X
Business

Gross conclusions come from gross statistics

One thing health IT can do, over the next decade, is come up with answers to these questions. We can collect data on what people actually eat less intrusively. We can collect more data on the impact of food on the body as that impact occurs.
Written by Dana Blankenhorn, Inactive

Let's return briefly to the story from last week which drew the most heat here, the meat will kill you story.

(Wikimedia got this picture of wurst from Tuppers at Flickr.)

This is an example of drawing a big conclusion based on big numbers.

There were about 500,000 people in the survey. Those who self-reported eating lots of cows and pigs had a higher death rate than those who claimed to eat just chicken and fish. Thus the conclusion, stay away from moo-cows and piggies.

But, as many commenters here asked, did they really account for all the variables?

For the conclusion to be accurate one would assume the two groups -- meaties and fishies -- were in other respects very similar.

But are they?

Think about it. These are self-reported descriptions of behavior. Are those who claim to be meat lovers living the same way as those who claim to be fish lovers?

Upon further review, I doubt it.

Folks who acknowledge eating a lot of beef and pork are also more likely to be sedentary, more likely to be overweight, and less likely to be paying attention to what they eat than those who say they eat just fish and chicken.

I did not say that liking steak makes you fat. But saying you like it may indicate a desire to tie on the old feed bag in front of big portions and chew happily.

On the other hand, if you say you eat a lot of fish and chicken, you may say a lot of other things that we would call health conscious. That you're walking where possible, that you're watching your calories, that you read a lot of health stories like this one which say eating fish and chicken makes you healthier.

Again, these are self-reported behaviors. The answer to the "beef or fish" question may indicate a lot of attitudes that have nothing to do with eating habits.

One thing health IT can do, over the next decade, is come up with answers to these questions. We can collect data on what people actually eat less intrusively. We can collect more data on the impact of food on the body as that impact occurs.

We need more study, and better health monitoring is going to give us better results. So if you had a steakburger last night you don't have to buy a cemetary plot today. Your friend the fish lover might just choke on a bone tomorrow.

Editorial standards