Life is Full of Worries

I've learned to not pay too much attention to "scientific studies". They warn you of dangers that would fill you with dread and send you scurrying to hide in the deepest darkest hole you could find. They warn you that eggs will kill you because of cholesterol then they tell you that if you don't eat enough eggs you will die because your tissues won't be plump and nice. They tell you to stop eating butter because it will kill you, then thirty years later they report that all those who fled butter for margarine had exposed themselves to unnecessary health hazards because of the petroleum products in margarine. Even too much water can kill you.

And here is a classic "latest research" where researchers reach polar opposite results (from an article in Seed magazine):

Two studies by Martin Voracek seem to uphold the notion that more intelligent people are more likely to commit suicide. Voracek looked at national suicide rates and average IQ, and found that countries with higher average IQs also had higher suicide rates. But a study released last week suggested the opposite might be true.

Even when there are real risks uncovered in studies, you don't always know about them. Years ago I was prescribed Vioxx by a doctor who probably wasn't told of the many side-effects of the drug. The drug companies pushed the drug because they could sell it at a high price and make a killing. In my case, aspirin -- very inexpensive -- probably should have been the drug of choice, but that wasn't what the doctor "prescribed".

In short, there are health hazards out there and if you take them seriously you will be paralyzed by fear. In short, a little ignorance helps keep us sane.

Now that you have digested the warnings... Here's the latest "research" report:

Obesity linked to brain shrinkage and dementia

Posted on: June 16, 2010 4:00 PM, by Mo

THE dangers of obesity are very well known. Being overweight is associated with an increased risk of coronary heart disease and stroke, the two leading causes of death in the Western world. Gout is more common in overweight people, with the risk of developing the condition increasing in parallel with body weight. Obese people are twice as likely to develop type 2 diabetes as those who are not overweight, and being overweight is also associated with several types of cancer. The list goes on...

Less well known is the effect of obesity on the brain. In the past few years, however, it has emerged that being overweight in middle age is linked to an increased risk of Alzheimer's Disease and other forms of dementia. Two new studies strengthen this association: the first, just published in the Annals of Neurology, shows that abdominal fat is linked to reduced brain volume in otherwise healthy middle-aged adults. The second, published last month in Proceedings of the National Academy of Sciences, shows that this reduction is associated with a common variant of an obesity-related gene.

There's more, go read the original article.

I don't doubt there is a link between obesity and dementia. But this kind of report bothers me because it doesn't come with enough numbers to put the "warning" into context. First of all, we know that aging increases the risk of dementia. We know that aging is associated with "middle age spread", i.e. most of us put on "a few extra pounds" as we age. Which came first, dementia because of the additional pounds or dementia because of aging?

I would be willing to bet that those with dementia are less physically active than those without, so I'm willing to bet that dementia causes obesity. But this report is based on a correlation between dementia and obesity and, as far as I can tell, has no theory to identify a causal direction. So the report can easily have mistaken correlation for causation. Issuing a breathless news release about the "connection" between obesity and dementia is not only foolish, it causes unnecessary harm by getting people worried that don't need to be.

If you go read the original posting you find that the study behind this conclusion watched 733 people. I don't know the actual rate of dementia, but I would assume it is something like 5% of the population over an entire lifetime. So watching 733 people for 5 years means that "on average" you would see 2 people with dementia. I may not be an expert in statistics, but doing a study where you expect to see 2 cases and discover 3 cases sounds terrible is you say "risks elevated by 50%". But a more honest assessment would be "having 2 or 3 cases is statistically insignificant".

In other words, I would get excited about this "latest study" if it had reviewed 40,000 cases where you would expect to see 100 dementia cases and if they discovered 120, then that would be statistically significant. But even then, I would be cautious. I'm not an expert in statistics, but this warning from the Wikipedia article on statistical significance is relevant:

A common misconception is that a statistically significant result is always of practical significance, or demonstrates a large effect in the population. Unfortunately, this problem is commonly encountered in scientific writing.[5] Given a sufficiently large sample, extremely small and non-notable differences can be found to be statistically significant, and statistical significance says nothing about the practical significance of a difference.

In short, I will use my usual rule about breathless reporting of the "latest scientific finding": Roll over and go back to sleep. It probably isn't significant. If it really, really is important, then I will hear about it a context other than "the latest scientific report" coming from the media (or a blog).