seems to contradict itself with aggravating regularity. You stop using butter and instead start spreading margarine on your toast, only to learn later that margarine can be just as bad for you as butter. After switching to bran muﬃns for breakfast because high-ﬁber diets supposedly prevent colon cancer, you hear about a big study showing that ﬁber doesn’t prevent colon cancer. Early research shows that coﬀee drinking increases the chances of developing pancreatic cancer, while later research shows that coﬀee drinking is harmless and may even have some beneﬁts. Some studies ﬁnd that eating ﬁsh prevents heart attacks, others don’t. These ﬂip-ﬂops are so confusing and so common that a negative report on vitamin E and beta-carotene once goaded Boston Globe columnist Ellen Goodman to write, “There seems to be some sort of planned obsolescence now to medical news. Today’s cure is tomorrow’s poison pellet. Fresh research has a sell-by date that is shorter than the one on the cereal box.”The sheer volume of information doesn’t help. Fifty years ago, medical researchers mostly ignored nutrition. For example, the longest study of health in the United States, the legendary and ongoing Framingham Heart Study, collected hardly any data on diet when it was started in 1949. Over the years, though, the trickle of information on diet and health has swelled into a fast-ﬂowing torrent.It’s only natural that people want to know the latest results, whether they are looking for ways to ﬁne-tune their diets or for that single magic key—the right food or vitamin or supplement—that will open the door to the longest, healthiest possible life. The media are only too happy to cater to this interest and serve up a steady stew of health news.The problem is that newspapers, television, the Internet, and other news venues often turn the baby steps of scientiﬁc research into “major advances,” “breakthroughs,” and “possible cures” or highlight the confusing contradictions. This makes getting health news seem like reading pages torn at random from a site.
REPLACING EDUCATED GUESSES WITH EVIDENCE
Another reason for the contradictions is that weighty recommendations about diet were often based on thin evidence. The thinking behind these early recommendations was that since people were going to eat no matter what, guidelines based on intelligent guesses were better than no guidelines at all. That’s actually a reasonable approach when there isn’t much evidence. Unfortunately, these recommendations never carried warning labels like “Educated Guess, Subject to Change,” and after being repeated thousands of times they acquired the ring of truth.
When researchers began learning of the dangers of saturated fat, for example, many recommended that people switch from butter, which is high in saturated fat, to low- saturated-fat margarine. This recommendation made sense, even though there were no studies showing that people who ate margarine instead of butter had fewer heart attacks. Then along came studies showing that margarine eaters didn’t fare better in the heart-attack department than butter eaters. To a scientist, this is the normal path of scientiﬁc progress—a recommendation based on a good guess is tested and toppled by one based on good science.
To the rest of the world, though, it is a frustrating contradiction.
The amount and quality of sound scientiﬁc information on diet and health have grown enormously over the past twenty years. That makes today’s evidence-based recommendations much more certain and much less likely to need radical changes than those made two decades ago. As the quest for new and better knowledge about diet and health continues, rest assured that even today’s recommendations will probably be subject to some ﬁne- tuning.
CONTRADICTIONS ARE INEVITABLE
Medical science has its own special rhythm, one that doesn’t ﬁt with the media’s need to tell compelling but simple stories. Eﬀorts to present “balanced” stories by quoting opposing views can sometimes confuse things even further.
For nutrition research, the rhythm is more a cha-cha— two steps forward and one step back—than a straight-ahead march. If you look at the day-to-day results reported more like sports scores than scientiﬁc research, it’s easy to wonder why researchers can’t get it right the ﬁrst time.
They can’t because these conﬂicts and contradictions are the way science works. It happens this way in every ﬁeld, from archaeology to zoology, nuclear physics to nutrition. Men and women carry out studies and report their results. Evidence accumulates. Like dropping stones onto an old- fashioned scale, the weight of evidence gradually tips the balance in favor of one idea over another. It is only when this happens that you should make changes in your life.
The size of the stone clearly makes a diﬀerence. As we describe on pages 30–33, most studies are like sand grains or small pebbles. Very few are like boulders.
learn more: Eat, Drink, and Be Healthy!
WORKING WITH REAL PEOPLE POSES SPECIAL CHALLENGES
Nutrition research seems to generate more than its share of contradictory results. That’s partly because the media pay special attention to nutrition (because of the public’s interest), while inorganic chemistry, geology, and many other disciplines escape this daily scrutiny. It’s also because nutrition scientists usually can’t exert the same kind of control over their research subjects as can chemists or zoologists. Instead they must work with unpredictable, independent, mostly uncontrollable subjects—people.
Here are a few of the challenges that face nutrition researchers:
• People don’t eat “human chow” meal after meal after meal. Instead diets change from day to day, week to week, and season to season. What you usually eat now is probably a bit (or maybe a lot) diﬀerent from what you used to eat two years ago or will eat two years from now. These changes are driven by personal taste, cultural changes, improvements in agriculture and technology, and changes in work and family life. They may also be due to disease or aging.
• Many studies depend on people accurately reporting what they eat, a challenging task. (Try remembering exactly what you ate one day last week.) Despite this diﬃculty, people are fairly accurate about reporting their longer-term eating pattern. But because they aren’t perfect, there’s almost always some uncertainty in linking diet and disease.
• The foods you eat each day contain thousands of diﬀerent natural chemicals, some known and well studied, some known and unstudied, many completely unknown and unmeasurable. So far, we’ve ﬁgured out what only a small percentage of them do in the body. Collecting information on others, and discovering how food compounds interact, is an important job for the future.
• Calculating the nutrients a person gets from the foods she or he eats—how much saturated fat, ﬁber, vitamin E, and so on—is tricky since it depends on sometimes sketchy information about food composition.
• Almost everyone eats some fat, ﬁber, sugar, starches, fruits, vegetables, vitamins, and so forth. That means nutrition researchers are faced with the more diﬃcult task of measuring how much of something is eaten, not just whether it is part of the diet.
• Heart disease, cancer, diabetes, osteoporosis, cataracts, and other chronic diseases almost always develop over many years. They also have other causes beside diet, including genes, physical activity, smoking, stress, and other factors yet to be identiﬁed.
DIFFERENT METHODS FOR DIFFERENT PROBLEMS
To get around these problems, nutrition scientists use a variety of research methods.
• Randomized trials. The “gold standard” by which other studies are usually judged is the randomized trial. In these carefully controlled studies, half of a group of volunteers is randomly assigned to the experimental diet or treatment, and the other half is assigned to the standard diet or treatment (the control) or possibly to no treatment at all. After a preset time, the number of people in the control group who have developed the predetermined “endpoint”— death, heart attack, broken hip, and so on—is compared with the number in the experimental group.
For example, say you want to know if vitamin C prevents age-related memory loss. You would round up a large group of volunteers, then randomly assign some to take a daily vitamin C tablet while the others take an identical tablet that contains an inactive ingredient that tastes like vitamin C (a placebo). After ten or twenty years, you would compare the percentage of people in the vitamin C group who have experienced memory loss with the percentage in the placebo group.
learn more: Building a Better Pyramid
This kind of study has plenty of advantages. If it is large enough, the randomization process does a good job of making sure the people in the experimental group are very, very similar to those in the control group in terms of age, health, exercise, and other possibly important factors. So the only thing diﬀerent between the two groups is the diet or treatment. Unfortunately, randomized trials are often impossible to do when it comes to nutrition. Getting people to ﬁx and eat special meals for a long time is diﬃcult. So is getting people to take a vitamin pill or placebo for maybe a decade or more. Given the large number of volunteers needed, the cost of running a randomized trial can be astronomical. The Women’s Health Initiative, which is primarily testing the impact of reducing dietary fat to 20 percent of calories and increasing fruits and vegetables on the development of breast cancer, will cost more than $1 billion and still probably won’t yield clear answers on this important question.
• Cohort studies. The next best method involves following large groups of what epidemiologists call “free-living humans”—regular people like you—for long periods of time. These cohort studies start with a group of people who often have something in common, like an occupation or place of residence. They are asked about their diets, smoking and drinking habits, education, occupation, medical conditions, and other possibly relevant things. The group is then followed for a period of time, ideally a decade or more, either directly with occasional checkups and mailed questionnaires, or by monitoring death certiﬁcates. Once the study has gone on long enough, researchers can examine the accumulated information to test a variety of hypotheses. They could, for example, determine if people in the cohort who eat the most ﬁber have diﬀerent rates of colon cancer than those who eat the least ﬁber, or if those who consume the most folate, an important B vitamin, have lower rates of heart disease than those who consume the least folate. Such long-term studies have yielded some of the best insights so far into the link between diet and health. By gathering information at the beginning, before speciﬁc diseases have occurred, cohort studies avoid the skewed recall sometimes seen among people who develop a particular disease—and who would like to ﬁnd an explanation for it. Cohort studies such as the Nurses’ Health Study, the Health Professionals Follow-up Study, and others use a carefully tested questionnaire to determine what the participants eat and ask them to ﬁll it out several times over the course of the study. This reduces errors and also lets researchers look at changes in diet over time.
• Case-control studies. In this type of study, a researcher gathers information from a group of people who have developed a particular disease (the cases) and a similar group of people who are free of that disease (the controls) and compares the two groups for diﬀerences in diet, exercise, or whatever variable he or she is interested in. Case-control studies are eﬀective tools when that variable is clear-cut—say, all-or-nothing things like cigarette smoking or occupation. They don’t work as well for diet, when only small diﬀerences are likely to be seen from person to person. Case-control studies are also more prone to error and bias than cohort studies. Because case-control studies can be done quickly and inexpensively, they supplied the evidence for many of the early recommendations about diet and health. As information emerges from cohort studies, though, we are ﬁnding that the conclusions from case-control studies were often oﬀ the mark.