We identified no association between consumer knowledge, attitudes and behaviours relating to salt and actual levels of salt intake. Furthermore, while most participants identified that salt is detrimental to health, and large numbers reported actions to address their salt consumption, the great majority of this sample of the Australian population continued to consume salt at a level above national and international recommendations.
The levels of knowledge, attitudes and behaviours of the participants are similar to those reported in a prior survey of a national sample of Australian consumers done in 2007 [15, 16] and a more recent study conducted in metropolitan Melbourne . Both these studies found that the majority of participants were aware of the harmful effect of salt on health but also showed that few were able to identify the recommended upper daily intake and revealed limited knowledge of the main foods that contribute to salt in the diet. Those Australian studies are also likely to have enrolled participants with comparable average levels and patterns of salt consumption to those reported here. Similar surveys have also been done for populations in five countries in the Americas (Argentina, Canada, Chile, Costa Rica, and Ecuador)  with similar findings, suggesting that the results found in this most recent Australian study are not atypical.
Our examination of the association between knowledge, attitudes and behaviours and actual levels of salt consumption is novel and raises questions about the likely capacity of purely educational interventions to change individual or community salt consumption levels. Given the high levels of knowledge and positive attitudes to salt reduction recorded in our survey, the most likely explanation for high salt consumption levels across the community is that there are many barriers to changing behaviour. The adverse nature of the food environment which comprises heavily advertised, low-cost foods high in salt, and without adequate labelling of salt levels on packaging, is likely to be a key factor inhibiting reductions in salt consumption amongst even well-informed individuals .
In contrast, several intervention trials have demonstrated effects of nutrition education on salt consumption in selected groups of individuals [20–22]. The reason why these trials had positive effects is most likely because of the high intensity of the interventions applied, which typically included multiple one-on-one consultations with participants, small group activities or provision of reduced salt foodstuffs. While clearly effective, this type of approach is not feasible at the population level because of the resources required  and the observed benefits cannot reasonably be generalised to community settings where average exposure of individuals to educational interventions targeting salt is so much less. It is of note that in the UK, Japan and Finland, where population-wide salt reduction has been achieved, community education was underpinned by programs that changed the average salt levels in key foods, used salt warning labels to reinforce the health promotion messages and/or altered the broader food environment in some other way [24–26].
This study benefits from assessment of salt consumption based upon 24-hour urine collections and the use of standardised questions about knowledge, attitudes and behaviours related to salt. While the questionnaires have not undergone rigorous test-retest evaluation these types of survey outcomes are widely considered valid within the field. The cross-sectional design means that there remains some uncertainty about cause and effect in the results observed although when considered in the broader context, the findings are not inconsistent with the existing literature in this field. Standard checks for completeness of the specimens based upon urine volume and urine creatinine excretion were done but it remains likely that some urine samples were over-collections and others were under-collections. Para-aminobenzoic acid (PABA) has been used as a marker for completeness in overseas studies , but it is not without limitations  and currently not approved for use in Australia . We believe that any errors in the included data are likely to be random and while this should not affect the mean group level estimates of salt intake, it could have reduced the power to detect associations between exposures and outcomes. In addition, because only one estimate was collected for each individual, and because there is significant within person variation in salt consumption from day to day, the variability of salt consumption within the population is likely to have been over-estimated. In turn, this means that the estimated proportions of individuals meeting the various salt targets (all of which lie below the population mean) are likely to have been over-estimated. It is likely that social approval bias [30, 31] resulted in more socially desirable responses but we do not believe this is likely to have happened differentially across population subgroups and think it is unlikely to have substantively confounded our results. The survey was restricted to one regional town with a below average socio-economic index for area (SEIFA) score which raises concerns about the generalizability of the data. The SEIFA score is derived from attributes that reflect disadvantage such as low income, low educational attainment, high unemployment and jobs in relatively unskilled occupations . However, comparable levels of knowledge and reported behaviours in other Australian surveys [32–38] using a range of different instruments [15–17] suggest that the findings are likely also to be valid outside of Lithgow. The constancy of the results across unadjusted and confounder-adjusted models, and in each of the study sub-populations, also argues for a fairly robust finding.
In conclusion, many of the participants surveyed were knowledgeable about the adverse effects of salt and reported efforts to ameliorate the risks caused by excess salt consumption. The persisting high levels of salt consumption across the population and the absence of any detectable association of knowledge levels with actual salt consumption are a significant concern. These data strongly suggest that education alone will be ineffective in reducing population salt consumption levels and that education programs must be supported by interventions that change the food environment in ways that encourage population-wide behaviour change. The food industry will play a central role in achieving population-wide salt reduction through the process of food reformulation, which has a key advantage over targeted behavioural and education interventions in that it can be delivered and sustained at scale  This is by no means a message unique to salt reduction efforts, with programs targeting obesity and alcohol consumption identifying similar limitations for programs based purely on education [40, 41].