Beyond the Headlines

Sorting Out Nutrition Science

Epidemiologic or population-based studies, which are often the source of nutrition news, look at very large groups of people and are also observational by nature. For example, an epidemiologic study looking at the habits of Kenyan distance runners over the last 20 years could find that those runners who ate greater amounts of ugali (carbohydrate-rich stew) each day had a better chance of winning an Olympic medal. That is, the study found an association between eating more ugali and winning a medal. These results, from observing hundreds or even thousands of Kenyan runners, however, do not prove that eating more ugali was the cause of winning a medal.

Only controlled experiments can determine cause and effect. Here, study subjects are selected according to particular characteristics and then randomly assigned to either a control group or an experimental group. This random assignment is paramount as it helps ensure that variables that may affect the outcome of the study are distributed equally among the groups. In our case, male marathoners ages 30-45 would be recruited, screened and then randomly assigned to one of two groups; half would be given Supplement X and half would serve as the control group (they would be given a placebo: a sugar pill or fake treatment). The men themselves would be "blind," that is, they would not know to which group they have been assigned.

To further strengthen the results of the study, the researcher wouldn’t know which group was getting what until all the data had been gathered and analyzed, thus eliminating (as much as possible) researcher bias. This gold standard in research is called a double blind, placebo-controlled randomized clinical trial. Unfortunately these types of trials aren’t always feasible. For example, it may be hard to hide the fact that one group is getting a certain food compared to the control group, or what workouts they are doing. In our case, if the fastest male marathoners turn out to have taken supplement X, we can then more confidently claim that running faster in the marathon (the effect) can be attributed to taking Supplement X (the cause).

Bottom line: Because two things are related, it does not mean that one is the cause and the other the effect. Observing what someone else does and reporting on it can only reveal associations—it cannot prove cause and effect.

Strategy #3: Understand what the numbers really mean.

For people who live by the numbers, runners often get thrown off course by percentages, statistics and proportions related to risk. It’s important to differentiate between the relative numbers in terms of the risk (or success) for a particular outcome and the actual (raw) data. For example, when estimating the absolute risk of a healthy woman developing breast cancer, the statistic often quoted is 1 out of 8. This "high" 12.5% risk of developing cancer causes much anxiety among women of all ages and running abilities. However, this is the lifetime risk for a woman who lives to be 90! Your absolute risk depends on your age: 1 in 2,500 in your 20s, 1 in 232 in your 30s, 1 in 55 in your 40s and so on.

Bottom line: Get past the attention-grabbing headlines and sensational statistics. Although the relative risk (or success) of a particular outcome may appear high, check what the actual numbers show.

Suzanne Girard Eberle, M.S., R.D., author of Endurance Sports Nutrition, lives and runs in Portland, OR.