4 ms·
This is just based on pure nutritional content of underlying ingredients. For example, a fruit and vegetable smoothie scored 98 out of 100, with 100 being the
by sudden_dystopia 4y ago
This is just based on pure nutritional content of underlying ingredients. For example, a fruit and vegetable smoothie scored 98 out of 100, with 100 being the healthiest. But I don’t see that they accounted for how your body interprets a beverage vs a food. Your body understands what to do if you eat an orange. But if you drink a glass of orange juice, which contains the juice of multiple oranges, does your body really understand that context without the fiber and need for a different digestive process? And wouldn’t drinking multiple oranges spike your insulin since it will all hit your system at once due to lack of need for solids digestion compared to eating 5 oranges? Wouldn’t this be just like taking meds and supplements with food is more efficient as it provides context for your body as to what to do with the incoming molecules?
- s0rce 4y agoA smoothie should have all the fiber (unless you use juice), assuming its just blended. Isn't it just basically pre-chewed? Although chewing does do stuff as you have enzymes in your mouth.
- wyre 4y agoThey only difference between eating and drinking our foods (assuming nutritionally equal) is that chewing stimulates the enzymes in the stomach for digestion. Some meds and supplements need to be eaten with food because they are fat soluble. Afaik the stomach don’t need ‘context’ it take all the nutrients it can and excretes the rest.
- elil17 4y agoThe model does look at the "context." It looks at things like NOVA class, specific lipid composition, fiber:carb ratio, etc.