Study here.
Roseto’s mortality and heart disease data were compared to those of four nearby towns, Nazareth, Bangor, Stroudsburg, and East Stroudsburg. Roseto had 1630 residents. The other towns had no less than 5766. Roseto was at least three and a half times smaller than the next smallest town in the study.
You can see that while Roseto did indeed have lower rates of heart attacks, they had a roughly similar rate of arteriosclerotic heart disease mortality. These people were not free of heart disease. Also, realize the numbers you see are extrapolated rates per 100,000 people, which would tend to amplify any faults in their sampling.
We are given only the most general dietary information, although we are also given anecdotal accounts of their use of prosciutto and lard. Gary Taubes likes anecdotal evidence so he did mention that, but he forgot to mention the next sentence. They were overweight. Ignore the awkward, right, Mr Taubes? Why not also say that ham and lard made these people fat?
Roseto became well-known for this paragraph. The people there seemed to have a very healthy social atmosphere. Roseto became known as a place where happiness made people a bit healthier.
Here is a newspaper article which makes that clear. One of the study collaborators, Stewart Wolf, is quoted. He would come to be the man most identified with the Roseto story. Wolf said that heart disease likely had many contributory factors. He didn’t pretend he had simplified heart disease and overturned the rest of the scientific literature.
In this article, he is more clear that he thought the Roseto phenomenon was chiefly about a relative lack of psychosocial stress. He did not defend lard.
Ancel Keys did have a say on Roseto. He made the point that “those of us who emphasize the importance of diet have not claimed that dietary fat is the only etiological factor and, conceivably, the influence of the diet could be overcome by other influences.” Mr Taubes, to characterize Ancel Keys as holding any other position is a misrepresentation of his position. It is a straw man. Keys also wondered if the population there had been stable, if the age matching was appropriate for the other towns, and if they had better medical care there. Moreover, the conditions in Roseto over time had not been measured, and even if they were to be, the town was so small that such data might not be very useful.
Stewart Wolf responded to Keys. He said, and Mr Taubes, pay attention, “no claim was made that diet is unimportant in the pathogenesis of coronary atherosclerosis.” Pause this and read this as many times as necessary, Mr Taubes. This means that even one of the authors of this study wouldn’t agree with your conclusions about it. Wolf said research in Roseto would be ongoing, and it was.
Wolf published more research on Roseto in 1974. Here he directly identified a closely knit social pattern as the unique feature of Rosetan life. Does Taubes think we should find it surprising that happier and less stressed people have fewer heart attacks?
Wolf also noted that they did have a lower smoking rate in Roseto than the other communities used for comparison in the original study. Perhaps smoking behavior is important for heart disease risk, Mr Taubes. Will you concede that? They also had a lower rate of diabetes in Rosetto than in Nazareth.
In 1992, Wolf studied Roseto one more time. The observed lower death rate in Roseto had been real, but their mortality rates eventually did rise as their happy way of life began to change.
Taubes wants you to think that nutritional epidemiology is a “pseudoscience at best.” He says that observations of free-living people don’t count for much but controlled experiments do. Yet he gives us a list of weak observational studies in his book, some altogether lacking in dietary data, and ignores tightly controlled metabolic ward experiments. He isn’t even interested in appearing to have consistent ideas and principles. His blatant hypocrisy is remarkable.
Now believe me, I do understand why he wants to blow off the whole concept of nutritional epidemiology. It produces inconvenient findings like this all the time. Ancel Keys may have “explained” away a study or two but at least he didn’t have to try to explain away a whole method of scientific investigation.
At the end of my look at these scattered quirky studies Taubes thinks are so significant, I want to show you an example of a paper that he should have taken more seriously. Do you see where he mentions men living for a year at a research station in Antarctica? This one was really interesting and he didn’t even try to explain it away. Unlike all those unsophisticated cross-sectional studies with their weak methods that he seems to like so much, this was a serious contribution to diet-heart research when it was published, and it is still has value for us today.
In this study conducted in Antarctica, researchers had unusually well-managed conditions. Young, healthy men were holed up, isolated in a research station in the freezing cold down there for an entire year. Therefore there was a minimum of variables beyond their control, which provided a great opportunity for nutrition research. These were not free-living obese and illiterate women in Trinidad.
Fat consumption was shown to raise cholesterol in the year of their study. The fat they ate was highly saturated.
Their unique situation allowed them to eliminate any effects of intra-individual variation. They took multiple blood tests to make sure the trends they saw were real. This is an example of a study that didn’t suffer from the usual faults. It’s not easy to criticize. That’s why you don’t hear about it much.
Mr Taubes, all studies aren’t of equal quality. The proponents of diet-heart like Ancel Keys evaluated the studies they saw critically and intelligently. The best research supported their views. In fact, the best research formed their views in the first place. They did not do what you do. They did not just list superficially contradictory studies of dubious merit and then simply declare the whole matter a wash. They sought to clarify. You attempt to obfuscate.
In my next videos, I will take a closer look at the major diet-heart intervention trials, including the one that Mr Taubes thinks should have ended the diet-heart idea all by itself.