Science By Headline
Chester J. Zelasko, Ph.D. | November 13, 2007

The past few weeks have been filled with more science by headline than I can ever remember. Readers Digest had an article titled “The Vitamin Myth” that prompted numerous questions about the safety of taking supplements, especially antioxidants (1). The American Institute of Cancer Research published a 537-page paper on factors associated with an increased risk of cancer (2). What led things off on the evening news from that report? The relationship between obesity and cancer. Not to be outdone, the Center for Science in the Public Interest, the “heart attack on a plate” people, updated information on Italian restaurants in a written publication and in video (3). If you read or listened to any of these, you’d have thought that record numbers of people were keeling over in the street from taking antioxidants, being overweight, and eating Italian food.

I’m going to take a different approach in this Newsletter. Instead of examining each article point by point, I’m going to provide some recommendations for scientists, authors, and expert commentators. By reading this Newsletter, you’ll understand what is real and what is hype in all these science-by-headline examples.

Recommendation 1: You must be able to explain the study. If you as a scientist cannot explain to some 15-year-olds how you analyze such dissimilar data, and make them understand it, you don’t get to be on television telling us we’re going to die if we take our vitamins.

One of the studies talked about in “The Vitamin Myth” was the study published in the May issue of Journal of the American Medical Association (4). This study has been commented on over and over in newspaper and magazine articles. It all revolves around a single conclusion from that study: those subjects who used antioxidants, specifically vitamin A, E, and beta-carotene, had a 16% increased risk of dying compared to those who didn’t. One would think that this study has significant merit because of the number of subjects--over 200,000. But that’s not exactly true. This was a meta-analysis of 68 studies. Some of the studies lasted a few months while others lasted five years and longer. Some had 25 subjects, others had close to 10,000. You don’t need to understand a thing about statistics to see that there might be some problem with this approach. It’s not just apples and oranges--more like apples and dragons.

The statistics used in many studies today are well beyond anything that many scientists can understand, let alone explain to the lay public; they’ve come from advanced mathematics and statistics. It’s one thing to say that the data were comparable because you accounted statistically for length of study, different doses of supplements, different ages, and different diseases. It’s a whole other thing to explain it to someone at a basic level--but if you can’t, stay in your lab.

Recommendation 2: You must translate the results you present into something the average person can understand and apply. If you’re the writer or the interviewer, you must insist that the scientist do that for you. In the publication by the AICR, the risk of six different types of cancers increased from 5% to 100% as body weight rises. That sounds ominous--but what does it really mean?

Here’s an example from that paper. If you’re a post-menopausal woman, your risk of breast cancer increases about 5% for every two units of Body Mass Index (BMI) or about 15-18 pounds depending on height. What is the risk of breast cancer for postmenopausal women? About one out of a thousand. Even if you went from a normal BMI of 25 to an obese BMI of 30, that would increase your risk 12.5%. That would mean instead of 1.000 out of a thousand it would be 1.125 out of a thousand, and that’s not nearly as scary. The skeptical person might speculate that scary numbers are more likely to get the scientist face time on TV. I’m not suggesting that women should gain weight or stop trying to lose it, but if you’re going to quote numbers, you must explain what they mean in the real world or you don’t get to say them.

Recommendation 3: Whether you’re a scientist or writer, you have to complete the story. Let’s go back to the 16% increase in mortality for antioxidant users. The studies used in that meta-analysis took place between 1990 and 2005. Because the studies took place within this time frame, there should have been some sort of increase in mortality--at least in the areas of heart disease and cancer, two areas specifically mentioned by the study’s authors.

Mortality statistics are usually presented in deaths per 100,000. That way, when the population increases or decreases, the stats are automatically adjusted for the change in numbers. In 1990, the mortality rate for cardiovascular disease (CVD) was 290 per 100,000 and 203 per 100,000 for cancer (5). In 2004, the latest year for mortality statistics, the mortality from CVD was 222 per 100,000 and 187 per 100,000 from cancer (6). That represents a decrease of 23% from CVD and 7.8% from cancer. Where is the 16% increase? Does this mean that in spite of trying to kill ourselves with supplements, medicine has become so good it’s saving us from ourselves? Don’t start the story if you can’t finish it.

Recommendation 4: Scientist, think tank, or author, back up what you say. I’ll give three examples--one from each of the articles.

In “The Vitamin Myth,” one of the experts says that you can get your vitamins and minerals from eating a well-balanced diet. There’s no evidence that’s true. Let me explain. I have yet to see the study that examines the food purchased, prepared, and consumed in real life to see if someone who eats a well-balanced diet actually gets all the vitamins and minerals they need. Food is transported to the market, it sits on shelves before purchase, it sits in refrigerators before it’s prepared, and finally it’s cooked. If you want to know what nutrients people are really getting from their diet, that’s the food that needs to be analyzed, not a diet made up of numbers from the USDA database. No one has done that, so we don’t really know what nutrition you’re getting from the foods you eat--even if you eat all the right foods.

In the AICR recommendations, the conclusion is that a reduction in weight will reduce the risk of cancer. There is no evidence to support that position. To my knowledge, no research group has had overweight people lose weight and then monitored them to see if there are fewer instances of cancer than expected over a period of years. They’re basing their conclusion on the data they present and assuming there’s a dose-response relationship between body weight and cancer; they deduce that if you lose the weight, you reduce the risk. You have to prove it, not deduce it--especially when you’re trying to set public-health policy for the world.

In the CSPI article, they imply that people eat Italian foods because they think it’s healthy. They also imply that people eat this way all the time. They provide no data on how often people eat at Italian restaurants, how often people share appetizers, entrees, and desserts, or how many people take some of their food home; they don’t analyze how often people cook Italian food at home. They provide no data, but they certainly make judgments about the dietary habits of Americans. If eating even the worst meal were a once-a-week occurrence, it might not be as bad as they portray. We’ll never know because that’s not what CSPI seems to be good at--they seem to excel at press conferences announcing that yet again the sky is falling.

Bottom Line: What does this mean for you? If you’re reading articles in the press or on the Internet, and the aspects that I’ve suggested are not explained, just move on to the sports scores or the gardening section. You’re going to get more information and better statistical analysis from those areas than you are from health articles that are incomplete. Yes, it’s easy to sit back and comment on what others haven’t done. What isn’t easy is reading the research and articles that these groups publish and giving you the real story. That’s what I do because they don’t or won’t. The goal for every healthcare professional should be to inform and educate, not mislead and frighten. We’re going to keep giving you the real deal at Better Life.

  1. Neena Samuel. The Vitamin Myth. Readers Digest. November, 2007.

  2. World Cancer Research Fund / American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. Washington DC: AICR, 2007.

  3. A 3,000-Calorie Dinner? Belly-ssimo!: CSPI Inspects Olive Garden & Romano’s Macaroni Grill. October 31, 2007.

  4. Bjelakovic, G et al. Mortality in Randomized Trials of Antioxidant Supplements for Primary and Secondary Prevention Systematic Review and Meta-analysis. JAMA. 2007;297:842-857.

  5. Leading Causes of Death, 1900-1998. National Center for Health Statistics.

  6. Arialdi M. Miniño, AM et all. Deaths: Final Data for 2004. Division of Vital Statistics.
BBBOnLine Reliability Seal © 2011 Better Life Unlimited™
A division of Better Life Institute © (BLI, Inc.)
 Contact Us  |  Privacy Policy
SecurityMetrics Credit Card Safe