One of the arguments used by those of us who are critical of Nutriscore is that it can be “a strainer” for a large number of products with a dubious relationship with health, which allows them to obtain ratings that the consumer associates with benefits (green colors and A categories and B). I explained the reasons for these results in this previous post and it is enough to take a critical look around the supermarket to relatively easily find striking examples of good Nutriscore evaluations, as can be seen, photos included, in this post or this other one.
Certainly going “on the hunt” in the supermarket is not a very scientific method of testing the rigor of a front-end labeling (FOPL) system, so it is important that there are studies that systematically look at the relationship between the Nutriscore rating and the quality of food or its relation to health. But using complementary and independent criteria to the Nutriscore system itself, something that is not usually done, since some of the studies that are usually cited in this sense (example) have used the FSA Score to measure the quality of the diet. That is, the same criteria as those used to develop the Nutriscore algorithm. And in addition, these studies usually take into account the entire diet, including fresh foods and not just processed ones. So its results cannot be considered too significant, as I explained in this previous post.
To try to clarify this panorama a bit, an interesting study has just been published analyzing the relationship between Nutriscore and NOVA, the food classification system based on its degree of processing (which I explained here). Although they are systems with different criteria – nutritional composition vs degree of processing – this comparison is very interesting for two reasons: First, because the Nutriscore label is only applicable to processed foods. And second, because there is increasing evidence that ultra-processed foods (NOVA level 4) are associated with poorer health (1,2,3,4,5).
This research, carried out by Spanish scientists and entitled “Two Dimensions of Nutritional Value: Nutri-Score and NOVA” (2021), has compared the data of thousands of products from Open Food Facts, the database contains information necessary to classify food through both systems. And it has obtained a few interesting results, all of them accessible in the original document, which is freely accessible.
I personally think that the most interesting result is the one that analyzes the percentage of foods of each level of NOVA processing that are in each Nutriscore group, which is shown in the following table:
The researchers include a figure (somewhat more complex to interpret) with the same results:
And if we focus on the ultra-processed (NOVA 4, right column of the table) we can graphically represent the percentage in each Nutriscore group:
As can be seen, uNot out of every four group A products (the best Nutriscore score) are ultra-processed. And half of those rated B (Nutriscore’s second best rating) are also ultra-processed.
Based on all the results, the authors include the following paragraph at the end of their paper:
“Food processing and the nutritional quality of food encompass different but complementary dimensions. All Nutri-Score categories include ultra-processed foods (…). Therefore, the information provided by the Nutri-Score is incomplete and does not identify all unhealthy foods…”
It should be pointed out that some of Nutricore’s defenders lately tend to insist that this system is not used to know if a food is healthy or not. But I don’t think it’s necessary to explain how the average consumer will interpret a green A or B, right?