Swapping the UK’s 2004/5 nutrient profiling model for the updated 2018 version would pull roughly a quarter more grocery products into advertising and promotions restrictions, according to new University of Oxford research. Soft drinks, breakfast cereals and yogurts would feel the change most, driven almost entirely by a shift in how sugar is scored.

A stricter yardstick

The UK Nutrient Profiling Model (NPM) scores products on their nutrient content to determine which foods and drinks count as “less healthy”. Under current legislation, a product falls within the advertising and promotions restrictions if it sits in one of the specified food or drink categories and is shown to be high in saturated fat, salt or sugar (HFSS) when scored using the NPM 2004/5 and the 2011 Technical Guidance.

The model was first developed in 2004/5. Following the Scientific Advisory Committee on Nutrition’s 2015 update to dietary advice, particularly on free sugars, the Government accepted the recommendations and commissioned a review. The resulting update, NPM 2018, made minor adjustments to the fibre and salt thresholds, switched from total sugars to free sugars, and changed the scoring thresholds. The new NPM was published in January 2026, and in March the Government launched a consultation on applying it to the restrictions.

To test what that would mean in practice, researchers at the Nuffield Department of Primary Care Health Sciences, University of Oxford, led by Dr Barthélemy Sarda, Prof Peter Scarborough and Dr Lauren Bandy, applied both models to a large dataset of supermarket products. The project was funded by the NIHR Policy Research Programme. The aims and research questions were co-developed with the Office for Health Improvement and Disparities, while the analyses were conducted independently.

The headline numbers

The dataset covered products sold by Tesco, Sainsbury’s, Asda, Ocado and M&S. In total, 15,514 unique products fell into one of the thirteen in-scope legislative categories.

Under NPM 2004/5, 58% of those products failed and would be subject to the restrictions. Under NPM 2018, that rises to 72%. Even so, the two models agreed on the classification of the vast majority of products (84%). Of the 2,498 products that were classified differently, 93% moved from passing under the old model to failing under the new one.

Sugar does the heavy lifting

The researchers attribute the gap primarily to the revised sugar component. Replacing total sugars with free sugars, combined with stricter scoring thresholds, meant many products picked up more points for sugar under the new model.

The effect is starkest in three categories:

• Soft drinks: pass rate falls from 66% under NPM 2004/5 to 6% under NPM 2018
• Breakfast cereals: from 79% to 39%
• Yogurts: from 78% to 36%

Fibre offers a partial offset

The picture is not one-directional. The maximum points available for fibre doubled under NPM 2018, from 5 to 10, and the study found that this can offset some of the extra sugar points. Breakfast cereals and savoury snacks in particular scored more points for fibre under the new model.

Life on the borderline

The share of failing products sitting close to the pass/fail line, defined as scores of 1–3 for drinks and 4–8 for foods, remained relatively stable, edging up from 10% overall under NPM 2004/5 to 13% under NPM 2018. Some categories moved more sharply, however: breakfast cereals, savoury snacks and ice cream all saw a larger proportion of their failing products close to the border.

How it compares internationally

The researchers also compared NPM 2018 against other classification systems, including UK multiple traffic light (MTL) labels, Nutri-Score and the Chilean Warning Label. NPM 2018 proved broadly comparable across all of them. The highest agreement was with Nutri-Score, where a pass was defined as a rating of A or B.

Why free sugar estimates matter

Free sugars and fruit, nut and vegetable content are not reported directly on labels, so the team estimated them from ingredients lists using an existing algorithm that calculates the percentage composition of each ingredient. Sensitivity testing suggests the approach is robust at an overall level:

• Varying the estimated free sugar score by plus or minus 2 points changed the overall proportion of failing products by only plus or minus 2.7%, although some categories were more sensitive than others.
• Where the degree of processing of fruit and vegetable ingredients was unspecified, the main analysis assumed ingredients were whole or intact and so did not contribute free sugars. Treating them all as fully processed, and therefore as contributing free sugars, reduced the proportion of products passing NPM 2018 by 3.2% for yoghurts, 2.6% for pizzas and 1.6% for ready meals.

A further test showed how much the free sugar question matters. If total sugars were assumed to equal free sugars under NPM 2018, classifications became stricter across every category, with much bigger differences for yoghurts, breakfast cereals, pizzas and ready meals.

The authors say this underlines the importance of estimating free sugar content correctly, particularly in those categories. They also caution that products whose estimated free sugar content lies close to the pass/fail border could be affected by incorrect estimation.

Limits of the dataset

The authors set out several limitations. The analysis does not cover discount retailers or the out-of-home sector, including restaurants, takeaways and canteens. Products with incomplete nutrition information or poor-quality ingredients data were excluded. Fibre, which is not subject to mandatory labelling, was imputed using a category median where missing. And the data were not weighted by sales or market share, so the results describe the range of products on shelf rather than what shoppers actually buy.

Where the pressure lands

For the study’s authors, the conclusion is clear: replacing the current model with NPM 2018 would increase the number of grocery products subject to the advertising and promotions restrictions by around a quarter, with the greatest changes in soft drinks, breakfast cereals and yogurts. They present the findings as an evidence base for understanding how updating the NPM might affect the scope of existing legislation.

For manufacturers and retailers, the research points to where scrutiny is likely to concentrate. Sugar is the lever that moves classifications, so categories dominated by sweetened products are the most exposed. The doubled fibre allowance, meanwhile, shows that the new model does not score sugar in isolation, a point of particular relevance to cereals and savoury snacks. With a growing share of failing products sitting near the threshold in categories such as cereals, savoury snacks and ice cream, small compositional differences could matter. And because free sugars have to be estimated from ingredients rather than read off a label, the accuracy of that estimate could determine on which side of the line some products fall.