by Marion Nestle
Jan 1 2008

The first topic in 2008: food scoring systems

My first question in the new year is from “fretful reader” who asks: “Esteemed Wise Woman Whose Writing Lights A Fire Under Me: …today’s [San Francisco] Chronicle has a story about the ONQI(overall nutritional quality index)…which purports to ‘make nutrition easy’. My college education (about 30 yrs old, damn near antiquated) is
inadequate to the task of combining “positive nutrients” , “negative nutrients”, dividing them, and why didn’t they remember to subtract the number of ingredients on the list altogether…as a way of penalizing the ‘foods’ that have those scary long lists in a designed to be unreadable, vertical typeface? Does it sound like I’m irritable? Probably.”

Dear irritable, fretful: Me too. I’m not much for scoring systems of any kind on food. I don’t think you need a score to know whether you are eating a junk food or not and is a slightly better junk food better for you? I can’t remember who started these things but PepsiCo has its Smart Spots and Kraft has its Sensible Solutions and companies like those can set up their own criteria for what is and is not “healthier.” It’s a lot of fun to go to supermarkets and look to see which products qualify. Kraft’s Lunchables are a good place to start. See if you can tell the difference between products that do and do not qualify. Hannaford supermarkets got some independent nutrition researchers to develop criteria for awarding one , two, or three stars to healthier products and guess what: less than one quarter of nearly 30,000 products qualified for even one star, and most of those were fruits and vegetables in the produce section. So when the criteria are tough, hardly anything qualifies. So now Dr. David Katz at Yale has gotten a committee together to develop his own set. You have to have a degree in mathematics to understand it but that doesn’t really matter. Do you really need a scoring system to tell you that General Mills’ Wheaties (score: 246.2403) is better than Barbara’s Puffins Peanut Butter (9.937892) or Kellogg’s Cocoa Krispies (0.476746)? Never mind the apparent but misleading precision of the 4 to 6 decimal places. All of these are low scores. The problem with these systems is that the criteria are arbitrary and make some highly processed foods look better than others. This is a great marketing tool but will it help people eat more healthfully? I doubt it. I take an extreme position on all such systems. They should not be allowed. If we must have them, the FDA needs to step in and set up one set of criteria. And I don’t envy the committee that has to do that. So I am adding one more item to my list of “rules” for supermarket shopping in What to Eat. If it has a self-endorsement of nutritional quality, don’t buy it; such things are about marketing, not health.

  • Dear Marion, & ‘Fretful’-

    First, I want to note that Dr. Nestle and I are friends and colleagues of long-standing, and I comment here with her prior knowledge, and at her invitation. I both like and deeply respect Dr. Nestle. But with regard to nutrition scoring in general, and the Overall Nutritional Quality Index (which I invented) in particular, I disagree.

    First, Dr. Nestle mentions home-grown nutrition scoring systems used by food companies, and tosses them in with systems such as the ONQI, and Hannaford’s Guiding Stars. This is misleading, since an objective, universal, science-based system of nutritional scoring is specifically designed to overcome the various and serious shortfalls of food company systems:

    (1) Food company systems represent a conflict of interest, since those selling the food are the ones to say what is ‘better’ for you. (2) Food company systems are not universal- they apply only to the product line of the company in question. There is no way to compare a Pepsi Smart Spot to a Unilever Health Check, or a Kraft Sensible Solution. Yet that is exactly what a shopper looking for the best choice within a product category, not within a company’s inventory, would need to do.

    As for ‘Fretful’s’ perspective that we don’t need such a system: if she/he does not, she/he is a rare bird indeed! Of course no one needs nutrition scoring to know that produce is better for us than candy (although a system as robust as the ONQI does allow for the comparison of apples to oranges! For what it’s worth, oranges win- although both, of course, score very high). But consider this: when we used the Overall Nutritional Quality Index to score foods, we found that a cereal with 1/3 less sugar scored worse than its regular counterpart, because the modified cereal had less fiber, more salt, more saturated fat, and more calories (the banner ad on the front of the package did not mention those!…); we found that fat-reduced peanut butter scored worse than its regular counterpart, due largely to differences in added salt and sugar; we found that light mayonnaise scored worse than regular…and so on.

    Dr. Nestle appears to be saying these are all junk foods, so who cares? And Fretful seems to be saying: I don’t need your help. But consider this: who is the message “1/3 less sugar!” on the front of a cereal package for? It is for an average, but health-consious consumer. It is for a busy mom or dad who who care about what they feed their kids. And it is a message designed by smart people on Madison Avenue to deceive. It conveys the message “better for you,” when the product is not.

    I doubt that either Dr. Nestle or Fretful would consider a genuine multigrain bread, rich in fiber and nutrients, a “junk” food. Yet sitting on the bread aisle of every supermarket in the US are breads adorned with front-of-package banner ads that read “Multigrain,” “12-grain,” etc.- with images of wheat fields into the bargain. Many of these contain multiple REFINED grains! Yes, they are “muiltigrain,” but they are NOT whole grain, and not nearly as good for you as a far humbler, generic whole wheat bread you may overlook entirely. Again, the person being mislead here is not a person unconcerned about nutrition or looking for one variety of junk food vs. another. The person being duped is a health-conscious mom or dad, trying to do the right thing- but without a PhD in nutritional biochemistry, a marketing degree, or a lifetime to spend in the grocery store scrutinizing the fine print.

    There are, in my opinion, two ways to improve the quality of the diet: one is a pattern shift, and the other is one informed food choice at a time. A pattern shift would require substituting whole categories of food for others. So, for example, “eat more fruits and vegetables” falls in this domain. The advice is excellent, but the uptake is poor. Not to be too glib, but at our current rate of progress toward Healthy People 2010 objectives for fruit and vegetable intake in the United States, we’ll by lucky to get there in time for New Year’s Eve, 3010!

    So while we should push for more fruit and vegetable intake (and adopt policies to support the goal), we clearly need to supplement the pattern-shift approach. I maintain that dietary quality may be improved dramatically one informed food choice at a time. Choose genuinely better-for-you bread, breakfast cereal, pasta sauce, salad dressing, spreads, yogurt, granola bars, crackers, chips, etc.- and tally up the effects across the expanse of your shopping cart, and diet. You may achieve dramatic reductions in salt, sugar, trans fat, and saturated fat intake- and equally dramatic increases in fiber, healthful oils, vitamins, minerals, and antioxidants. You should, of course, still eat more fruits and vegetables, too! But everyone, not matter their intake of produce, eats some items that come out of bags, boxes, bottles, jars and cans. If you are informed and empowered to choose better nutrition every time, yes, it matters.

    Now, as for the ONQI: this system was developed by a dozen among the top nutrition scientists in North America, many of them as close to Dr. Nestle as to me (see for more information). The work was not for any industry interest. It was pure science, of the highest caliber, and took 15 people (three doctoral-level staff of mine supported the work of the panel) 2 years. There was nothing “arbitrary” about the ONQI science. The panel of experts carefully scrutinized and considered the research literature, with particular reliance on such sources as the Dietary Reference Intakes of the Institute of Medicine; the 2005 Dietary Guidelines for Americans; and so on.

    Is the ONQI complicated? Yes. In includes more than 30 nutrient entries, ranging from the obvious- sodium, calicum fiber- to the less so: omega-3 fatty acids; the glycemic load; the levels of 9 individual essential amino acids to allow for scoring of the biological quality of protein. But I the engine of your car is complicated, too. Yet it contains neither more, nor fewer, moving parts than are necessary to take you reliably from A to B. Under the hood, the ONQI is complicated, but neither more nor less so than required to represent the best science, and the best thinking of an international team of top nutrition scientists.

    But the system is turn-key, at-a-glance simple at point of purchase. The system will use a 1-100 scale, with 100 being the most nutritious food in the supermarket, and 1 the least (we currently have over 15,000 food items scors on this 1-100 scale). The raw scores at which Dr. Nestle pokes fun are not intended for use by the public- we simply posted them early on the web so that experts would get early, general information about the system. The ONQI scores in supermarkets will not have any decimal points, I assure you- and are, in fact, based on market testing that didn’t begin until the algorithm had been finalized.

    As for the concern raised about the system being proprietary: the ONQI was developed with no commercial interest in view. It was supported by the non-profit Griffin Hospital, in Derby, CT- and when finished, offered first to the US FDA. The FDA, while very supportive of the work and the sophistication of the ONQI, encouraged commercialization as the only way to get the ONQI into consumer hands any time soon.

    So now, yes, there are business interests involved. And there is intellectual property. But the fundamental workings of the ONQI algorithm were shared with scientists at a conference held in Washington, DC on 11/30/07 devoted to the purpose of transparency. Dr. Nestle was invited, but unable to attend. The ONQI will soon be published in the peer-reviewed literature as well. Not every last detail of the algorithm will be shared, but more than enough for those even with somewhat less expertise than Dr. Nestle to judge the reliability and robustness of the system. Of the 100 or so scientists who attended the November conference, not a one complained that they had insufficient detail to judge the algorithm.

    Nor will the algorithm ‘rest’ on its laurels. A team of scientists, entirely independent of the ONQI commercialization effort and with no financial interest in it, will oversee maintenance and updates of the algorithm going forward. Among other things, plans are in place to test the performance of the ONQI relative to health outcomes using data from large cohort studies at Havard. Nothing about the system is being taken for granted. If it can be improved over time, it will be.

    There is, finally, a problem on which we can all agree: there are too many such systems out there. But the ONQI- which will be launched in thousands of supermarkets throughout the US in 2008, and be available to all for free use on the Internet as well- is designed to be suitable as ‘the’ industry standard. But the only way to solve the ‘tower of Babel’ problem (too many concflicting claims about nutrition) is to contribute to it temporarily, by putting out one more system with the potential to ‘win’ over the others. Whether that proves to be the ONQI, or some other system- time will tell. But I strongly support a single, science-based system, and my own efforts are means to such ends.

    There is limited information about the ONQI at; we’ll be adding more soon. You can sign up there to get e-mail updates.

    With best wishes to all who happen by for a healthy, well-nourished 2008-

    David L. Katz, MD, MPH, FACPM, FACP
    Director, Prevention Research Center
    Yale University School of Medicine
    Inventor, the Overall Nutritional Quality Index

  • Pingback: What to Eat » More on nutrition scoring: David Katz responds()

  • Pingback: Bookmarks Tagged Irritable()

  • Mark

    There’s a vast literature and prior art on nutrient density (and its inverse, caloric density) scoring, and at this point it’s mostly reinventing the wheel. There was a good roundup on a dozen or so of the more interesting systems in one of the nutrition journals a couple of year’s back (I’m on vacation so I don’t have the reference at hand).

    We use the RRR or 3R system at CalorieLab, because it only requires the FDA’s Nutrition Facts data, which lets us apply it to packaged food and most restaurant foods, rather than just to foods in the USDA database.

    As we say in our FAQ, “Nutrient density ratings necessarily simplify a complex subject and should be viewed in that light.”

  • In response to Mark- the article you are referring to is by Dr. Adam Drewnowski, and was published in the American Journal of Clinical Nutrition. A more current analysis, entitled “Nutrient Profiling: coparision and critical anlysis of existing systems,” was published by Azais-Braesco et al. in Public Health Nutrition late last year.

    I disagree that the wheel should not be reinvented. The first wheels were likely stone or bone; then wood. Vulcanized rubber is a 20th century invention. The wheel, and the mousetrap, have in fact been reinvented, and greatly improved, many times over.

    The ONQI does the same for nutrient scoring. It uses the literature to which you refer as the starting point, and builds from there- incorporating a great deal of epidemiology along with nutrition science. And it includes many more nutrients than the nutrition facts panel, requiring us to engage in multiple licensing arrangements to procure the data we need to ‘fuel’ the algorithm (and always meeting or exceeding the standards of the FDA and USDA – we know, because we checked). We have done so, and have the means to generate the over 30 nutrients we need to produce an ONQI score for over 120,000 foods, and that figure will go up. We will be able to generate a score for virtually any item in the food supply, and any recipe. In fact, strictly as a by-product of generating ONQI scores, we will be producing the world’s richest library of nutrient data- over 100 nutrients for each of as many as 200,000 food items.

    I am glad my car does not run on stone wheels. Just because nutrition scoring systems existed before does not mean that better ones cannot be produced. With all due respect, that kind of thinking places a low ceiling over the restive aspirations of human ingenuity. Usually by climbing onto the shoulders of those who came before, we can always reach new heights.


  • Pingback: » The first topic in 2008: food scoring systems()

  • Pingback: Pages tagged "fretful"()

  • Pingback: Peach “Lasagne” « Salivation Army()

  • I merely wished to officially say “Hi there” to everyone here. In my opinion , that this appears like an unusually appealing place to be online.

    I cannot wait to begin. Do you have any sort of guidance for someone in the beginning stages? Any certain section that you would advise more versus others to begin?

    I want to to share with you a quote that i have often found to be tremendously motivational in order to start formal introductions:

    [quote]Never believe in mirrors or newspapers. ~Tom Stoppard[/quote]