Junkfood Science: If only it were true...

January 10, 2008

If only it were true...

News around the world has reported that “healthy habits can add 14 years to your life.” As incredible as that sounds, more incredible is that not a single health or medical reporter appears to have read the study behind the headlines and accurately report its findings.

This phenomenon embodies an old adage that a very wise doctor and researcher reminded me of just this morning:

“A falsehood will fly from Maine to Georgia, while truth is pulling her boots on.”unknown author, Boston Commercial Gazette, 1820 [That is, if the truth can afford boots.]

Within hours this week, television newscasters, as well as some 500 published articles — for scientists, medical professionals, nursing professionals, business professionals and consumers — were all reporting the same interpretation of this study. The script was provided by the publication’s press release: “4 health behaviors can add 14 extra years of life.”

“New Year resolutions could add years to your life “ — New Scientist

“4 Steps to Living 14 Years Longer” — WebMD

“Study reveals path to longer life” — Nursing Practice

“Healthy Living Worth 14 Years” — Staff Nurse

“Small Lifestyle Changes Can Boost Longevity” — Business Week

“Healthy Habits Can Mean 14 Extra Years” — Associated Press

“Academics find formula for 14 extra years of life” — Guardian

“Healthy life 'can give you another 14 years” — Telegraph

Simply making extraordinary statements about a study, however, does not make them true — no matter how much we might want them to be. And who wouldn’t want to believe that by doing just four easy things we could add 14 years to our lives?

If it sounds to good to be true....

For this study, Cambridge researchers with the Medical Research Council (the primary sponsor) and Institute of Public Health at the University of Cambridge School of Clinical Medicine, used the Norfolk, UK participants in the European Prospective Investigation into Cancer and Nutrition (EPIC) database. JFS readers will remember (here, here, here) that EPIC is an enormous database of self-reported diet and lifestyle questionnaires and medical histories; body measurements; and lab results gathered on 521,000 Europeans in 10 countries. The EPIC database has been dredged by researchers looking for endless correlations between lifestyles and disease.

In 1993-7, Norfolk adults (ages 45-79) had completed one-time questionnaires, been measured and had labwork drawn. For this paper, the data on 20,244 healthy participants without histories of heart disease, cancer or strokes at the time of enrollment were used. The researchers selected four health behaviors as reported at the time of enrollment — smoking, physical activity, alcohol intake, and plasma vitamin C levels (as a marker for fruit and vegetable intake). They labeled as ‘healthy’: nonsmoking, vitamin C levels ≥50nmol/l, active (a nonsedentary job or 30 minutes of leisure activity a day), and 1-14 alcoholic drinks a week*. The researchers then tallied the death certificates from the Office of National Statistcs through 2006, and used disease billing claim codes (International Classification of Disease, ICD) to learn the underlying causes of death. Using computer modeling, they looked for correlations between each of these health behaviors and all-cause mortality and certain disease-specific deaths.

Their key finding, which has not been reported, was they were unable to find a tenable correlation between any of the health behaviors and mortality: all-cause, cardiovascular disease, cancer or any other cause of death. The relative risks all hugged either side of 1 — null findings.

Their working premise, however, was that they knew “lifestyles such as smoking, diet and physical activity have a major influence on health.” They then tallied the health behaviors and looked for correlations. Compared with those who’d reported doing four healthy behaviors more than a decade earlier, they found no tenable correlation with higher all-cause mortality among those reporting only three healthy behaviors more than a decade earlier ... nor among those reporting only two healthy behaviors ... nor among those reporting only one healthy behavior.

Only among those who reported no healthy behavior at all at the time of enrollment (49 deaths) did the relative risks for mortality go above chance or statistical error. We have no idea, however, what countless possible confounding factors might actually point to a cause for this correlation.

In other words, reports of higher relative risks associated with not engaging in four healthy behaviors were based on 49 people, 0.2% of the cohort. But the absolute (actual) risk of dying differed only 0.2% between those doing zero and those doing all four healthy behaviors (0.25% and 0.05%, respectively).

Increasing life expectancies

Extrapolating statistical relative risks associated with healthy behaviors derived from a data dredge of a population, even if they were significant, into added years of life is where things become especially dubious. It is popular to believe that we can increase our lifespan, even prevent chronic diseases and aging, by following some ideal lifestyle or diet. These core premises were examined at length in “Anti-aging diets,” when fifty-one of the world’s most recognized scientists in the field of human aging examined the evidence and concluded such beliefs have little to no scientific basis. “Our language on this matter must be unambiguous: there are no lifestyle changes, surgical procedures, vitamins, antioxidants, hormones or techniques of genetic engineering available today that have been demonstrated to influence the processes of aging,” they concluded.

Getting enough to eat and enjoying a variety of foods to prevent deficiencies and enjoying life is certainly good. But the idea that we can add years to our life by eating or engaging in certain ‘healthy’ behaviors continues to be popularized, despite no credible evidence, such as the eleven behaviors that doctors say can add 30 years to our life!

Despite mathematical possibilities of population data, “they ignore the biological forces that influence the length of life of individuals in genetically heterogeneous populations,” wrote Dr. S. Jay Olshansky and colleagues with the Center on Aging at the University of Chicago. It goes beyond the fact that infectious diseases, social and political unrest, accidents, natural disasters and other events limit the reality of life expectancies. As they explained:

[W]hen life expectancy at birth is 50, it takes an estimated 4.1% reduction in total mortality at every age to raise life expectancy 1 year, a mortality scenario similar to that experienced by French females at the beginning of the 20th century. By contrast, raising life expectancy from 80 to 81 years requires a 9.1% reduction in total mortality at every age. The mortality reductions at every age required to achieve a 1-year increase in life expectancy at birth today are more than twice those needed to achieve the same gain early in the 20th century.

Dramatic improvements in life expectancies seen in most of the 20th century, achieved by reductions in child deaths from infectious and parasitic diseases, for instance, are unlikely to be repeated, they noted. Future gains will come at a slower pace and as cures and treatments for major diseases that come with aging, such as cancer, are discovered. “There are no life-style changes, surgical procedures, vitamins, antioxidants, hormones, or techniques of genetic engineering available today with the capacity to repeat the gains in life expectancy that were achieved during the 20th century... Although death rates among long-lived populations continue to decline at middle and older ages, the chance of achieving a life expectancy at birth of 90 years and older has not changed appreciably over the last decade.”

As seen in mortality trends since 1985, they said, “future gains in life expectancy will be measured in days or months rather than years.”

Even decades ago when death rates from cancers and heart disease were higher than they are today, potential gains in life expectancies among populations by eliminating the major causes of death were far less significant than might be guessed. For example, researchers examining Finland data, reported in the International Journal of Cancer that totally eliminating ALL forms of cancer and all cancer-related diseases among the entire population would add 2.38 days. The total elimination of all lung cancer would gain 8.9 hours. The total elimination of all breast cancer would gain 4.6 hours.

Another older report examined potential gains in life expectancies in Texas. They found that eliminating half of all cardiovascular disease in the entire population (ages 15 to 70) would add 6 to 12 months to life expectancies of the population.They calculated that eliminating half of all cancer could theoretically add 3 months to the life expectancies of the population. “Even with a scientific breakthrough in combating those causes of death, it appears that future gains in life expectancies for working ages will not be spectacular,” they cautioned.

So when you hear claims that some special lifestyle or food can add years to your life, let alone decades, it’s not founded on science.

Closing thoughts

Going back to the original Cambridge study behind this week’s headlines, there was no need to have even read the study or understood the science to recognize that the headlines were unsound. What was the dead give away? The “most common source of unsound health interventions:” Correlations are not causation. The first line of the abstract said: “We examined the prospective relationship between lifestyle and mortality in a prospective population...”

Yet a correlation found among a population was turned into a cause and then transformed into a means to a longer life. [And a means to sell health policies, healthy eating and exercise programs, and to blame those who die young as it being their own fault.]

Of course, we can also fall back on another axiom: The saturation of media coverage is usually inversely proportional to the soundness of the science. When one study — out of the thousands of new studies released every single day — is reported in every media outlet, on the exact same day, all saying the exact same thing, this is primarily evidence of brilliant marketing. :)

© 2008 Sandy Szwarc

* As they noted, the body of evidence has consistently shown the relationships between alcohol consumption and mortality to have a wide “u” shape, with lower risks for a wide range of consumptions, and higher risks only at both extremes. Internationally, upper-limit alcohol recommendations for health range from “five drinks daily for men and three drinks daily for women in France; to two drinks daily for men and one for women in the United States. In the UK, the recommendations are up to 21 drinks weekly for men and 14 drinks weekly for women.”

Bookmark and Share