Food Groups in the National Nutrition Survey II
For the nutrition guidelines of the German Nutrition Society (DGE) , “closeness to usual consumption” is the most important criterion. That consumption was determined on the basis of the National Nutrition Survey II (Nationale Verzehrsstudie II, NVS II) of 2008.
Now, recording food statistically is difficult. Anyone who has ever tried to track their own food intake meticulously will know what I am talking about: what, exactly, is “buttered bread”? I know people who spread the butter so thinly that you can hardly tell which side of the slice was buttered. Others cut the butter off the block and lay it on the bread as a slab - the bread is basically a handle, so that the butter doesn’t melt in your hands.
It is a Herculean task to do all of this at population level and to process it statistically. The Max Rubner-Institut (MRI) and its collaboration partners took it on, and in 2008, based on the data collected, published the latest National Nutrition Survey to date - the foundation of the DGE nutrition recommendations.
The two - the NVS II and the DGE - refer to each other, and this is where it gets tricky. I don’t know exactly whom my criticism is aimed at: in the end it comes down to circular reasoning: <rant>The DGE says how we should eat, then the NVS II, with reference to the DGE, says how we actually eat, followed by the DGE saying: we should change that as little as possible. And that is how science is done</rant>.
Since I eat completely differently from what the DGE guidelines advise, I regularly find myself having to explain why I disregard them. Some of the arguments touch on ethics, others on health, others on feasibility or even politics. Scientific studies are cited again and again.
Today is about the foundations of those studies: what struck me when I looked at the compilation of the data that the DGE uses as the basis for its recommendation. In particular, it is about the wording in the summaries of the food groups, which strikes me in several places as a little too simplified.
Not every thought fits into one article, and by and large I am impressed by the work: as I said, it is a Herculean task. Nevertheless, I notice issues that I would like to discuss.
This is not a scientific paper, and I am quite deliberately doing some cherry-picking - today in part 2 of the final report of the NVS II(1). I am not covering part 1, because it mainly describes the study design, the socio-economic structure, knowledge and other results not directly related to the nutritional value of food.
And in case a passage seems familiar: yes, I already had this same document in my hands in Milking the Data. Back then I only skimmed off the cream, today you get the whole jug. Full-fat, naturally.
Business as usual is the most important aspect of the DGE nutrition guidelines
What Does the DGE Recommend? Chapter 1
Behind the DGE dietary guidelines: scientific foundations and the choice of goals and solution spaces. How are the recommendations weighted?
Read more →General state of health, Chapter 2.13
I find this section interesting because the DGE explicitly addresses its guidelines to healthy people. One aspect that stands out: self-reported state of health, divided into 5 levels. First it says “76% of all participants gave the highest ratings, very good and good, for their state of health” - the table below shows us that only 17.5% chose “very good”, 58.5% “only” good - so they felt there was room for improvement. As for the reasons, and where the line between “healthy” and “not healthy” runs, one can only speculate on such a one-dimensional scale.
The following paragraph then turns to age and ends with “unsurprisingly, the number of responses for the ratings ‘mediocre’, ‘poor’ and ‘very poor’ increases with age for both sexes”.
Here, the MRI describes the situation and self-assessment of the participants whose diets are recorded in the study. The DGE derives its recommendations for healthy people from the dietary data collected from people with quite mixed health.
The DGE nutrition guidelines are made for healthy people
DGE Guidelines. A Contradiction?
Why the recommendations of the German Nutrition Society don’t actually contradict my completely different personal experience
Read more →Food consumption of the German population by sex and age group
Chapter 4 makes me curious - because depending on where I read about the NVS II, prepared foods are either broken down into their components or summarised in categories by their main ingredient. Here it says:
The recipes mentioned during the survey were assigned to the various food groups according to their main ingredient (e.g. dishes based on cereals and cereal products)
Following on from my article about hamburgers, I ask myself: is a hamburger a meat dish, because of the patty? Or a dish based on grains or bread, because it is, after all, a kind of sandwich?
The table in the appendix leaves this open as well - it also offers dishes based on bread, e.g. sandwiches, club sandwich with tuna, Toast Hawaii, bread dumpling (Serviettenkloß). For dishes based on meat, meatballs (Frikadelle), pork schnitzel and goulash are named as examples.
Spoiler: the answer comes later in this article.
What really took my fancy, though, are the subsections here, which approach the descriptions of the various food groups in very different ways - some soberly describe the group and its share of consumption, others recommend it warmly and mention the DGE recommendations while doing so, and yet others praise its exceptionally high nutritional value and bioavailability, without making a recommendation for it.
This is also where it gets blurry whether my criticism is aimed at the MRI or at the DGE. The MRI wrote it, but in a context that was set by the DGE and later reused by it.
Curious? Let’s look at them separately:
Bread and cereal products (baked goods), Chapter 4.1.1
The food groups bread, baked goods, cereals/cereal products and the dishes made from them are among the most important in terms of quantity consumed. For men, at more than 300 g/day they are indeed the most consumed groups, while for women they are roughly comparable to the consumption of the food groups vegetables, fruit, and milk and cheese
We note: cereals are a main component of the diet; the NVS II makes no statements in this chapter about their nutritional value or their general worth in the diet, issues no warning and no recommendation, but merely soberly describes their share of the diet.
That is close to what I expected: after all, this is about describing the foods that are part of the food group.
The chapter on cereals is one of those that have no reference to recommendations and guidelines and say nothing either positive or negative about this food. The soberest description, very neutral. (That was a compliment.)
Alternatively, I offer my own views and findings here as a cross-reference:
Vegetables, mushrooms, legumes, Chapter 4.1.2
Vegetables are an important part of a healthy and balanced diet. Vegetables have a high content of vitamins, minerals and secondary plant compounds. They are also rich in fibre and low in fat. As a general rule, vegetables, like fruit, should be on the menu several times a day. The DGE advocates eating three portions of vegetables (400 g/day)
The quality of the description of this food group is already completely different from that of cereals. Not only are nutrients named that one can possibly obtain from these foods: the DGE recommendation is explicitly mentioned too, and another food group that is entirely beside the point here is brought in as well: fruit.
That strikes me as a politically inserted paragraph, because this study is about the figures for current consumption, not about communicating and repeating the guideline recommendations.
I like to poke fun at the term “fruitsandvegetables”, which is used in everyday speech as if fruits were just vegetables with better PR. In a scientific work that describes food groups explicitly, though, “fruit” is out of place when speaking about vegetables.
Potatoes, Chapter 4.1.3
The potato is one of the typical German staple foods. It supplies above all large amounts of starch and high-quality protein, and is a main source of vitamin C. The food group potatoes also includes potato products such as chips (French fries), potato pancakes and mashed potato, but not crisps.
Other articles on this site deal with my opinion on starch, so I won’t go into it further here. What slightly baffles me is the praise of vitamin C right next to the various foods in which most of it is lost through processing.
I have never seen anyone at the chip shop with the intention of raising their vitamin C level. Likewise, protein has never been the focus of anyone eating potatoes. No wonder: at 2 g of protein per 100 g of (boiled) potato, you would have to eat considerable quantities of potatoes - and the starch that comes with them - to get anywhere near covering your protein requirement. That is simply not what the potato is known for. (You can verify this in the calculator for essential amino acids.) The protein may be of high quality, but it only occurs in almost homeopathic doses: for each gram of protein you’d get 8 grams of carbohydrates.
Yes, yes, all right - I did warn you that I am cherry-picking. When such different foods are lumped together in one category, and then specific nutrients are praised on top of that, it practically invites it.
A potato is, ultimately, just sugar too
Sugar - What Exactly Is It?
A long answer to a trivial question - sugar is a great deal more than the stuff sold under that name in the supermarket. And it’s not necessarily sweet
Read more →Fruit and fruit products, Chapter 4.1.4
Here we have the group again that has already sneaked into the vegetables, with similar editorial content that goes far beyond describing and defining this group. What is not mentioned is also interesting:
A high consumption of fruit goes hand in hand with a high intake of vitamins (e.g. vitamin C, folic acid and β-carotene), minerals (e.g. potassium, calcium, phosphorus) and secondary plant compounds. This has many health benefits. The DGE advocates two portions of (fresh) fruit a day, which is 250 g. To make the recommendation easier to implement, one portion of fruit can be replaced by fruit juice. However, this should not be done every day
Not mentioned: sugar. Particularly since fruit juice is explicitly mentioned, which as a rule contains about as much sugar as soft drinks. Later on, the term “fruit juice/nectar” is also used explicitly - nectar is fruit juice diluted with water, to which sugar is added so that it doesn’t taste watered down. Confusingly, both will meet us again in a later chapter.
Nuts and seeds, Chapter 4.1.5
Botanically speaking, nuts also count as fruit. They do contain relatively a lot of fat, but because of their favourable fatty acid composition and their high content of fibre, vitamins and minerals they are to be classed as a valuable part of the diet. […] roasted and salted nuts and seeds [were] assigned to the snack foods […].
So snack foods are explicitly considered separately. That is interesting: I assume that this is because of the higher salt content, since otherwise a category “based on…” analogous to the other food groups would have been appropriate too.
A little surprising, however, is that in the appendix (Table A2) “roasted soybeans” is explicitly listed under “vegetables/legumes” as an example of a vegetable, although I know them as a snack as well.
Well, fine - that is not Germany’s favourite snack, so it hardly distorts the statistics - but the distinction seems arbitrary.
Fats, Chapter 4.1.6
First, fats are explained in general: animal vs. vegetable, saturated vs. unsaturated, etc. - it strikes me that this category is a motley assortment of substances - more a chemical class of substances than a food group. But if you classify fats by use and not by origin, it fits again. Nevertheless, a categorisation that, in my eyes, somehow feels out of place next to the other food groups:
This food group essentially covers fats that were used as spreads on bread. Fats that were used to prepare foods or dishes are included in the corresponding food groups or dishes.
And indeed, Table A2 in the appendix lists only the subgroups “butter” and “margarine”.
Now it doesn’t matter which camp you belong to, or whether you consider butter or margarine the work of the devil: I don’t know what this combination in one food group is meant to tell me. The next paragraph also lists the average men and women who eat both butter and margarine - logical, that is what an average is(2).
After the explanation of the vitamin content of fruit and potatoes, I miss the mention here that fats too (butter naturally, margarine when fortified) contain important vitamins, which fruit(3), by its nature, cannot contain because they are fat-soluble. Only with the fish chapter are they named.
Milk, dairy products and cheese, Chapter 4.1.7
Milk, dairy products and cheese are important suppliers of protein, vitamins and the mineral calcium
Further on, what strikes me is how variable this food group is:
Dishes based on milk and dairy products are essentially understood to mean rice pudding, ready-to-eat muesli or cereals with milk or yoghurt, […] and pudding[…]. The main ingredient of these dishes is milk[…]
Since flavoured milk drinks are listed as well, heavily sweetened confectionery is also lumped in here, because it is made from milk: not just fruit juice, chocolate milk too has a sugar content comparable to soft drinks, and in the supermarket, comparing the size of the shelf for (sweetened) fruit yoghurt with the one for plain yoghurt is very revealing.
So one could in good conscience file some of the dairy products under confectionery. Here, though: chocolate milk is cheese(4).
Eggs, Chapter 4.1.8
Eggs are of essential importance in the diet and cuisine of many countries because of their nutrient content and their positive properties (e.g. emulsifying effect). […] The protein in hen’s eggs has a high biological value and, in combination with other foods such as potatoes, can thus raise the overall protein quality of a meal
The description of the nutrient content of eggs that has stuck in my memory most vividly is this: “By definition, an egg contains all the nutrients necessary for constructing a complete animal organism”. That is a little more dramatic - the study puts it more soberly.
This sober account is - actually - very on-brand for the purpose of this study. But since the DGE’s recommendations have already found their way in for vegetables - there with a specific quantity recommendation - it is striking that nothing is said here about the DGE at the time recommending an upper limit of three eggs a week for this food. Unlike today (2024 upper limit: one egg a week), though, that included all eggs processed into other dishes. Going by that, eggs must be dangerous - which does not emerge from this description (“high biological value”)(5).
Ironically, then, the NVS II recommends pairing potatoes with precisely the food that the DGE strictly limits. The potato on its own already gets praised - with egg, it gets no blessing from the DGE. 🤷
Meat, sausages and meat products, Chapter 4.1.9
Meat is an important source of B vitamins (B6, B12, niacin) as well as iron, zinc, selenium and potassium. In addition, the bioavailability of the minerals in meat and meat products is higher than in cereals and other plant foods. Furthermore, meat contains high-quality protein and thus contributes to the protein supply.
All of this is a very appreciative description of a food that is worth considering for the diet.
If one were to continue this sentence with the DGE recommendation - as with vegetables - it would read: “…the DGE recommends limiting consumption of this highly nutrient-dense food to 43 g per day"(6) - so little that the amount is actually given per week, because the daily amount is so laughably low. In fact, the DGE guideline allows or accepts more free sugars than meat per week(7) - and that despite the high praise sung here by the NVS II for meat’s nutrient density and bioavailability.
Just as with eggs, the DGE sets an upper limit for meat as well, and thereby implicitly presents this food, described so positively, as dangerous.
Later in the chapter, a question I asked earlier gets answered:
Dishes based on meat are essentially understood to mean sausage salads, meat-containing fast food such as hamburgers or meat with sauce.
The hypervegetarian(8) dish “hamburger” (see the article linked above) is therefore not a dish based on bread, but one based on meat. I fear, then, that - in case hamburgers should turn out not to be healthy - meat gets the blame, and not the chips, the sugary sauce, the fizzy drink or the bun that typically go with it.
This paragraph gives no information on how that ultimately affects the results, or on whether the ingredients are nevertheless properly broken down.
Fish, Chapter 4.1.10
The description of fish sounds similar to that of meat: all great, highly available and high-quality - and the direct conclusion is the same as for meat: going by the summary of this category, fish seems eminently advisable to eat.
Despite these apparently similar properties of fish and meat, the DGE issues a recommendation for fish, but sets an upper limit for meat. This contrast is not addressed, although the description gives no reasons at all for avoiding either of the two food groups.
Incidentally, this is the first place where fat-soluble vitamins are named:
According to their habitat, fish can be divided into sea and freshwater fish, or, according to their fat content, into fatty and lean fish. Fish is particularly rich in valuable polyunsaturated fatty acids. Fatty fish in particular is rich in n-3 fatty acids. In addition, fish provides biologically high-quality and easily digestible protein. Fatty fish also provide the vitamins A and D in particular, and sea fish the mineral iodine in particular.
Soups and stews, Chapter 4.1.11
In this section we now enter the realm of culture and cooking methods. The examples listed show that this group seems to be entirely independent of the ingredients and is distinguished only by the high water content that all of these dishes share:
Soup is one of the most traditional and best-loved foods in Germany. Every region has its own soup culture; in Bavaria, for example, liver dumpling soup is popular. Classic German soups include bread soup, pea soup, potato soup and lentil stew.
The nutrient profile of all these examples is likely to be extremely different. It is a mystery to me why a pea soup is not a “dish based on legumes”, and a bread soup not a “dish based on bread/cereals”.
Don’t worry, I won’t suggest a new category “dishes based on water”.
Sauces and seasonings, Chapter 4.1.12
Another motley assortment of things that could not be more different:
The food group sauces and seasoning ingredients includes, besides hot and cold sauces, condiments such as mustard, vinegar, soy sauce or dried spices (appendix Table A.2). Condiments in particular are usually used only in very small quantities
Incidentally, consumption averages 24 (women) to 33 (men) grams per day - when sauce and dried spices are lumped together at this order of magnitude, I honestly don’t know what the food group is supposed to tell us either.
Confectionery, Chapter 4.1.13
Under the heading confectionery, besides sweets, ice cream, sweet spreads and sweeteners are also included.
Sweeteners stand out: by that, I take it they mean artificial sweeteners, which are needed in extremely small quantities because their sweetening power is considerably higher than that of sugar (which can also be described as a sweetener). Here, “milligram ingredients” are measured together with ones that are sold in kilogram packs. A rounding error for a substance with a much-debated effect on the organism.
What is especially interesting, though, is another statement that my mother, at least, never made: “Don’t eat so many sweets, they contain too much fat”:
Confectionery contains a lot of sugar and in some cases also a lot of fat, which is why it should only be eaten in moderation.
Snack foods, Chapter 4.1.14
The food group of snack foods essentially comprises potato-based snacks, salty small baked goods, salted and roasted nuts and seeds, as well as peanut puffs, popcorn and the like. Snack foods are mostly products that are relatively rich in fat and salt and should be eaten only occasionally and in smaller quantities
I already went along with the line drawn between nuts and seeds and snack foods earlier, and asked why roasted soybeans are explicitly not part of it here.
I notice: here too, consumption recommendations tag along again, this time in the form of the recommended limit. I would like to see the popular lumping together of “salt and fat” broken up, just like the aforementioned “fruitsandvegetables” or “sugar and fat” for confectionery: the fat content of roasted nuts is not substantially different from that of unroasted nuts. The fat quality may well differ markedly, but that is not at all covered by the word “rich in fat”.
Drinks, Chapter 4.1.15
I’ll look only at the non-alcoholic drinks:
The food group of non-alcoholic drinks includes water, coffee and tea, but also fruit and vegetable juices/nectars, lemonades or fruit juice drinks. Thanks to their 100 per cent fruit content, fruit juices are to be rated higher in quality than nectars. For fruit nectar the minimum fruit content is 50%, and up to 20% sugar may be added. Lemonades and fizzy drinks, by contrast, usually have a very high proportion of added sugar. The DGE recommends drinking 1.5 litres of fluid daily, with water and low-calorie drinks to be preferred.
Fruit juices and nectars were already mentioned under fruit, so I am baffled by their separate listing here. A slight correction, too: depending on the fruit, fruit nectar may contain as little as 25% fruit. Fruit juice drinks are listed, but their regulation is not described separately - I would call them sugar water: the required fruit content drops, depending on the type of fruit, to 6-30%, and flavourings and acidifiers are permitted - not to mention the added sugar.
Given all the DGE consumption recommendations contained in the other chapters, I am surprised by the restrained way in which sugar bombs like these are treated here, simply advising a preference for low-calorie drinks and leaving the rest of the exercise to the reader.
Now that I mention it, I’ll have to look up what the DGE says about low-calorie drinks sweetened with artificial sweeteners. And we now have another contender for the new category that I promised not to suggest: dishes based on water…
Incidentally, my bafflement at the double listing of fruit juices cleared up on rereading - under “fruit” it says “excluding juice”, but replacing part of it with juice is allowed. Something new comes up, though: the DGE allows part of the fruit to be replaced by fruit juice, while the NVS II explicitly includes nectar as well. It remains unclear whether the two are talking about the same thing.
Other, Chapter 4.1.16
Under “Other” we find a rather creepy assortment:
Some foods could not be assigned to the aforementioned groups. These include soy products, vegetarian spreads and desserts with unassignable ingredients. The mean consumption of these foods is, however, very low (soy products 2-3 g/day, vegetarian spreads and other sweet dishes <1 g/day).
In addition, protein and energy supplements as well as formula diets were recorded. For 67 men and 30 women there is information on products from this group. While for the men these are predominantly mixed drinks based on protein powder, more than half of the women take formula diets.
The creepiest component, in my eyes, formula diets, needs to be put in perspective, though. “More than half of the women” in this case means “more than 15”: 8,278 women and 7,093 men took part in the NVS II survey.
Slightly baffling, as a finishing touch, is the mention of soy products here, after legumes as such were already included in the chapter on vegetables. Could it mean heavily processed soy? But then, we know that legumes have to be processed to a greater or lesser degree before they are fit to eat.
The necessary processing of legumes, using kidney beans as an example
Superfood or Antinutrient?
Foods praised for their ingredients often bring the antidote along too. How do antinutrients work? Can they be rendered harmless? A selection.
Read more →Conclusion
When I look at this list in relation to my own ketogenic diet, I wonder how my data would have been sorted: the ketogenic diet is very high in fat and low in carbohydrates. Even so, the fats category might well be empty for me: because I don’t eat bread (remember: the handle for the butter), I simply use no spreading fat - but I don’t skimp on butter when I make an omelette.
What bothers me most, though, are the selective circular references:
The NVS II describes current consumption. Claims and recommendations of the DGE have no business there to begin with - or should be classified consistently as description. The contrast between the recommended amount for vegetables and the upper limit for meat and eggs, set against the description of these categories, is astonishing(9).
So is the DGE’s contrast between meat and fish, which does not emerge at all from the NVS II description. The authors involved obviously knew the DGE guidelines, but either nobody noticed the difference, or it wasn’t worth mentioning to them.
Since I consider the difference decisive, this does not match my expectations. If you’re going to argue in circles, at least do it everywhere. Better still: nowhere.
As a closing word, what stays with me is the bitter - or bittersweet - taste of sorting things into more or less capricious food groups. Perhaps best summed up as:
Chocolate milk is cheese(4).
