Eat This, Never Eat That, Live Forever. Really?

Apparently, the secret to living to 100 has been hiding in my refrigerator.

Or perhaps my coffee cup.

Depending on what I read this week, blueberries will protect my brain, coffee will extend my life, walnuts will save my heart, olive oil will ward off dementia, and dark chocolate will perform some other minor miracle.

Unfortunately, by next week one of them may be trying to kill me.

We've all seen the headlines:

Eat This One Food Every Day to Live Longer.

This Common Ingredient May Increase Your Risk of Dementia.

The Drink That Could Add Years to Your Life.

The Food Cardiologists Want You to Stop Eating.

And my favorite variety: Scientists Have Discovered...

Usually followed by something human beings have been eating for 4,000 years.

I've spent about 50 years in and around media and marketing, so perhaps I'm more ... well, a LOT more cynical about these headlines than most people. I know what a headline is supposed to do. It is supposed to make you stop. And these days, above all, it is supposed to make you click.

“Eating a varied diet containing fruits, vegetables, legumes, whole grains and nuts is associated with better long-term health” may be reasonably accurate.

It's also a terrible headline.

“THIS ONE FRUIT MAY HELP YOU LIVE TO 100” is a much better headline, guaranteed to grab attention. And therein lies the problem.

Carl Sagan and the Blueberry

Carl Sagan famously said: “Extraordinary claims require extraordinary evidence.”

I think about that line surprisingly often when reading nutrition news.

If someone tells me blueberries contain fiber, vitamins and phytochemicals and are a nutritious food, I don't consider that an extraordinary claim.

If someone tells me eating blueberries will prevent dementia, that's a different matter.

That's an extraordinary claim. Show me the extraordinary evidence.

Not one study. Not a study in mice. Not a questionnaire asking 600 people what they remember eating over the past year. Not an association in which people who ate the most blueberries happened to have a lower incidence of dementia 20 years later.

Those things may all be interesting. They may even be important pieces of a much larger scientific puzzle. But they aren't proof that blueberries prevent dementia.

The distinction tends to disappear somewhere between the scientific paper and the headline.

“Associated With” Is Doing a Lot of Work Here

Nutrition research is extraordinarily difficult.

You can put a laboratory mouse in a controlled environment and determine exactly what it eats. Human beings are less cooperative.

We eat thousands of different foods in thousands of combinations over decades. We exercise or don't. We smoke or don't. We sleep differently. We have different incomes, educations, occupations, genes, neighborhoods and, and perhaps most importantly, access to medical care.

And then researchers sometimes ask us what we ate.

If you've ever had trouble remembering what you had for lunch last Tuesday, you can see one of the problems.

Suppose researchers follow 20,000 people and discover that those who eat berries several times a week develop dementia less frequently.

That's interesting. But what have we learned?

Perhaps berries protect the brain. Or perhaps people who regularly buy fresh berries also tend to exercise more, smoke less, have higher incomes, get better medical care, eat more vegetables and generally take better care of themselves.

Researchers attempt to adjust statistically for these differences.

But human beings are complicated creatures, and statistical adjustment is not a magic wand.

This is why two words ought to appear much more prominently in health reporting:

“associated with.”

Instead we frequently get:

BLUEBERRIES PREVENT DEMENTIA.

The first one gets closer to describing what the research actually showed. The second one gets clicks. And, clicks equal dollars.

Then There's Relative Risk

This may be my favorite trick—not necessarily because anyone is deliberately trying to deceive us, but because it produces wonderfully dramatic numbers.

Suppose your risk of developing some condition over a particular period is 2 in 1,000.

Now suppose people who eat our 'Miracle Food' have a risk of 1 in 1,000.

The relative reduction is 50 percent.

So which headline would you rather publish?

FOOD X CUTS DISEASE RISK BY 50%!

Or:

Disease occurred in one fewer person per thousand.

Same numbers. Very different impression.

Whenever I see “reduces your risk by 37 percent,” I now want to know:

Thirty-seven percent of what?

Absolute risk isn't always the better or only useful measure. Relative risk can tell researchers something important. But readers deserve enough information to understand what that impressive percentage actually means.

The Superfood Problem

I have another problem with the whole idea of a “superfood.”

There are certainly foods that are particularly nutritious.

But calling one of them a superfood encourages a peculiar way of thinking about nutrition—as though health were something we could sprinkle on top of whatever else we're eating.

Poor diet + blueberries = healthy diet.

It doesn't work that way. Nor does the reverse:

Healthy diet + one supposedly “bad” ingredient = catastrophe.

Foods matter. But no single food is the diet.

That may be the least exciting nutritional advice ever devised.

It also may be among the most useful.

Compared With What?

Here's another question I'd like to see attached to every nutritional claim:

Compared with what?

Suppose a study finds that people who eat more beans have better cardiovascular health. Fine. What did the beans replace?

If they replaced bacon and sausage, that's one experiment.

If they replaced broccoli, that's quite another.

We don't eat nutrients in isolation. We eat meals. And when we eat more of one thing, we usually eat less of something else.

“Eat more X” therefore tells us remarkably little unless we know what X is replacing.

My Five-Minute Nutrition Headline Test

I've developed a little mental checklist for these stories.

When I'm told that some food will save me—or kill me—I want to know:

1. Was the research done in humans?

2. Was it an actual experiment, or did researchers observe an association?

3. How big was the effect in absolute terms?

4. Was the study really about this one food, or about an overall dietary pattern?

5. Has anyone else found the same thing?

   
   

And I'd add Sagan's question: How extraordinary is the claim, and how extraordinary is the evidence?

None of this means dismissing nutrition research. Quite the opposite.

Good nutrition research has taught us a great deal.

But science normally advances through accumulation. One study adds a tile to a mosaic. Eventually enough tiles may produce a recognizable picture.

The headline business has an unfortunate tendency to pick up one tile and announce that we've discovered the Mona Lisa.

The Boring News May Be the Good News

After decades of nutritional whiplash, perhaps there's something reassuring about how unexciting much of the broad dietary advice actually is.

Eat a variety of vegetables and fruits. Eat legumes, whole grains and nuts. Favor foods that aren't heavily processed. Don't consume enormous amounts of sugar, sodium and saturated fat. Don't eat more calories than your body needs. Move your body.

No miracle ingredient. No forbidden berry. No vegetable that will guarantee you'll blow out 100 candles with your mind perfectly intact. And no need to throw out everything in the pantry because somebody published a study on Tuesday.

By all means, eat the blueberries. They're delicious.

Just don't ask them to make you immortal.