What goes into a food health star rating – and crucially, what doesn’t

Examine, a free weekly newsletter covering science with a sceptical, evidence-based eye, is sent every Tuesday. You’re reading an excerpt – sign up to get the whole newsletter in your inbox.

Last month, Kate Aubusson and I covered a new paper arguing that multinational corporations were the world’s deadliest “vectors of disease”: corporate mosquitos spreading obesity, lung cancer and microplastics.

One line jumped out to a lot of readers: a mention that the food health star system had been co-designed by the food industry and was, to quote a professor of food politics, “deeply flawed”.

Health stars have been with us for more than 10 years now. About four in every 10 products wear them on the front of the box (they will become mandatory in the next few years).

My morning bowl of muesli gets 4.5 stars. My favourite muesli bar gets four. I get to feel smug about my good choices. A lot of readers apparently feel the same.

So … what do you mean, the star ratings are “deeply flawed”?

Stars of the show

Health star ratings have a torturous history. Originally, they were meant to be mandatory traffic lights; that later became the star-ratings system.

The food industry was initially on board, but some soon got cold feet. They and public health advocates lobbied ministers and the media, and eventually, in 2014, the government approved a non-mandatory Health Star Ratings system.

Health stars are self-determined by food manufacturers in a multi-step calculation.

Step 1: Are you selling water or fresh fruit or vegetables? No need to go further, you get an automatic five-star rating

Step 2: Work out which category your product is – food or beverage, dairy or non-dairy, fat or oil or cheese.

Step 3: Put in your “negative points” – the energy, saturated fat, sugar and sodium levels.

Step 4: Put in your “positive points” – fibre, protein and the percentage of fruits, nuts, vegetables and legumes your product contains. And voila! You get between 0.5 and five health stars.

Why does protein get extra points, especially given Australians tend to eat enough protein already, and not enough of a range of other nutrients? The theory was animal-based protein (meat) tends to be high in other key nutrients such as iron and calcium; we’ll see the problems this creates in a moment.

You can play with the calculator here, making up all sorts of Franken-foods. You’ll quickly see how increasing protein or fibre levels can ramp up your health star rating.

Do they work?

Your new Franken-food is a clue to a key criticism of health star ratings – that they fall victim to “nutritionism”, focusing more on individual nutrients within a food than the actual healthiness of that food.

“It’s not really assessing the form of the foods,” says Dr Sarah Dickie, a public health nutritionist and researcher at Monash University. “A lot of ultra-processed foods can get through this system because they can add isolated forms of fibre and protein. Protein shakes have five stars.”

“It’s this idea you can have high salt, sugar, fat, and if you add enough of something else good it masks the idea that it’s high in sugar – but no, it’s still high in sugar.”

And food companies have “gamed” the star system, public health researchers allege.

“The worst offenders … basically offset their high salt and sugar content by pumping things full of protein,” says Associate Professor Alexandra Jones, leader of the food governance program at the George Institute for Global Health, an independent medical research organisation.

“We particularly see this in protein bars and breakfast cereals, because they are the kind of food where you can do that.” When measures become targets, you can get perverse results.

Dr Duncan Craig, director of health at the Australian Food and Grocery Council, did not respond directly to the allegations the food industry was gaming the system, but did defend the stars.

“Health Star Rating is built on decades of nutrition science. A higher Health Star Rating reflects higher levels of positive nutrition – including fibre, protein, fruit, vegetables, nut and legume content and lower levels of saturated fat, sugar and sodium,” he said in an emailed statement.

Part of the point of health star ratings was to encourage the food industry to make healthier foods – is it really fair to call that process “gaming”?

Then there’s the ultra-processed problem. Almost half the food we eat is industrially formulated, using precise combinations of nutrients chemically extracted from source foods, plus chemical additives.

A very-large meta-analysis published in the BMJ in 2024 linked exposure to these foods to higher rates of diabetes, cardiovascular mortality and negative mental health outcomes.

The argument here is these processed foods are bad for us, not just because of what they contain but also because of how they are made. The star algorithm doesn’t consider that.

Associate Professor Phillip Baker, who researches ultra-processed foods at the University of Sydney, points to liquid-breakfast drinks, which often get high health stars. “Yet it’s full of all sorts of food additives,” he says. “It’s about as ultra-processed as you can get. It’s not really food.”

The Grocery Council’s Craig rejected the suggestion the stars should consider whether a food is ultra-processed.

“There is no single agreed scientific definition for this term, and the level of processing is a poor and often misleading guide for determining healthier dietary choices.”

Everyone I spoke to agrees the ratings can be improved. But are they fatally flawed?

As with most measures, the test we should apply to health stars is whether they actually tell us something meaningful about the healthiness of food.

In a 2018 paper, the George Institute’s Alexandra Jones showed that they do. Her team compared the health star ratings of 47,116 products with their “core” or “discretionary” status under the Australian Dietary Guidelines. Core foods form the basis of the guidelines’ healthy diet; discretionary foods are not necessary for health, and are high in salt, fat, sugars or alcohol.

Overall, “core” foods in Jones’ study had a median star rating of four and discretionary foods a median rating of two. The exceptions tended to be snacks with high amounts of salt and dried, fried or pickled vegetables, which tended to score highly despite being high in negative nutrients.

“In contrast to intense media attention on occasional anomalies, this large quantitative analysis suggests that the scope of genuine misalignment … is very small,” the paper concludes.

Because health stars are not mandatory, food companies – unsurprisingly – are much less likely to put low ratings on the box, evidence suggests. That feels like the real system gaming.

And it’s useful to know – because if you’re considering a box of food without a rating, there’s a good chance it’s not real healthy.

But Jones says we shouldn’t get too hung up on scouring packages for star ratings. “What we care about with diets is overall dietary patterns – so I usually tell people to remember most of the best stuff (fresh fruit and veg) doesn’t come in a pack and won’t have stars. Use stars once you’ve made your shopping list to pick a better yoghurt, dip, sauce, what not. And don’t get too hung up on the rest.”

The Examine newsletter explains and analyses science with a rigorous focus on the evidence. Sign up to get it each week.