How can nutritional science be made more accurate?

[sanantone]

I'm used to the social sciences having a lot of grey areas because human behavior is complex, but I expected nutrition to be more easily understood because it's based on physiology and biochemistry. I think the problem with nutrition, as with the social sciences, is complexity and individual differences. In my medical sciences program, we learned that looking at stomach contents is a highly inaccurate way of estimating the postmortem interval (time since death) because people digest food at different rates, and different foods break down at different rates. 

As I was listening to an NPR program the other day on diets, I realized that I had to throw everything I knew about nutrition out the window. We've learned that the problems fats and cholesterol cause have been overstated. We've learned that sugars, simple and complex, are a bigger problem than fats. The calories in/calories out advice given to lose weight is bad advice. Different people eating different foods at different times will retain different amounts of calories. A study even found that eating day-old, leftover pasta results in less calorie retention than eating fresh pasta even though the person is eating the same number of calories.

I think these problems are seen all throughout the medical sciences, actually. A lot of medications that were seen as safe have been later found to be dangerous. Not too long ago, I had a doctor recommend melatonin for my sleeping problems, and I told him that I was already taking it, and it wasn't really helping. He was shocked. I was shocked that he was shocked. Melatonin is simply a hormone that regulates your body's clock; it tells your body that it's time to go to sleep. If you have trouble staying asleep, then melatonin may not help. You'd be surprised at how often doctors are following trendy news headlines as much as we do. I won't gross you out with all of the grey areas I've learned about in pathology.