Eating Out is Killing You

“Our findings from this large nationally representative sample of US adults show that frequent consumption of meals prepared away from home is significantly associated with increased risk of all-cause mortality.” 

Yang Du, MD, Ph.D. candidate, Department of Epidemiology, University of Iowa

How did Dr. Du and the rest of the team come to this conclusion? The study used data from NHANES, an ongoing cross-sectional study of Americans involving questionnaires and physical examinations. The data used in the research came from 35,000 respondents, age twenty or more, between 1999 and 2014, who reported dining out.

The responses to 

• “On average, how many times per week do you eat meals that were prepared in a restaurant?”

or

• “During the past 7 days, how many meals did you get that were prepared away from home in places such as restaurants, fast food places, food stands, grocery stores, or from vending machines?”

were recorded (the wording changed over those fifteen years). The usual demographics and risk factors were also obtained, including the nutritional quality of the participants' meals – using the HEI, the healthy eating index.

The Results

Participants were followed for an average of eight years or more, and during this period, there were 511 cardiovascular deaths and 638 deaths from cancer. 

• Compared to those who dined out less than once a week, individuals dining out two or more times per day were 50% more likely to die.
• The risk was 18% greater for cardiovascular deaths and 67% greater for cancer-related deaths.

Those numbers sound bad, maybe keeping the restaurants closed would be good for us. Just kidding! But those numbers are the empirically derived findings. I do not doubt the authors’ data collection or calculations. But before we abandon Shake Shack or Popeye’s Louisiana Kitchen, let’s consider Dr. Du’s words, especially “significantly associated.”

The Limitations

The discussion of limitation always begins with the study’s strength. In this case, 

“The major strength of this study was the use of data from a large and nationally representative sample of US adults, which facilitates the generalizability of the findings to the general US adult population.”

It is a large study, so the statistical analysis is adequately powered, no need to p-hack for a result. But several of the limitations question the validity of data. Specifically, the diet and dining out information is both self-reported and only known at the baseline. What the participants ate or how often they dined out over the next median eight years is unknown. I believe the researchers' data collection, I just have significant reservations of whether that data reflects eating habits throughout a minimum of 8 and a maximum of 15 years. Ask yourself, leaving 2020 aside, have your eating habits changed since 2000? Besides, as the authors point out, they have no information on the meals eaten outside the household other than they were consumed. 

It is the underlying ambiguity of the empiric data they collected, based on their assumption that this was reflective over time, where I beg to differ. And with that link in the chain of association weakened or broken, how accurate can the description of a “significant association” be?

The Discussion

In the study’s discussion, the authors offer some possible underlying reasons for their findings. While they bolster their thoughts by referencing other studies, these are not empirical facts; they are opinions. Opinions are subject to all our human biases. Their thoughts are the story they tell, to themselves and us, to tie the information together. It may make perfect sense, but it is a story.

Among their considerations, that food prepared at home is more nutritious than that dining out. (I would not like to compare some of the meals made by my mother to those of Eric Ripert of Le Bernardin). Of course, those having to take nearly every meal outside the home may well have stressful jobs, which also have been shown to increase mortality. The one that most caught my attention was this, 

“… a recent study found that people who dine out had higher levels of toxic chemicals, including phthalates, than those who ate home-cooked meals.”

Right up front, let me say, I have no idea whether this is correct or not, don’t write to tell me I am either chemophobic or chemophilic. But I did look at the study they cited to support those words. A third of participants in the cited work were under age 20, an age group not included in the current study, so there is a bit of apple-orange comparison. And phthalate levels, converted to androgen-disruptor units [1], were actually measured with calculations based on what the participant had actually eaten the day before. By the way, much of the elevated levels were among the younger participants, where dining out meant the school cafeteria. So maybe Union Square Café is OK, but my local elementary school cafeteria is the problem. By the way, the most significant source of phthalates in the study, sandwiches, which evidently includes cheeseburgers. (As compared to bologna, on white bread, with mayonnaise, that I was given as a child)

Here is my take-away. There may be an association between dining out and mortality, but this study doesn’t strengthen or create any linkage. The scientific data is sound only if you accept the underlying assumption that your eating habits do not change over time. And without that, any conclusion is unjustified. 

[1] Do you believe that describing phthalate levels as androgen-disruptor levels reflects a bias in the study?

Source: Association Between Frequency of Eating Away-From-Home Meals and Risk of All Cause and Cause-Specific Mortality Journal of the Academy of Nutrition and Dietetics DOI:10.1016/j.jand.2021.01.012