Dear Dr. Molly,
How am I supposed to know what is good for me and what is bad for me? Whenever I read the paper or a magazine, there are results from a new study that contradict that last study. What gives?

Pullman, Washington

When there is a topic of great interest, several groups will often design experiments to learn more about that topic. These decisions are made simultaneously and it can take several years for the results from these studies to be reported. When designing experiments, researchers must take several factors into account. Among these might be: the group of people who will be tested, the duration of the experiment, the outcome that is to be measured, and how that outcome will be measured. As one might imagine, the choices made in each of these categories have the potential to affect the final conclusions of the study. For example, one experiment might test how a high fat diet affects the waistline of middle-aged women over five years while another study might test a high fat diet on sleep quality in adolescent boys over one week. These two studies could be interpreted quite differently and different conclusions regarding high fat diets might be reported.

One real life scenario in which confusing results have been reported is in the consumption of coffee. Coffee has been blamed for transient increases blood pressure (Childs E and de Wit H) and increased cholesterol (Higdon JV and Frei B). Other experiments, however, have shown that female lifelong coffee drinkers perform better on cognitive tests late in life (Johnson-Kozlow M, et al.). Finally, recent studies have shown that coffee may be a valuable source of antioxidants (Halvorsen BL, et al.) and folic acid (Child Health Alert) and might even prevent certain diseases (Higdon JV and Frei B). This plethora of results is clearly confusing, but if the different experimental questions and designs are taken into account, it is understandable how these different results came to be. Coffee has many different aspects (caffeine, antioxidants, processing chemicals) that could be used as the basis of experimental questions. The populations of individuals tested in these examples range from general populations of caffeine and non-caffeine users to aged women and pregnant women. The experimental time frame ranged anywhere from one day to a lifetime. The outcomes were measured by surveys, blood tests, laboratory tests and clinical tests of cognition. Each study provides unique information, but the results might seem contradictory when synopsized into a headlines. If you are curious, read more about the details of the study to understand the context of the results. And, with coffee, the real take-home message is what we have heard many times before: anything is okay within moderation.

Child Health Alert. 2007 Mar;25:5-6.

Childs E and de Wit H. Psychopharmacology (Berl). 2006 May;185(4):514-523.

Halvorsen BL, et al. Am J Clin Nutr. 2006 Jul;84(1):95-15.

Higdon JV and Frei B. Crit Rev Food Sci Nutr. 2006;46(2):101-123.

Johnson-Kozlow M, et al. Am J Epidemiol. 2002 Nov 1;156(9):842-850.