Is Tufts University’s Food Compass Nutrient Profiling System ‘Broken’? 

In late 2021, Tufts University’s Friedman School of Nutrition released their Food Compass nutrient profiling system in the prestigious journal Nature Food

Food Compass uses a quantitative ranking algorithm across nine domains to rank foods from 1 to 100. 

As the abstract notes, Tufts Food Compass ‘aims to discriminate healthfulness of foods for front of package labelling, warning labels, taxation, company ratings, and more.’

Food Compass is not just another theoretical research paper.  As you will see, the Food Compass will likely form the foundation of future food public policy in the US and the rest of the world over the next fifty years.  So, it’s pretty important! 

However, since its release, many commentators have noted that the Tufts University Food Compass’s nutrient profiling system provides surprising and non-sensical results, with many ultra-processed foods ranking above minimally processed, nutrient-dense options. 

Sadly, the Food Compass may have been distorted by the magnetic pull of financial conflicts of interest and belief-based nutrition. 

In this article, we review Food Compass’s nine ranking factors against 125,761 days of food data from 34,519 to understand if the Food Compass aligns with the parameters that help us increase satiety and avoid obesity.

What Is a Food Compass Score?

The Tufts University Food Compass ranking system evaluates foods across nine domains to assign a Food Compass Score between 1 and 100. 

The chart below shows the weightings given to each of the nine domains used in the Food Compass:

  1. Nutrient ratios
  2. Vitamins
  3. Minerals
  4. Food-based ingredients
  5. Additives
  6. Processing
  7. Specific lipids
  8. Fibre and protein
  9. Phytochemicals

As represented in the diagrams below, a food scores and ranks more highly if it scores well across all nine domains, and vice versa. 

  • The chart on the left scores well in vitamins and minerals but poorly in the other domains, which gives it a low score of 23. 
  • Meanwhile, the food shown on the right scores well in food-based ingredients and several other domains, so its compass score of 83 is much higher. 

Selecting these parameters and their weightings is critically important for ensuring they provide sensible results that align with greater satiety and improved metabolic health. 

We’ve been on a five-year mission to design and build a data-driven system using large datasets to quantify the foods and meals that cause us to overeat.  From there, we’ve developed a ranking system that would allow us to identify ideal foods and meals that align with specific goals. 

To learn more about our nutrient-focused food ranking system, see:

Who Created the Food Compass? 

The Food Compass nutrient profiling system was developed by the Friedman School of Nutrition Science and Policy at Tufts UniversityProfessor Dariush Mozaffarian led the project, ranking 8,032 foods that took three years to complete. 

Tufts University touts the Food Compass as the ‘most comprehensive and science-based nutrient profiling system to date that clears up confusion to benefit consumers and policymakers.’ 

In May 2022, it was announced that Mozaffarian was named co-chair of the White House Conference on Hunger, Nutrition, and Health planned to be held in September 2022. 

The conference is billed as the most significant food policy event in over fifty years.  The previous occasion was in 1969, under the leadership of President Nixon and senators George McGovern and Bob Dole. 

The chart below shows what has happened to our food system since the historic McGovern Report, which suggested Americans eat less fat and cholesterol and more complex carbs and fibre.  

Over the past 50 years, the diet of the US and the rest of the world has become a more refined conglomerate of processed carbohydrates and ‘heart healthy’ industrial seed oils that have fuelled our obesity epidemic. 

With Mozafarrian at the helm of the upcoming congress, the Food Compass will likely be enshrined into public policy later this year.  Hence, Food Compass is not just another research paper; it’s an important document worth reviewing in detail. 

The McGovern Commission and the Dietary Goals for Americans that followed in 1977 were based on research and beliefs about nutrition during that time.  The guidelines were not tested together as interventions to see if they improved satiety or health. 

Unfortunately, as you will see, many of those beliefs have not changed in the last half-century and are set to be locked in for the next fifty years once Food Compass becomes public policy.   

For a data-driven review of the current dietary guidelines, see: 

Food Compass Food List vs Satiety Index Score

If you’re simply reading this to understand what you should eat, the chart below shows how the Food Compass ‘healthfullness’ rating compares to our Satiety Index Score. 

Food Compass’s goal is to define a food’s ‘healthfulness’ based on the parameters we’ll review below.  Meanwhile, the Satiety Index highlights foods that will satisfy your cravings with fewer calories by providing the nutrients you require. 

  • In the bottom left, we see both systems have a low rank for butter, coconut oil and lard. 
  • Both systems rank non-starchy green vegetables like spinach, rocket and broccoli in the top right.  Seafood also scores well in both systems.
  • Along the bottom, we can see that Food Compass gives a significant bump to refined fats like olive oil and mayonnaise get a significant bump because they contain unsaturated fats, even though they provide little nutritional value. 
  • Along the right-hand side, many fruits score 100% even though they provide minimal nutrition other than a few vitamins
  • Meanwhile, towards the top left of the chart, we see that Food Compass provides a lower weighting for animal-based foods due to their cholesterol content. 

The Benefits of a Nutrient Profiling System

Confusion and conflicts of interest abound in the field of nutrition.  Hence, creating an unbiased quantitative ranking system that food manufacturers can’t manipulate is a  worthy quest to combat the exploding diabesity epidemic.  

Considering many modern interfaces like Google, Instagram, Facebook, and TikTok are optimised with quantitative and complex algorithms, it makes sense that we should also be able to systematise the parameters of the food we eat based on what our bodies require to precisely satisfy our cravings.  Doesn’t it?

With the explosion of voice search, it won’t be long before people come home and say, ‘Google/Alexa, what should I eat tonight or this week?’ and Amazon or Uber Eats will arrive on their doorstep with dinner or the groceries for the next few days. 

Unfortunately, most ranking systems are built to advertise to YOU more precisely and end with YOU as the product.  Because they do this so effectively, Facebook, Instagram, and Google have quickly become some of the wealthiest companies. 

Similarly, if the food industry finances a food ranking system, it will likely guide you to buy more of the products it wants to sell you. 

In the following sections, we’ll peek under the hood of Food Compass to see how each ranking factor stacks up against our satiety analysis. 

Ready for a data-driven nutrition adventure? 

Let’s go!

The Ranking Factors

The table below shows the nine domains used by the Food Compass System nutrient profiling system, which includes 35 different ranking criteria.  Note that the three domains towards the right (specific lipids, protein and fibre, and phytochemicals) have been given half the weighting of the other five domains. 

On the surface, these all sound logical.  But will this complex system help you identify the foods containing the essential nutrients required to let you thrive and feel satisfied?

In a recent review of the Food Compass nutrient profiling system, Ty Beal and colleagues noted that ‘a primary limitation of Food Compass is that the weighting of attributes and domains does not represent a balanced reflection of available evidence and sometimes appears arbitrary and insufficiently transparent.’ 

We heartily agree!  And by the end of this article, we think you will too.

Let’s look at the details and how they align with our data.

Fibre and Protein

Let’s start with what we believe the big guns are: protein and fibre. 

Protein (%)

Protein % refers to the per cent of total calories from protein. 

Data from our analysis of our 125,761 days of food data from 32,192 Nutrient Optimiser users shows that protein percentage dramatically impacts the amount of food you tend to eat.  This is illustrated in the protein % vs calorie intake chart below.

People consuming a diet with the lowest percentage of protein tend to eat twice as many calories as people eating a very high protein %! 

We created the nutrient-calorie satiety response curves like the one shown above for protein by dividing the data into ‘buckets’ (e.g., 10 – 15% protein) and calculating the average daily calories consumed by those people. 

With such a large dataset for both macros and micros, we’ve gained some powerful insights into how the quantifiable parameters of our food align with satiety and how much we eat!

Our appetite balances our requirements for nutrients like protein and energy.  Hence, it’s critical to get adequate protein and nutrients without overconsuming energy

The frequency distribution chart below illustrates how we tend to gravitate towards an intake in the middle of high and low protein extremes.  The protein % of our Optimiers tends to be around 30% which is much higher than the US population average of approximately 13% protein.

While increasing protein % by dialling back energy from fat and carbohydrates sounds simple, it’s not easy to do in practice.  We are creatures of instinct and habit, and old habits die hard.  In our Macros Masterclass, we guide people to ensure they get adequate protein while progressively dialling back the energy from fat and carbohydrates in their diet. 

ALL organisms eat until they get enough protein.  So, if we only have access to low-protein foods, we eat more until we satisfy our bodies’ requirements.  Conversely, if we only have access to higher protein % foods, we innately consume fewer calories

Our multiple analysis of several datasets aligns with Professors Raubenheimer and Simpson’s Protein Leverage Hypothesis.  From slime to insects, animals in the wild, and humans, protein leverage has been observed consistently in multiple organisms.  The chart below shows the summary of protein % vs total energy intake in humans. 

Findings from other human studies align with observations from our data analysis.  Increasing satiety and a lower overall calorie intake align with a modest increase in absolute protein intake but a significant reduction in energy from fat and carbs.  Hence, this equates to a higher protein %. 

Total Protein vs Protein %

Before we go on, I want to address one common point of confusion.

Upping our protein to coincide with protein leverage theory and a higher protein % does not simply mean we need to eat MORE protein (i.e., in grams).  Simply eating more protein would just increase our calorie intake.  For example, you can’t eat more low-protein % foods like doughnuts, bacon, or butter to get your protein. 

The chart below shows that a higher absolute protein intake (i.e., in grams) aligns with greater energy consumption.  The average intake of our Optimisers is a moderate 110 g per day. 

Rather than eating more of the same foods to get your protein, you have to change what you eat.  In practice, this requires us to reduce our consumption of easily-accessible energy from fat and carbs and only modestly increase our absolute protein intake. 

Fibre

Fibre also has a positive impact on satiety.  To the left of the chart, we can see that a little bit of fibre is better than none.  From there, satiety increases as we get more fibre per calorie.  Fibre—or the lack of—works as an excellent indicator of whether a food is processed or refined; the less fibre a carb-rich food has, the more likely it is to be processed (i.e., non-starchy vegetables vs refined sugar and flour).  

While fibre is substantial for satiety, we see that protein has three times the satiety impact that fibre.  When we run the data through a multivariate regression analysis, we found that increasing protein from 18% to 44% corresponds with a 34% reduction in calories, whereas moving from 2% to 9% fibre aligns with a more minimal 10% reduction in calories. 

Food Compass Weighting for Fibre and Protein

Good news: the Food Compass nutrient profiling system includes fibre and protein in its ranking factors. 

Bad news: the weighting for these critical parameters is very low.

Protein and fibre—the most critical factors influencing satiety—have not only been grouped together in one domain, but they have been given half the weighting of the other five. 

The chart below shows the relative weighting given to each ranking factor (domain) used by Food Compass.  The little brown slice in the top left represents protein and fibre combined.  Their system only weighted 6.7% to protein and fibre!

If the goal behind the Food Compass was to actually empower people to eat less and reverse the diabesity epidemic, the fact that some of the highest-ranking foods (e.g., celery juice, Cheerios, grapefruit, avocado, and chocolate almond milk) contain negligible protein suggests that they might have dropped the ball in their weightings.

Fibre and protein certainly deserve to be included!  However, Food Compass appears to have massively underweighted the most critical factors.  As you will see later in this article, adding all the other parameters makes the most important factors largely irrelevant!

Why Protein is Often Neglected in Our Modern Food System

The word protein is derived from the Greek word Proteus, meaning ‘of first importance’.  Proteus was also the firstborn son of Poseidon. 

Although the importance of protein is well-known, high-protein foods tend to be more expensive and take longer to grow.  As a result, profit margins are lower.  Hence, a system that guides consumers away from protein is good news for food manufacturers. 

In modelling or forecasting—my primary domain as an engineer—complex black box systems often leads to garbage outputs, especially if the outcomes aren’t compatible with real-world data.  In other words, we get less accurate data if we have too many different scenarios of people eating in the real world.

A complex system with many moving parts is not necessarily better than a more straightforward system that works.  You can’t just throw a laundry list of your favourite indicators in a box, shake it up, and hope for the best! 

When developing any model, you must identify the most critical factors and ensure that the weightings align with real-world data.  Otherwise, your model will make you more confidently wrong! 

The Punchline

Before we go any further, I want to give you the punchline of our satiety analysis and this whole article:

Nutrient density is the icing on the cake of satiety and Nutritional Optimisation.  However, getting adequate protein without excess energy from fat and carbs is the cake itself!

Consuming adequate protein without excess energy should be your first fundamental step in optimising your nutrition and increasing your satiety. 

If you read no further, this is the simple take-home message:  A higher protein % is the most potent food parameter that aligns with increased satiety and a lower calorie intake.  

If you want to learn more about the icing on the cake and leveraging nutrient density in your favour, read on.

Minerals

It’s encouraging to see that some critical essential minerals were also included in the Food Compass.  That is,

Our analysis also shows that foods and meals that contain more minerals per calorie satisfy our cravings to varying degrees and improve satiety.

The charts below show the satiety response curves for the minerals considered by Food Compass.  In each graph, you will notice that getting more of each mineral per calorie tends to align with a lower energy intake. 

However, you should also note that these relationships are not linear.  Once we get enough of a nutrient, our cravings tend to decrease for it and send us in search of foods that contain other nutrients we need to thrive, like a self-guided nutrient-seeking missile. 

Note:  While considered a ranking factor by Food Compass, available food databases do not commonly measure iodine because it is highly variable depending on the region the food is grown.  Hence, we don’t believe we should include iodine as a ranking factor because of the lack of reliable data.

Which Minerals Will Satisfy Your Appetite the Most? 

The table below shows protein, fibre, and minerals ranked in terms of satiety response.  While protein still dominates, foods containing more minerals like potassium and calcium align with a significant increase in satiety.    

parameter satiety response
protein %-52.0%
potassium-47.0%
fibre-43.3%
calcium-37.0%
phosphorus-35.0%
selenium-33.0%
folate-27.0%
sodium-26.0%
magnesium-25.0%
iron-23.0%
zinc-21.0%
riboflavin-21.0%
manganese-20.0%
copper-13.0%

It’s also worth noting that minerals like potassium and calcium have declined significantly in our food system over the past century.  We now have to consume a lot more food than we would have in the 1940s to get enough of these minerals.  So it makes sense that we would have the strongest cravings for them. 

Multivariate Analysis

To understand the degree to which nutrients impact our cravings when considered together, I have run a multivariate analysis of the essential minerals.  I also included protein and fibre, which we already established, as having a solid relationship with satiety. 

As the table below shows, potassium, calcium, and sodium have a statistically significant relationship with overall calorie intake.  Using the parameters of this analysis, increasing your diet’s mineral content in addition to protein and fibre aligns with a 14% additional reduction in calorie intake due to our cravings for potassium, calcium and sodium. 

 P-value15th85thCalories%
protein %019%43%-449-28.4%
potassium1.19 E-701,8395,880-103-6.5%
calcium2.84 E-594581,832-83-5.2%
fibre (%)4.03 E-282.1%8.6%-58-3.7%
sodium4.69 E-151,4414,967-37-2.4%
total-730-46.2%

To summarise—for now, anyway—in addition to protein and fibre, the core electrolytes are critical for people looking to satiate their appetite, eat less, and reverse conditions related to the diabesity epidemic

Vitamins

Next, let’s look at the satiety impact of vitamins

Below, we’ve included the nutrient-satiety response curves for the vitamins considered by Food Compass.  Like protein and minerals, vitamins affect satiety positively. 

However, their positive return only lasts to a point.  The satiety benefit dissipates when we exceed the amounts we can obtain from whole food.  In fact, it often rebounds! 

Just because your Cheerios are fortified with a smattering of synthetic vitamins to replace the nutrients lost in processing doesn’t mean they’re more satisfying or healthier for you overall.  Usually, it only guarantees your pee will be a bit more yellow.

Multivariate Analysis

When we include vitamins in the multivariate analysis, we sometimes find that riboflavin (B2) is statistically significant.  In addition, depending on the other nutrients we included, folate (B9) also appears significant from time to time. 

What About Supplements and Fortification?  

Unfortunately, our satiety response to vitamins is likely confounded by the isolated synthetic vitamins used to fortify otherwise nutrient-poor processed foods.  Furthermore, the multivariate analysis assumes a linear relationship. 

So, the analysis will be skewed by supplementation and fortification.   If we could get a dataset without fortification, we would likely see a more significant satiety response to vitamins. 

While fortification of the food system may benefit people unable to get adequate nutrients from whole foods, supplements and fortification seem to provide a disbenefit when consumed at higher levels. 

After deficiencies like pellagra and beriberi became more prevalent during the depression due to food scarcity and a dependency on nutrient-poor foods, the United States started fortifying foods like rice, flour, and cereals in the 1930s.

Interestingly, experiments were undertaken on livestock in the 1960s to examine the impact of nutrients on health and livelihood.  These studies found cattle failed to thrive, got sick, lost their hair, and stopped growing when fed processed grain feed alone and not given time on pasture to get their full spectrum of nutrients.  In other words, the after-effects were not good for business! 

However, animals no longer needed pasture or supplemental nourishment when vitamins were added to their feed.  They actually grew fatter with less feed in less time and were no longer interested in the supplemental feed!

Coincidently—or not! —soon after the fortification of animal feed became widespread to increase profits, the fortification of human food also increased throughout the 1970s.  If we plot the trendlines for food fortification alongside the obesity epidemic, we see that they seem to coincide. 

This is illustrated in the charts below that show the dramatic increase in fortified nutrients like niacin, iron, thiamine, and folate in our food system that mainly took place during the 1940s and 1970s. 

One critical role of the B vitamin family is to help us utilise carbs and fat for energy.  Thus, fortifying foods with B vitamins enables us to use energy in otherwise nutrient-poor foods and store them as fat.  But unfortunately, it seems one caveat to consuming these fortified foods is that we lose our cravings for whole foods that typically provide B vitamins, like meat, eggs, seafood, fruits, and vegetables. 

Our satiety analysis also suggests that supraphysiological quantities of vitamins only achievable with fortified foods do not improve satiety.  Instead, we actually eat more of them!  Instead of seeking out the foods that naturally contain vitamins, we are more content to continue chowing down on fortified processed foods (like Cheerios and Mini Wheats which rank well in the Food Compass system, largely due to fortification) rather than seeking our vegetables, seafood and meat that naturally contain those vitamins.   

For more details about the potential downsides of fortification, check out Mark Schatzker’s The End of Craving or this podcast with him.

Nutrient Ratios

Next, we’ll look at the nutrient ratios used by Food Compass, like:

  • the potassium:sodium ratio,
  • the fibre:carbohydrate ratio, and
  • the unsaturated:saturated fat ratio. 

Nutrient ratios are an interesting topic that is often overlooked.  Many dietary nutrients compete for absorption, so having too much of one nutrient from supplements or fortification can inhibit the absorption or utilisation of another. 

Considering the micronutrient ratios becomes important if your diet relies heavily on supplements and fortified processed foods.  However, if you get most of your nutrients from a variety of whole foods containing the full spectrum of nutrients in the forms and ratios your body understands, you needn’t worry as much about nutrient ratios; they tend to look after themselves.  

For more details on micronutrient ratios, see Micronutrient Balance Ratios: Do They Matter and How Can I Manage Them?

Potassium:Sodium Ratio

As mentioned above, our nutrient-satiety analysis indicated that sodium AND potassium positively influence satiety.  This is because we eat less when we consume more foods that naturally contain more of these minerals per calorie. 

As the charts below show, the availability of both of these critical minerals has declined significantly in recent years as our soils have become increasingly depleted.  Moreover, our previous analysis indicates that climbing obesity rates have correlated with declining mineral levels.  For more details, check out The Biggest Trends in Nutrition.

However, there is an upper limit to the benefits of these nutrients.  In other words, you won’t crave more of a nutrient you’re already getting plenty of.  Based on the satiety analysis, we have established an Optimal Nutrient Intake (ONI) of:

  • 4 grams of sodium per 2000 calories, and
  • 6 grams of potassium per 2000 calories. 

This limit has been set based:

  • The 85th percentile intake from our Optimiser data, and
  • The intake at which the effects of our nutrient-satiety charts show they start dissipating. 

We wanted to ensure that the ONIs were achievable with whole food.  While it’s a stretch target, it is possible!  However, exceeding the 85th percentile intake does not yield any additional benefit.  Hence, we’ve set the ONI based on the point at which there is no further gain. 

The Optimal Nutrient Intake targets serve as both stretch targets and limits.  Even though processed food manufacturers hijack our cravings for salt, you don’t score higher if you consume more than the Optimal Nutrient Intake for sodium (we’ll discuss this later in more detail).

There is plenty of controversy around sodium, and most recommend trying to minimise it.  However, sodium is not necessarily a nutrient we need to minimise or avoid at all costs, but we still need to keep it balanced with potassium

For more details on how much salt we really need, check out How Many Grams of Sodium Do You Need Per Day?

Hence, considering the potassium:sodium ratio is reasonable, mainly because most people don’t get enough potassium.

The chart below shows the distribution of potassium:sodium ratio in our 125,761 days of data.  The median potassium:sodium ratio of our Optimisers using Nutrient Optimiser is 1.6 (i.e., on average, they consume slightly more potassium than sodium.  

The nutrient-satiety chart below shows that people consuming more potassium than sodium tend to consume fewer calories.   But in the big scheme of things, the calorie reduction is relatively small compared to the decline we see from other nutrients like potassium and sodium themselves. 

Potassium is crucial, and you don’t need to go out of your way to get more than four grams of sodium per day.  However, there doesn’t appear to be any disbenefit in consuming more sodium.   

Fibre:Carb Ratio

Next, we come to the fibre:carb ratio, which the Food Compass system uses to measure refining and processing.  The chart below shows the satiety response to carbohydrates and non-fibre carbohydrates alongside one another.  As noted earlier, fibre tends to have a positive satiety response, and we eat less when our carbohydrates are more fibrous. 

Similar to fat, carbohydrates are just another energy source.  Hence, it makes sense to eat more when our food contains more non-fibre carbohydrates.  We can see that we consume the most energy aligns when about half of our energy comes from non-fibre carbohydrates, and the remainder comes from fat.  However, it’s not a simple linear relationship. 

  • Zero-carb is not better than lower-carb. 
  • Lower-carb (i.e., 10 – 20% total carbs or 20 – 30% net carbs) tends to align with the lowest overall energy intake.
  • On the right-hand side of the chart, we see that a very high-carb, low-fat diet corresponds to a slightly lower calorie intake than when we consume fat and carbs together.

Looking at the satiety response curve for fibre:carb ratio, we see that we benefit from consuming more fibrous carbohydrates up to around 0.3.  While this is a valuable determinant for measuring how processed a food is, you’re unlikely to get the same benefit if you just add a fibre supplement to an ultra-processed diet.

Saturated:Unsaturated Fat

Things start to get interesting when we look closely at different types of fat. 

Fat is an easily-accessible energy source and is the most calorically-dense macronutrient.  While it shouldn’t be too surprising to see that more energy from fat corresponds with a higher energy intake, it’s important to note that this is not a linear relationship. 

Similar to carbohydrates, less fat is not necessarily better.   Because protein often comes packaged with fat, eating a very low-fat diet might force us to sacrifice some of our protein intake, which would negatively impact satiety.

The satiety response to saturated fat is even more interesting.  Up until around 30% of our calories come from saturated fats, more saturated fat aligns with a higher calorie intake.  However, our energy consumption tends to decrease beyond this point.  This is likely because foods like hamburgers and chips fried in lard containing saturated fat as ingredients provide more highly-palatable energy.  However, saturated fat from whole foods like steak and eggs that contain more protein is not as much of a concern. 

Interestingly, the US intake of saturated fat was approximately 42% of total calories in the 1940s.  This is shown towards the right of the chart below.  However, with the ever-climbing increase in industrial seed oils in our food system, our average intake has drifted back to 31%.  Ironically, this percentage aligns with the maximum calorie intake. 

Meanwhile, a greater intake of monounsaturated fat aligns with a more substantial calorie intake, particularly once we exceed 15% 

For more details, see Monounsaturated Fat: Is It So ‘Good’ After All?

Processed plant oils that consist mainly of monounsaturated fat have become ubiquitous in our food system as cheap ingredients for energy-dense, nutrient-poor junk foods.  As the chart below shows, added fats and oils contributed to most of the calorie increase over the past half-century. 

Although the sources of saturated fat like butter, dairy and lard have remained relatively unchanged over the past 50 years, data from the USDA Economic Research Service shows that there has been a massive influx of ‘salad and cooking oils’

Since we worked out how to extract oil from industrial crops at scale in 1908, the fat in our food system has increased by more than six hundred calories per person per day!

The chart below from Our World in Data shows that the production of industrial seed oils like palm, soybean, sunflower, and rapeseed has exploded from 20 million tons to 160 million tons per year! 

Given that these cheap, nutrient-poor oils have ominously come to dominate our food system within the last 50 years, does it really make sense to give them bonus points in the Food Compass nutrient profiling system while penalising saturated fat which has remained relatively unchanged? 

The chart below shows no simple linear relationship between calories and saturated fat vs unsaturated fat.  Instead, the maximum calorie intake aligns with a blend of saturated and unsaturated fat, which is uncommon in whole foods. 

Summary

To summarise our findings on nutrient ratios, we see that:

  • A higher potassium:sodium ratio aligns with greater satiety.  However, it does not appear to add any additional benefit as a ranking factor if we don’t prioritise excess sodium.  
  • The fibre:carb ratio does have a beneficial positive correlation with satiety.  It is also a valuable differentiator between refined vs unprocessed carbs. 
  • The saturated:unsaturated fat ratio does not appear to be helpful.  However, it seems we should minimise processed foods that contain a blend of the two. 

Food-Based Ingredients

Food Compass has used several ranking factors based on the ingredients they contain, which I’ll offer some thoughts on individually in the following section. 

Fruits

Fruits are good energy sources that are hard to overeat by themselves, but most don’t contain much in the way of nutrients per calorie. 

For example, let’s look at the nutrient profile of watermelon, which ranks highly in the Food Compass chart system.  Watermelon provides a burst of vitamin A and C but negligible amounts of the other essential nutrients. 

If your blood sugars aren’t well-controlled, fruit is also not ideal; large blood glucose swings tend to increase appetite when your blood sugar comes crashing down. 

Fruit contains most of its energy as fructose, which can contribute to non-alcoholic fatty liver disease (NAFLD) if consumed in large quantities.   A recent study showed that consuming more than four servings of fruit daily exacerbated steatosis, dyslipidaemia, and glycemic control. 

Despite the shortcomings, quite a few fruits like apricots, blackberries, cantaloupe, cherries, grapefruit, lemon etc., score a 100 in the Food Compass ranking system.

Fruit and vegetable juices like celery and tomato are also ranked well per the Tufts Food Compass list.  However, they’re absorbed very quickly and provide little satiety.

While fruit is undoubtedly better than ultra-processed junk food, there doesn’t appear to be any uniquely redeeming features to fruit that make it THAT much more nutritious if your foods contain it! 

Non-Starchy Vegetables

Non-starchy vegetables are included as a separate ranking factor by Food Compass.  However, because they rank highly in terms of nutrients per calorie, non-starchy vegetables stand on their own merits without needing to be propped up with bonus points from an arbitrary food category.  

Beans and Legumes

As the example below shows, soybeans, beans, and legumes are excellent energy sources.  In addition, they might provide some protein if you don’t eat meat, dairy, or seafood.  However, I don’t believe they warrant bonus points to distort the food ranking system further to ensure that they rank better than seafood and animal-based protein sources, which tend to be more bioavailable and more nutritious overall.  

Whole Grains vs Refined Grains

Whole grains are certainly more nutritious than refined grains, which have been stripped of important nutrients during processing to make them more palatable.  Despite this being used as a ranking factor, even fortified cereals like Cheerios scored incredibly well?!?!  I believe this is another unwarranted category that skews the Food Compass rankings when fibre and nutrient density would better differentiate. 

Nuts and Seeds

Nuts and seeds are yet another criteria the Food Compass system uses to rank foods.  While they are a great energy source--if you need it, that is—most people find them easy to overeat.  If they are believed to be a ‘superfood’, this compounds!

With some extra weighting on the nuts category, it might seem bizarre that chocolate-covered almonds made it to the top with a score of 78!  I would love it if chocolate-coated almonds were a high-satiety superfood, but the reality is I can eat this carb-and-fat combo all day and never feel full.

Seafood

In terms of nutrient density, seafood tends to rank highly.  It is an excellent source of protein, omega-3 fatty acids, and a range of vitamins and minerals.  However, it doesn’t need any additional help from any ranking factor!  If we just let its nutrients speak for themselves, seafood does super well with respect to protein, vitamins, minerals, and essential fatty acids.  

Yogurt

Yogurt can be a great source of calcium and bioavailable protein.  However, it’s somewhat ironic that it’s been included as a specific ranking factor considering Danone funded Food Compass. 

Perhaps they needed to add a new factor to reverse the unnecessary anti-saturated fat and cholesterol ranking factors that would otherwise penalise dairy.   

Plant Oils

As we discussed above, counting monounsaturated ‘plant oils’ extracted from industrial crops via hydrogenation as positive is unlikely to add significant value to a food ranking system.  In fact, I think it is ludicrous given that modern ultra-processed food is packed with empty calories from ‘heart healthy vegetable oils’ like palm oil, canola, soybean and rapeed oil! 

If you wanted to name a single smoking gun of the diabesity epidemic, you could reasonably point the finger at the 600 or so extra calories that have been added to our food system from plant oils alone. 

Red or Processed Meat

Red meat is an emotive subject tied to sustainability and ethical beliefs.  However, others on the opposite side of the belief system (i.e., carnivores) are eating nothing but steak and thriving. 

Red meat can be a cost-effective source of protein for many people who are not getting enough.  This is especially true in developing countries, where people can’t afford to eat a vegan diet AND buy supplements to avoid deficiencies. 

Processed meat is certainly not ideal.  However, it tends to rank lower than whole foods when viewed through the lens of nutrient density anyway.   The image below from the CSIRO shows where Australians get their protein from.  If you applied a nutrient-density filter to these protein sources, the heavily processed protein sources would fall to the bottom of the pile because they are diluted with refined carbs and industrial seed oils.  There is no need for an additional weighting factor to demote red meat. 

Summary

Overall, these qualitative value judgements are unlikely to add additional value beyond focusing on nutritional quality.  Instead, they simply reinforce the ‘conventional wisdom’ our food guidelines from fifty years ago directed us to eat. 

Weighting these belief-based factors above parameters that we have statistically significant data for (e.g., protein %, fibre, minerals and vitamins) dilutes the usefulness Food Compass nutritional profiling system, which is why so many highly ranked foods appear non-sensical.

Specific Lipids

The specific lipids considered by Food Compass include:

  • Dietary cholesterol,
  • Alpha-linolenic acid (ALA),
  • Omega-3 fatty acids,
  • Medium-chain fatty acids (MCT), and
  • Trans fats. 

Omega-3 Fatty Acids

Our analysis shows that foods containing more omega-3 fatty acids positively impact satiety.  Seafood is high in omega-3s, protein, and a range of other beneficial vitamins and minerals.  Because omega-3 is an essential nutrient that we appear to crave more of, it should be incorporated.  

Trans Fats

The satiety response to trans fats is interesting.  None seems to be better than some, but eating more of them does not necessarily mean you will consume more calories. 

Focusing on nutrient-dense, minimally processed foods won’t consume substantial amounts of dangerous trans fats.  However, trans fats from whole foods sources aren’t necessarily a concern.  

ALA and MCTs

While they may be beneficial, data on alpha-linolenic acid and medium-chain fatty acids are generally hard to come by in food databases.  Hence, if it is not regularly measured in food, it’s probably not good to use as a ranking factor for widespread use.

Cholesterol

You may be surprised to see that a greater dietary cholesterol intake positively impacts satiety.  Yep, you heard that right!

While not officially deemed an essential nutrient, some dietary cholesterol has beneficial effects on the body, like hormone synthesis.  As the chart below shows, our cholesterol intake has declined since the release of the 1977 Dietary Goals for Americans that recommended we consume less fat and cholesterol and the explosion of widespread industrial farming.  Sadly, obesity has exploded since then. 

Despite the 2015 US Dietary Guidelines removing cholesterol as a nutrient of concern, Food Compass still used it as a negative ranking factor nearly six years later.

While there’s no need to go out of your way to consume more cholesterol, you also don’t have to fear nutrient-dense whole foods like liver, eggs, and beef that contain it.

Multivariate Analysis

Interestingly, cholesterol and folate remain statistically significant nutrients that we appear to crave more of when we include all the essential nutrients in a multivariate analysis.   As the table below shows, moving from low to high cholesterol foods aligns with a massive eight per cent reduction in calories! 

 P-value15th85thcalories%
protein %1.4818E-25419%44%-390-25.0%
cholesterol1.70602E-382291091-124-7.9%
sodium1.00873E-2614045015-101-6.4%
folate1.59019E-19162993-93-6.0%
fibre (g/2000 cal)2.20695E-181043-92-5.9%
potassium1.34862E-1018066051-77-4.9%
calcium1.69546E-134421883-76-4.8%
total-952-61%

While increasing protein % from low to high aligns with a 25% reduction in calories, dialling all the other statistically significant nutrients collectively aligns with a further 36% reduction in calories!

Excitingly, this deeper understanding of how various nutrients impact our craving and satiety provides an exciting alternative to the Food Compass, which has used a range of belief-based ranking factors that have not been tested as a whole.  Later, we’ll look at how the Food Compass scores align with these data-driven nutrient-based satiety index rankings. 

For more details on nutrient-dense, high satiety food and meals that will satisfy your cravings with fewer calories, see: 

Summary

  • It makes sense to include omega-3 as a beneficial ranking factor. 
  • Given the lack of data on trans fats, ALAs, and MCTs, adding them as individual ranking factors seems unnecessary.
  • Since cholesterol is no longer considered a nutrient of concern, it should not be penalised as a negative ranking factor.  In fact, our satiety analysis indicates that we crave foods containing cholesterol and will eat more calories if our food contains less of it.  

Additives

Food Compass has also considered the following additives as individual ranking factors:

  • Added sugar,
  • Nitrites,
  • Artificial sweeteners, flavours, or colours,
  • Partially hydrogenated oils,
  • Interesterified hydrogenated oils, and
  • High-fructose corn syrup. 

Additives are typically not ideal and seldomly found in nutrient-dense whole foods.  While we can fortify foods with a smattering of vitamins and minerals to add to the front of the box, foods that are nutritious across the spectrum are low in these additives. 

These subjective parameters are unlikely to provide significant additional value in a ranking system.

If I  had my way, I would also include fortification as a negative ranking factor.  Any food that requires fortification is likely a low protein, low fibre conglomerate of refined flour, sugar and industrial seed oils.   Instead, it appears that many fortified foods like breakfast cereals rank very well in the Food Compass system because they are fortified?!?!

Processing

Food Compass has also considered processing as an additional ranking factor. 

NOVA Classification

Quantifying foods based on the level of processing may be helpful.  However, it is unlikely to add additional value once we first quantify foods that naturally contain a broad spectrum of nutrients, especially protein and fibre. 

Ultra-processed foods are typically produced using large-scale industrial, agricultural practices fueled by fossil fuel-based fertilisers.  These practices grow crops quickly, but they accumulate fewer nutrients from the depleted soils they’re grown in year after year.  Nutrient density alone tends to eliminate these quickly.

Fermentation

Fermentation is generally a good thing.  For example, in our Micros Masterclass, many people use sauerkraut to optimise their vitamin C or kombucha to boost their calcium intake.  However, it is not necessarily useful as an additional ranking factor. 

Frying

Again, frying is just a cooking method that becomes irrelevant once we consider nutrient density.  Frying simply results in the accumulation of lots of fat, which dilutes nutrient density.  Fried foods automatically fall to the bottom of the pile when you consider nutrients per calorie.

Phytochemicals

Food Compass also includes phytochemicals or compounds like flavonoids and carotenoids found in foods.  While these are non-essential, they may be beneficial for health.  But unfortunately, these are not commonly measured and thus not available in most nutrition databases. 

Joel Furhman’s ANDI system relies heavily on parameters like phytochemicals.  Because they are predominantly—if not exclusively—found in plant foods, they tend to bias heavily towards a plant-based diet.

There is plenty of disagreement within the always-warring dietary camps about whether these non-essential plant compounds are beneficial or detrimental.  Unfortunately, enough quantitative data is not yet available to understand the complex effects these non-essential nutrients have on the body.

As a general rule, you will get plenty of beneficial non-essential nutrients like phytochemicals as bonuses if you consume plenty of nutrient-dense whole foods.

Should You Use Food Compass-Based Foods and Recipes?

By now, we hope you see that Food Compass uses somewhat meaningless criteria to rank foods that they do nothing to optimise nutrient density.  Additionally, some of the ranking factors they consider ‘positive’ might have negative implications on the number of nutrients you’re consuming and keep you from feeling full.  As a result, you might continue to overeat and remain one of the many trapped in the diabesity epidemic.

To help you apply nutrient density at a bio-individual level, we took the data and created the micronutrient fingerprint to gamify the process of Nutritional Optimisation.  By assisting people in getting more of the nutrients they aren’t currently getting enough of, they can create a well-balanced diet at a micronutrient level—no matter where those nutrients come from!  We firmly believe your diet doesn’t need a name—just enough nutrients!

The chart below shows an example of one of Nutrient Optimiser’s micronutrient fingerprint charts, which we use in our Micros Masterclass.  

The black line represents 100% Optimal Nutrient Intake (ONI) for each essential micronutrient.  While the Recommended Dietary Intake (RDI), Adequate Intake (AI), and Estimated Average Requirement (EAR) are intakes to evade deficiency, our ONIs are data-driven nutrient stretch targets for optimal health. 

If you can fill the area to the left of the black line without using fortified foods or supplements, you get a perfect 100% score.  To do this, you must chase the foods that contain more of the nutrients you currently lack.  Consuming more than 100% of the black line does not boost your overall score further; you need to seek other nutrients!

In our Micros Masterclass, Nutrient Optimiser gives precise food and meal recommendations to help people dial in their nutrient density AND satiety.  This is because our appetite shuts off when we start to give our bodies the nutrients they need to function optimally.

Beyond protein leverage, we see that people who pursue a higher Diet Quality Score are satisfied with fewer calories time and time again.  As an example, we included the chart below to show Diet Quality Score vs calorie intake.  As we can see, people tend to eat less as people increase the nutrient density of their diet. 

Summary

Our review of the Food Compass nutrient profiling system against our very own data-driven satiety analysis suggests that:

  • Protein and fibre are included, but they are vastly underweighted if Food Compass’s goal is to promote nutrient density and satiety so we can begin reversing the diabesity epidemic. 
  • Promoting unsaturated fats and demoting cholesterol and saturated fat only perpetuates the increased use of cheap and refined vegetable oils that have started to dominate our food system.
  • Including other factors like nutrient ratios, food-based ingredients, specific lipids, additives, and processing only dilutes the system and skews it towards more processed foods. 
  • Simply thinking in terms of essential nutrients—including fibre and amino acids—per calorie is more than adequate for ranking foods and meals that optimise satiety and support metabolic health. 

More

3 thoughts on “Is Tufts University’s Food Compass Nutrient Profiling System ‘Broken’? ”

  1. Thank you for donning your scuba gear and taking the deep dive on this for us. 🙂

    In a word, YES, the compass is broken! I read the Eureka Alerts paper on this when it first came out, looked at it, and said, “Oh no–yet ANOTHER food selection gimmick that tells us what we alreay know (or should know)”. Obviously, eating REAL foods is going to nourish us better than junk foods, and we don’t need another cutesy graphic telling us this. Apparently this is what you get with grant money! Anybody who actually cares about their health is way WAY beyond this point, and as you pointed out at length, this “compass” doesn’t go nearly into enough depth to be truly effective. Perhaps another grant or two would make that happen.

    Meanwhile, the graphics dept. has once again cranked out an image telling us to avoid junk foods in favor of whole foods. Duh!

    It’s funny–researchers STILL don’t seem to know enough about our food and how to put nutrition to its highest and best use, or are at least are unwilling to put it in print (“Oh–too complicated.”). Let’s see someone put THAT into one simple cutesy graphic! You have to use multiple graphics to do it, so what are THEIR chances?

  2. Hi,

    Thank you for this quite in depth work,

    I think the compass is limited but your tool might be too. I still firmly believe the glycemic/insulin inducing capability of the foods, they impact on the gut microbiota, their relation to inflammation and oxidative stress should be factored in any food.

    My dream would be to have a radar graph based on all the parameters with thresholds for the “limit this food” defined by age/sex/metabolic condition of each person (people with obesity might not need the same nutrients, energy or even food types as athletes).

    Might need a big data approach for this to setup z-scores for each parameter but would definitely be more helpfull.

    Now overall food must also be considered in terms of pleasure, satisfaction, user-friendliness (sweet potatoes are awesome but baking them makes them less appealing than a bunch of grapes for me).

    But again, to be used it has to be simple. So I guess just sticking to energy-controlled mostly plant-based diet with from time to time low fat animal proteins with a control on added sodium and a mass of spices and there you go. Supplementation in B12 in any from would be necessary and a focus on certain micronutrients might be necessary (zinc, selenium, iron) to avoid any troubles. Then from there go see a nutritionist to adapt on the daily scale.

    Cheers !

Comments are closed.