We need better nutrition studies. Here’s one solution

Pick a food (coffee? eggs?) and think of all the times you've heard there's research saying it’s good or bad or good again. Critics say nutrition science suffers from epidemiology based on people’s faulty recall of what they ate or how much. So what to do? Kevin Hall of the National Institute of Diabetes and Digestive and Kidney Diseases suggests updating World War II-era “domiciled feeding studies.” House people comfortably in centralized research facilities, feed them study food for weeks or months, and gather teams of scientists to study complex interactions between diet changes, the microbiome, and physiology, for example (think biomarkers for disease risk). Then take what’s learned to test in the real world. Food for thought.