Study proves it's not unhealthy to be overweight? Nah.

Researchers at the University of Manitoba have released a study on the relationship between body mass and health problems. Here's a typical mass-media report on the findings, headlined, "Overweight people don’t have bigger health problems, study finds."

It's an epidemiological study, which means researchers looked back at a bunch of historical data about people and then tried to draw associations between those people with certain characteristics or behaviors, and how they fared over time. Epidemiological studies are much cheaper and easier and usually quicker to do than almost any other type of study, and they seem impressive because they often involve large numbers of people. But they also are some of the least reliable studies, being typically plagued with just about every problem that a study can have, including confounders (what happens to people in the study may have little or nothing to do with the causes that are being looked at), confusion of cause and effect (what happens to people in the study may actually be causing the factors that are being studied, rather than being caused by them), unrepresentative samples (the people being studied may not be typical of the population we care about), trying to draw conclusions from small differences, and more. Most studies about the relationship between health and weight we hear about are epidemiological studies, and we've had ample evidence over time that these are some of the least reliable among generally unreliable epidemiological studies.

But there may be something else to worry about with the conclusions of this Manitoba study, something that makes me wonder if the findings are actually right--but not for the reasons implied in reports of the study. 

Last year I met with Oberlin College biologist Keith Tarvin, who studies foraging behavior in animals. Tarvin explained to me that throughout most of the animal kingdom, animals will generally put on excess weight when provided with plenty of calorie-rich food and not given any reason to be physically active, such as having to forage or hunt for food, or having to avoid predators. (Animals that need to keep weight down in order to survive are sometimes an exception--birds, for example, can't fly if they get too heavy.) Situations where animals lose weight, on the other hand, are usually associated with some pathological state or threat--as for example when fish keep weight off in order stay small to be less visible to and appealing to predators, or to slow their maturation in order to preserve fertility and lower energy needs during times of drought, extreme temperatures, food shortages or other environmental pressures.

In other words--and at this point I'm merely expressing a conjecture that came up in my chat with Tarvin---it may be that, weight-related health factors aside, having a stronger appetite is associated with being well-suited to the environment. Or to put it differently, keeping weight off when there is plenty of rich food around and no pressing reason to be physically active is, in a sense, a pathological state. Apply this theory to human society today--and at this point I'm degenerating into my own conjecture--and we might well predict that being overweight is associated with being, in some ways at least, healthier than those who keep weight off.

Does this mean we should stop worrying about being overweight? Absolutely not. To conclude as much would be confusing cause and effect. If this conjecture is right, it means that being overweight is in some ways an effect of being healthy; it doesn't mean that being overweight confers any health benefits. And in fact, the possible rightness of this conjecture should have no bearing whatsoever on the well-established fact that carrying excess fat is generally unhealthy, and that people improve their health and lower their health risks when they lose excess fat.

Even if it's true that people who are overweight are in many cases healthier in some ways than many people who are not overweight, these lucky overweight would still on average even further improve their health by losing the excess weight, and those who are not overweight would become less healthy by putting on excess weight. That doesn't mean that weight loss is a pathological state, even though it's usually associated with one in the animal kingdom. We should understand that the loss of excess weight is a good thing, even if, in a way, it is an unnatural thing in our society of plentiful, overly stimulating, calorie-dense food and sedentary lifestyles. That's why our goal should be, in effect, to change what's natural in our society--so that people are pushed by everything they see around them to avoiding excess weight, rather than being pushed to consume rich food and avoid physical activity as they are today.

To have even a decent chance of proving that losing excess weight will make you healthier--or to solidly prove just about any theory about the relationship between excess weight and health--we'd need a randomized controlled trial in which people are randomly assigned to either become fatter than they are, or lose weight, or maintain their weight. That trial is never going to happen, for what I hope are obvious reasons. And this brings us back to the problems with epidemiological studies like the one done in Manitoba. They rarely answer the questions we really want to ask. Instead, they give us answers that only raise more questions. There's nothing wrong with that--it's how science operates. We should be grateful scientists are conducting these studies, and appreciate the fact that they add to our knowledge base. But we should also be aware of the extreme limitations and qualifications that attach to the findings.

In the case of the Manitoba study, I think it's fair to say the study at best tells us little about the healthfulness of avoiding being overweight, and at worst is, at least in how it's being reported, extremely and even dangerously misleading. One of the biggest problems with the obesity crisis is the lack of awareness and motivation on the part of a sizable percentage of overweight people. When a study like this gets press claiming that being overweight seems to be as healthy as not being overweight, we take a big step backwards, and cause real damage that could in principle be measured in the loss of many years of life in the population, not to mention the drop in quality of life experienced on average by the overweight. I wish scientists and journalists would start taking these issues into account in their reports to journals and the mass media.