Our study quantified the relative contribution that food companies and their products make to household sodium purchases in Australia. Products from just a small number of food companies accounted for a large amount of the total sodium purchased, with the majority of sodium purchased from processed meat, bread, sauces and cheese. Most households had packaged food and beverage purchases without optimal sodium density. Lower income households in Australia were found to purchase significantly higher amounts of sodium from packaged foods and beverages compared to middle and higher income households.
A key finding from our analyses is that packaged foods and beverages produced by just 10 food companies (representing 0.7% of total number of companies) accounted for over half of all sodium purchases. This was largely attributable to the private-label products of three major supermarket retailers and was primarily driven by the large volume of these products purchased by Australian households, rather than a higher sodium content. These findings suggests that small reductions in the sodium content across key food categories and companies has considerable potential to create meaningful change to population salt intake levels, and that the Australian government should consider prioritizing engagement efforts toward key companies.
Another important finding is that at present, only six of the top 10 food categories that contribute most to household sodium purchases actually have reformulation targets defined by the Healthy Food Partnership . For some products, such as plain milk, reformulation targets would not be appropriate given these products do not contain added sodium. However, even assuming 100% adoption of the Healthy Food Partnership targets across all companies, it has been estimated this would only achieve an -9% reduction in sodium intakes (-212 mg/person/day) . Our findings reinforce the need for broadening the scope of the Healthy Food Partnership by adding further category-specific targets. In particular, our analyses suggest additional reformulation targets should be set for some processed vegetables (e.g. pickled vegetables), edible oils (e.g. butter and margarine), and spreads and dips. Previous analyses indicate that the sodium content of these types of packaged foods varies extensively across the even quite similar foods [10, 23, 24] indicating that reformulation is likely feasible from both a food technology and customer taste perspective . Furthermore, the UK government has already set sodium targets for these three categories , demonstrating the potential scope of a broader target range in Australia.
Our study also found that just a small number of food categories disproportionately accounted for total sodium purchases. This finding aligns with existing literature conducted in the US  and the UK . In the US, the top contributors to sodium purchases in 2014 were condiments/sauces/dips (%), mixed dishes, salty snacks, breads and processed meats  and in the UK, processed meat, bread, dairy products and sauces and spreads . While these findings are not directly comparable due to different classification systems used for categorizing products, they still highlight a number of key food categories globally that consistently contribute a large share to sodium purchases.
Prior research in Australia and globally has also demonstrated higher sodium intakes in individuals with a lower socio-economic status (SES) [29–32]. By using contemporary and objectively collected sodium purchase data, representative of Australian households, our findings suggest that differences in sodium consumption according to SES is likely at least partly driven by larger volumes of packaged foods and beverages purchased by lower-income households, rather than due to purchases of higher-sodium content products. Assuming purchasing patterns of packaged foods were to remain stable in the population, our findings suggest that reformulation of packaged products to a lower sodium content will likely result in greater proportional reduction in sodium intake in lower-income households, which could contribute to a reduction in cardiovascular disease-related health disparity .
A key strength of this study was the use of objective purchase data to assess household packaged food and beverage purchases in a nationwide sample of Australian households, which was comparable to the average Australian household size, household income and location. The use of continuously collected purchase information with brand- and product-specific nutrient data is an innovative approach to accurately and objectively estimate the contribution of major food companies and their products to household purchases of sodium. Furthermore, our findings at the food company and food category level provide detailed baseline data to allow us to track long-term trends in sodium purchases and monitor the impact of the Healthy Food Partnership on sodium reductions in the Australian food supply .
A limitation of the analyses is that under-reporting of purchases by the Nielsen Homescan panel is likely, with previous research suggesting under-reporting rates of 10–20% [35, 36], although we attempted to control for such underreporting by excluding households below spending threshold limits, which did not have a discernable impact on our findings. As we achieved an 89% match rate between FoodSwitch and the purchase quantity of products in the Nielsen dataset, our results further slightly underestimated true household sodium purchases. However, such under-estimates are unlikely to have affected the validity of our results as the main results of our paper involved ranking companies and food categories by their relative contribution. Given our study analyzed packaged food and beverages available in Australia and purchased by Australian households, our findings may not be generalizable to other countries. This study did not assess household food and beverage expenditure outside of the home, which is a growing portion of household food spending, and represents an important area for future research to understand population sodium exposure .