When I was director of the Congressional Budget Office, I was testifying so frequently to the Senate Finance Committee that the chairman granted me a special exception to the committee rules: He allowed me to drink diet soda rather than just water during hearings. At my peak, I was downing up to eight diet sodas a day.
My family did not think this was such a great idea, and at their urging I have largely eliminated my diet soda drinking. Recent data from Beverage Digest suggest others are cutting back also; consumption of diet sodas fell more than that of sugary sodas in 2013. “While the health risks of sugary sodas have been publicized for some time, the growing public aversion to diet drinks -- with many believing artificial sweeteners are also unhealthy -- has caught the industry somewhat off guard," the Wall Street Journal noted.
This raises two questions: Why is total consumption declining, and is drinking diet soda harmful to health? Although the data do suggest a change in attitudes toward diet sodas, one potentially underappreciated factor in the consumption data is the role of demographics. Consumption of diet soda is twice as high among non-Hispanic whites as among Hispanics, for example, as data from the National Center for Health Statistics show. As the share of non-Hispanic whites in the population declines and the share of Hispanics rises, one would expect diet soda consumption per person to fall.
Demographic shifts don't entirely explain the shift in soda drinking, however. Concern about health effects appears to be rising, too -- which is not surprising, given how much media attention has been trained on reports of harm.
A recent example is an analysis of the extent to which the consumption of diet drinks influences cardiovascular events -- a study that highlights both the insights and the shortcomings of research on diet soda. Ankur Vyas and colleagues at the University of Iowa examined the diet drink consumption of almost 60,000 older women and concluded that, “compared to women who never or only rarely consume diet drinks, those who consumed two or more a day were 30 percent more likely to suffer a cardiovascular event and 50 percent more likely to die from related disease.”
The Vyas study identifies a troubling correlation between diet soda drinking and significant health risks. Yet, as the authors themselves point out, this correlation proves nothing about causality.
To understand why this matters, consider that only about 5 percent of the women in the study were drinking two or more diet sodas a day. And these 5 percent were different from the rest of the women in many ways, in addition to their diet soda drinking: Almost twice as many of them were obese (42 percent versus 23 percent), and almost twice as many were smokers (almost 8 percent versus a little more than 4 percent). If overweight smokers are, for whatever reason, more likely than others to drink diet soda two or more times a day, that shouldn't lead us to believe it's the sodas that are increasing their risk of cardiovascular disease.
It is also interesting that the "regular" diet soda drinkers were just as likely as the other women to be daily drinkers of sugar-sweetened beverages. That underscores another concern: The study did not compare people who drank diet soda with those who drank sugared soda or other types of drinks; indeed, it is unclear what the people who “never or only rarely” consume diet soda are drinking.
The authors deployed a statistical model to adjust for the adverse health of effects of smoking and other variables, and they still found a correlation between diet soda consumption and cardiovascular risk. Yet even this residual correlation is only as sound as the model itself. Those who drank far more diet soda seem so unlike the others that one has to wonder whether all the differences have been fully controlled. That is not to say that diet sodas are necessarily safe, but just that this study doesn’t prove they are harmful, as its authors themselves (but not most news reports about the study) emphasize.
This conclusion is echoed in the wider literature about the health effects of diet soda. A careful analysis of the research by Richard Mattes of Purdue University and Barry Popkin of the University of North Carolina found that long-term randomized controlled trials are essential, because almost every existing study is subject to concerns about correlation versus causation.
I’m glad that my family succeeded in getting me to shift away from diet soda and toward tea. Yet I’d say the case is still out on whether my health was ever at risk.
(Peter Orszag is vice chairman of corporate and investment banking and chairman of the financial strategy and solutions group at Citigroup Inc. and a former director of the Office of Management and Budget in the Obama administration.)
To contact the writer of this article:
Peter Orszag at firstname.lastname@example.org.
To contact the editor responsible for this article:
Mary Duenwald at 1+212-205-0366 or email@example.com.