Nutritional signal and niose

From An Apple a Day, and Other Myths by George Johnson.

But there is a yawning divide between this nutritional folklore and science. During the last two decades the connection between the foods we eat and the cellular anarchy called cancer has been unraveling string by string.

[snip]

About all that can be said with any assurance is that controlling obesity is important, as it also is for heart disease, Type 2 diabetes, hypertension, stroke and other threats to life. Avoiding an excess of alcohol has clear benefits. But unless a person is seriously malnourished, the influence of specific foods is so weak that the signal is easily swamped by noise.

The situation seemed very different in 1997, when the World Cancer Research Fund and the American Institute for Cancer Research published a report, thick as a phone book, concluding that diets loaded with fruits and vegetables might reduce the overall incidence of cancer by more than 20 percent.

After reviewing more than 4,000 studies, the authors were persuaded that green vegetables helped ward off lung and stomach cancer. Colon and thyroid cancer might be avoided with broccoli, cabbage and brussels sprouts. Onions, tomatoes, garlic, carrots and citrus fruits all seemed to play important roles.

In 2007, a major follow-up all but reversed the findings. While some kinds of produce might have subtle benefits, the authors concluded, “in no case now is the evidence of protection judged to be convincing.”

The reason for the change was more thorough epidemiology. The earlier studies tended to be “retrospective,” relying on people to remember dietary details from the distant past. These results were often upended by “prospective” protocols, in which the health of large populations was followed in real time.

[snip]

With even the most rigorous studies, it is hard to adjust for what epidemiologists call confounding factors: Assiduous eaters of fruits and vegetables probably weigh less, exercise more often and are vigilant about their health in other ways. Some of this can be sorted out with randomized controlled trials, with two large groups of people arbitrarily assigned different diets. But such studies are expensive, and the rules are hard to enforce in the short term — and probably impossible over the many years it can take for cancer to develop.

All quite interesting for being said clearly in one place rather than accreting the observation in dribs and drabs.

Macroeconomics, nutrition and health, personal finance, education, life decisions, climatology; in all these fields, much of the yield of research of the past fifty years has not been simple answers to complex questions. It has been complex and uncertain answers to complex questions. Indeed, in many ways, the primary product of fifty years of research is 1) an increasing comprehension of just how complex the questions really are and 2) a dawning recognition that some of the questions cannot in a meaningful way be answered with precision in a timely fashion owing to the sheer plentitude of confounding factors and the smallness of the effect. In other words, there is a signal but it is very weak and usually drowned out by the system noise. By the time you are able, at great expense, to demonstrate that the signal is indeed real, it is either so small that the effort is not warranted or the system has changed so much that there is justifiable concern that the effect can no longer be replicated under current conditions. For example, while there are lessons to be learned, a developing country wanting to become rich cannot simply execute the Japanese strategies of 1950-60. The world has changed and what worked then is unlikely to yield the same results now.

All the investment in research is not wasted. It sets the parameters for future, more productive research. In all these fields there are also very useful heuristics. In food and nutrition, instead of a litany of particular foods to be eaten under particular circumstances, prepared in particular ways, and for particular anticipated benefits, there is now the luxury of a simple heuristic that works equally well: Eat in moderation and with variation, supplemented by moderate routine exercise and occasional extreme exercise.