A clustered-randomized controlled trial in which 45 food companies are the unit of randomization. The intervention companies will receive an advocacy program which comprises of commonly used advocacy actions, incorporating a theory of change model. The control companies will have no specific intervention targeted at them. The aim of this study is to quantify the effects of advocacy delivered by a local non-government organization on the salt content of food products produced or marketed by companies in Australia.

The product formulations of processed packaged foods frequently require added salt. Salt reduction in these foods is a focus of non-governmental organisations (NGOs) in an effort to reduce diet related disease. Evidence suggests that advocacy does have the potential to influence corporate behaviour but few robust data exist to describe the effects of NGO actions on food companies' salt reduction efforts.

A food composition database was used to select eligible food companies in Australia which were then classified into three strata based on company ownership, size of company and industry sector. Of the 45 food companies, 23 were randomised to the control group, and 22 to the intervention. The sample will provide 80% power to detect a difference of 50mg/100g in mean sodium levels assuming a mean of 430mg/100g, standard deviation of 300mg/100g and intracluster correlation coefficient (ICC) of 0.05 using a two-sided T-test with a significance level of 0.05.

The control group will have no specific intervention targeted at them but specific requests of the study team will be acted upon within the resources available. The intervention group will receive an advocacy program which comprises of commonly used advocacy actions, incorporating a theory of change model.

Data for the study will derive from periodic surveys of the characteristics of included companies, annual surveys of the composition of the processed foods they provide and an advocacy log recording all elements of the intervention program.

The study is being conducted by an Australian NGO over two years between December 2013-2015. Ethics approval to collect survey questionnaire and interview data from food companies has been obtained from the Human Research Ethics Committee at the University of Sydney.

This study will provide evidence about the potential for an Australian advocacy program to influence corporate behaviour and the quality of the processed food supply in Australia. Whether the program is effective or not the results, which use a novel experimental approach, will have important implications for the future of Australian efforts to reduce the large burden of disease caused by poor diet - a positive finding will highlight the need for investment in advocacy whilst a negative result will reinforce the importance of other, policy-based initiatives for the improvement of the food supply.

Inclusion Criteria:

The unit of randomization is a food company. Food companies were included if they had

• Australian based production, distribution or marketing of processed foods.
• 20 or more processed food items recorded in a food composition database (2011).
• the types of food they manufacturer are likely to contain added salt, sugars or saturated fats.

Exclusion Criteria:

• No identifiable Australian operation.
• Less than 20 processed food items recorded in a food composition database (2011).
• A company was known to be in receivership.