Food star ratings are actually fairly accurate, large audit concludes

Mr 2018 May29

They’ve come in for some bad press, but health star ratings are actually doing a fairly good job of defining what a nutritious packaged food is, a study reveals.

Sydney researchers have found that, in most cases, the five-star rating system aligns with the Australian Dietary Guidelines, and where they’re misleading, it’s the guidelines that have fallen down.

The health star rating scheme, introduced in 2014, has been widely criticised for infamously awarding four-plus stars to foods such as Milo, while foods such as plain Greek yoghurt only earn 1.5 stars.

However, the study of more than 47,000 products has discovered that the failures tend to lie with the guidelines’ definition of core and discretionary foods.

The researchers, from the George Institute for Global Health, defined the products as either core (23,460) or discretionary (23,656) using information from nutritional food labels to make the analysis.

They defined core foods as those in the dietary guidelines’ five food groups that form the basis of a healthy diet: vegetables and legumes, fruits, grains, dairy and alternatives, and lean meats and poultry. And discretionary foods were those not necessary for a healthy diet, including foods high in saturated fats, added sugars and/or salt or alcohol.

Overall, 87% of the reviewed products had a health star rating that aligned with their dietary classification. The remainder had star ratings that didn’t align with their core or discretionary definition.

Of these 6300 products, 83% were “ultimately determined to be Australian Dietary Guidelines failures, largely caused by challenges in defining foods as core or discretionary”, the researchers said.

“In contrast to intense media attention on occasional anomalies, this large quantitative analysis suggests that the scope of genuine misalignment between the Australian Dietary Guidelines and health star rating algorithm across the Australian food supply is very small,” they concluded.