Sunday, November 23, 2014

Nutrition Labels: Calories and Sugar

I am a little disappointed in the scientific rigor of the two studies I mention below, especially because they address two issues of importance to me - and you; restaurant menu and food package labeling. The study designs/methods make it hard, no impossible, to make causal inferences.  We can't say x caused y, in either study - but to be fair, the second study was only meant to describe peoples perceptions.  Unfortunately, the people surveyed are not representative of the larger US population.

The first study (from the University of Glasgow in Scotland) compared the average weight gain of two different groups of college students - during two different years.  The first group may or may not have eaten at the cafeteria when no calorie labels were present and the second group may or may not have eaten at the cafeteria when calorie labels were present.  The first group gained 8 pounds during 36 weeks (perhaps a semester) and the second group gained 4 pounds during the 36 weeks they were assessed.  There is absolutely no way to say that the calorie labels were responsible for the weight gain differences - none.  And I am 100% pro labeling. However, during the second year when calorie labels were on the menus, the customers did purchase less calories than when the labels were not on the menus - during this one time period.  That is important.

The second study - a very small, internet survey- strove to answer the question of whether or not Americans would be confused by the addition of an added sugar declaration on the Nutrition Facts Panel (the panel on the back of packaged foods).  The rationale for this study was the assertion by anti- added sugar labeling folks that consumers would not find the additional nutrition information helpful - on the contrary, naysayers say - added sugar information will confuse the public.

The assertion that more information is bad really gets under my skin, because again, I am pro labeling.  However, there is evidence from studies on nutrition information that suggests we cannot apply available nutrition information in the context of our personal, daily needs.  That is why the IOM and others have recommended more intuitive labels, ones that highlight a few nutrients of concern, like salt, sugar and overall calories- in color coded front of the package schemas.

But the FDA is considering changes to the Nutrition Facts Panel(NFP) on the back, including an added sugar declaration - read more here.  Ted Kyle and his colleague surveyed a small group of internet users, a couple hundred, and asked them if the added sugar section of the NFP was helpful or confusing.  The survey respondents looked at a make believe label before answering the question.  Most of the people who participated said that the additional information was helpful.  I think this is probably right, however, the survey doesn't really tell us anything about what people will do when the labels actually contain this information, we need a different kind of study for that.

You can read a popular press write up about the Glasgow study here, and an over view of the added sugar survey results can be found here.  The ConscienHealth blog written/maintained by Ted Kyle focuses on more rigorous studies than the added sugar survey  and is one of my favorites.

Tuesday, November 4, 2014

Prevent Obesity w/ More exercise - less fast food

Morales, Gordon-Larson and Guilkey (2014) conducted a simulation experiment where they statistically modeled several factors that are related to the probability of a person being obese in the future.  The model tested their assumptions by predicting who would become obese. You can read the study here .- Several statistical equations were used to simulate decisions about smoking, physical activity, starting a family, and fast food consumption.  In addition, the researchers added a neighborhood level factor.

The data used were real - information came from the National Study of Adolescent Health (USA) which contains several years of information following people from high school through young adult hood.  During those years, people in the study usually transition from a home environment chosen by their parents, to one they chose themselves.  The choice of environment is based on some unknown personal preferences or necessities of the person.  Since many researchers speculate that where a person lives has a direct impact on whether or not that person becomes obese, these researchers felt it important to capture neighborhood choice in their model predicting obesity.  In fact, Morales, et al knew exactly where the participants lived at all times and added the number of parks, complete streets and fast food restaurants in each persons neighborhood to their simulation model.

The results of the study are interesting and give public health advocates, researchers, policy makers and the general public some things to consider.  The researchers tracked people over time in their simulation model, too, so they carried forward things about the person that had been learned at the previous time point.  They found that the best or most powerful predictor of whether or not a person would be obese at time 2 or 3 or 4 was whether or not they were obese the time before that.  In fact, the researchers assert that the most important thing to do is prevent obesity in the first place, because a reversal just isn't seen.  In other words, if a person was obese at any of the previous times it is very unlikely that they will be normal weight at any point in the future.  I agree that prevention is the key.

Other findings that seemed important to me, include; the more fast foods a person consumed the more likely they were to be obese; and physical activity reduced the probability of obesity, but in a nuanced way.  The researchers found that at least 30 minutes of physical activity 5 days a week were needed to keep a person from becoming obese and that the greatest impact on weight - or the best chance of not being obese - came from maintaining that level of physical activity from adolescence (i.e., high school) throughout adulthood.

Neighborhood characteristics, a main focus of this study, did have an impact on the probability of being obese, but the magnitude of the effect was much smaller than physical activity and fast food consumption.