Making lifestyle choices with imperfect knowledge about what is healthy

Gary Taubes ends his opinion piece in today's New York Times with a rhetorical question:

Before I, for one, make another dietary resolution, I’d like to know that what I believe I know about a healthy diet is really so. Is that too much to ask?

after having spent the article presenting a persuasive case for the conclusion that "Yes, it is too much to ask."  His key point is that

Because the hypotheses [in nutrition] are ultimately about what happens to us over decades, meaningful trials are prohibitively expensive and exceedingly difficult. 

When I was interviewed recently about intermittent fasting, I explained that we don't really know whether it's healthy or not, but I follow it because the idea that it promotes health makes a lot of sense to me, given what I do know about the underlying science. Gary Taubes makes his own decisions. ("My own vote [for what most of us are eating too much of] sugars and refined grains; we all have our biases.") I don't think he's actually biased. His "vote" is based on a career of looking into the matter. It is not so much a bias as an informed guess.

As a scientist, I am aware that much of what we spend our time thinking about is not proven. In fact, I have argued that nothing in science is completely proven (outside of mathematics, and even within mathematics, a lot of time is spent considering the implications of conjectures and lemmas). There are cases, especially in the physical sciences, when residual doubt is absurdly insignificant, and there are cases, including nutritional science, when there is a lot of doubt.

When policy is developed based on science, it has to be based on the best available evidence and the best possible estimate of how likely it is that the prevailing wisdom is wrong.  If appropriate, we want confidence intervals. It's not always true that when there is smoke there is fire, but when there is smoke, you should definitely stop what you're doing and have a look (or get out).

In any case, I very much enjoyed this presentation of what we don't know, and I recommend it (here). One of the better bits:

Since the 1960s, nutrition science has been dominated by two conflicting observations. One is that we know how to eat healthy and maintain a healthy weight. The other is that the rapidly increasing rates of obesity and diabetes suggest that something about the conventional thinking is simply wrong.

I want to soundly endorse his conclusion:

Making meaningful inroads against obesity and diabetes on a population level requires that we know how to treat and prevent it on an individual level. We’re going to have to stop believing we know the answer, and challenge ourselves to come up with trials that do a better job of testing our beliefs.