Objective To evaluate population health benefits and cost-effectiveness of interventions for reducing salt in the diet.
Design Proportional multistate life-table modelling of cardiovascular disease and health sector cost outcomes over the lifetime of the Australian population in 2003.
Interventions The current Australian programme of incentives to the food industry for moderate reduction of salt in processed foods; a government mandate of moderate salt limits in processed foods; dietary advice for everyone at increased risk of cardiovascular disease and dietary advice for those at high risk.
Main Outcome Measures Costs measured in Australian dollars for the year 2003. Health outcomes measured in disability-adjusted life years (DALY) averted over the lifetime.
Results Mandatory and voluntary reductions in the salt content of processed food are cost-saving interventions under all modelled scenarios of discounting, costing and cardiovascular disease risk reversal (dominant cost-effectiveness ratios). Dietary advice targeting individuals is not cost-effective under any of the modelled scenarios, even if directed at those with highest blood pressure risk only (best case median cost-effectiveness A$100 000/DALY; 95% uncertainty interval A$64 000/DALY to A$180 000/DALY). Although the current programme that relies on voluntary action by the food industry is cost-effective, the population health benefits could be 20 times greater with government legislation on moderate salt limits in processed foods.
Conclusions Programmes to encourage the food industry to reduce salt in processed foods are highly recommended for improving population health and reducing health sector spending in the long term, but regulatory action from government may be needed to achieve the potential of significant improvements in population health.
- Public health
Statistics from Altmetric.com
Cardiovascular disease is the biggest cause of mortality in the world; more than 17 million people died of cardiovascular disease in 2004, and this is projected to increase to more than 23 million by 2030.1 High blood pressure is the leading cause of mortality from cardiovascular disease,2 with elevated risks of disease from a systolic blood pressure as low as 115 mm Hg.3 4 Diseases related to blood pressure above this level contribute approximately 10% of global healthcare expenditure.5 Diets high in sodium have been linked to high blood pressure levels6 7 and increased risk of stroke and cardiovascular outcomes.8 9 Many countries now recommend restricting daily sodium intake to 100 mmol (approximately 6 g of table salt) or less,10–12 but in a recent review of world salt levels, only seven out of the 25 countries reviewed met this goal.13 A forum on salt, chaired by the WHO in 2006, stressed the need for immediate action and urged countries to implement interventions to reduce dietary intake of salt.12
In countries where a western-style diet is eaten, approximately three-quarters of all dietary salt intake comes from salt added during food processing.14 15 Advice from a dietitian can be effective in reducing salt intake, but effects attenuate over time, and resulting changes in blood pressure may be too small to provide substantial health benefit at the population level.16 It is important, therefore, to work with the food industry in encouraging changes in food processing that will reduce population exposure to salt.17 Many countries, including Finland, the USA, the UK, Canada, France, Australia and New Zealand, have implemented salt reduction programmes based on product reformulation, labelling and/or public education to inform and encourage change in individual dietary choices and food industry practices.18 19 The population health benefits, however, are largely unknown; there is evidence that average sodium levels have successfully decreased in Finland and the UK, but increased in the USA, and there is a lack of data to determine trends in other countries.18
A number of studies have estimated that interventions aimed at changing food industry practices will be cost-effective or even cost-saving if salt reduction targets (eg, 15–50% reduction in average daily intake) can be achieved.20–23 Simulations in the USA even suggest that achieving a salt reduction of 3 g/day in the American diet (approximately 29% of daily intake for men and 41% for women) could benefit public health more than targeting tobacco, weight loss or cholesterol levels, and prevent as many deaths as antihypertensive drugs would prevent if given to everyone with hypertension.20 It is not clear, however, if salt reductions of this magnitude are achievable with current measures that rely on voluntary changes by food manufacturers and informed choice by consumers. It is argued that legislative enforcement of salt limits in food processing are needed to achieve meaningful change in population salt levels.24 25
In this research, we evaluate a range of salt reduction interventions, focusing on prevention of cardiovascular disease in Australia. We consider strategies ranging from those that aim to change individual dietary behaviour, to the current programme of incentives for voluntary changes by food manufacturers, to a more paternalistic approach with government legislation of more moderate salt levels in processed foods. For each strategy, we consider the potential benefits to population health and evaluate intervention cost-effectiveness to determine which approaches are a good investment for the heath sector.
We evaluate four salt reduction interventions: (a) the current practice ‘Tick’ programme19, which provides incentives for voluntary changes by food manufacturers; (b) legislation and enforcement to make Tick salt limits mandatory for food manufacturers; (c) dietary advice for everyone at increased risk of cardiovascular disease (systolic blood pressure >115 mm Hg); and (d) dietary advice for those at high risk (systolic blood pressure >140 mm Hg). We evaluate the cost-effectiveness of each intervention in comparison to what would occur if none of the four interventions for reducing dietary salt intake were in place. Interventions are evaluated using the Assessing Cost-Effectiveness method of economic evaluation,26 which has been widely used for evaluating interventions and informing policy in an Australian healthcare context.27–29
In the following sections we provide a brief summary of the intervention programmes and cost-effectiveness modelling methods. Further details of modelling methods and all intervention input parameters are provided in an appendix (available online only).
Food industry interventions
The Tick programme for salt reduction has been run by the Australian Heart Foundation (http://www.heartfoundation.org.au) since 1989. The programme encourages food manufacturers to reduce the salt content of products voluntarily, in return for the right to buy an endorsed logo that can be used on product packaging to achieve a marketing advantage. The royalties received from food manufacturers cover the costs of monitoring product compliance, programme administration, evaluation, education and health promotion.19 Similar programmes are running in New Zealand (Pick the Tick; http://www.nhf.org.nz), South Africa (Heart Mark programme; http://www.heartfoundation.co.za) and Canada (Health Check symbol; http://www.heartandstroke.com). The programme is preventive, thus the recommended limits on salt content are more moderate than the low levels recommended for people with pre-existing disease (<120 mg sodium per 100 g).30 In New Zealand the programme is estimated to have reduced the average sodium content of bread (<450 mg sodium per 100 g) by 26%, margarine (<400 mg sodium per 100 g) by 11% and breakfast cereal (<400 mg sodium per 100 g) by 61%.19
To evaluate the cost-effectiveness of the current Tick programme, we derive total sodium reduction, by age and sex, from the estimates of sodium reduction in each product, the tonnage of products sold,19 and average per capita daily consumption data.31 We determine the total amount of sodium removed from the food supply from the reduction in breads, margarines and breakfast cereals (the products for which we have evidence from the Tick study).19 From a comparison of reported daily intake32 and apparent consumption (the combined total of commercial production, home production, imports and opening stocks, adjusted for loss through exports, processed food manufacture, non-food use, waste and closing stocks),33 we estimate that 20% of manufactured bread, margarine and breakfast cereal products are not consumed. Programme costs are evaluated from the number of participating products19 and average annual cost per product. We do not include any food manufacturer costs associated with adding less salt.
To evaluate the cost-effectiveness of developing and enforcing legislation to make salt limits mandatory, we determine the total sodium reduction that would occur in the population if all cereal, bread and margarine manufacturers made the salt reductions that have been observed in the Tick evaluation study.31 Costs of changing legislation to enforce salt limits are derived from WHO estimates of the number of resources used in making and enforcing legislative changes34 and the unit costs of the resources.35
Dietary advice interventions
Providing dietary advice aims to change individual behaviour around eating and purchasing food. A Cochrane review of seven randomised controlled trials (2166 participants) found that counselling from a dietitian is effective in reducing salt intake and lowering blood pressure in both normotensive and hypertensive participants.16 Meta-analysis showed a reduction in urinary sodium excretion of 48.9 mmol/24 h (95% CI 32.5 to 65.4) at 6–12 months after the start of the intervention and 35.5 mmol/24 h (95% CI 23.9 to 47.2) at 13–60 months, but no significant effect thereafter.16
We evaluate cost-effectiveness of two intervention scenarios: offering dietary advice to everyone at risk (systolic blood pressure >115 mm Hg) and offering dietary advice to people at high risk (systolic blood pressure >140 mm Hg). Four of the seven trials in the meta-analysis (which include 96% of overall participant numbers at 12 months and 100% of overall participant numbers at 13–60 months) are of similar intervention design. We base our calculation of resource use on the TOHP II trial,36 which best represents the meta-analysis trials in terms of costs (neither most nor least resource intensive) and effects (similar sodium reduction to overall meta-analysis results). The resources include mail-out promotion, an individual counselling visit and 17 group sessions for each eligible participant. We assume that the costs of recruitment and dietitian delivery of counselling sessions accrue to the government. We also include the costs to patients for their time and travel in attending intervention sessions.
The distribution of systolic blood pressure is derived at each age and sex from measurements in Australia's AusDiab survey.37 The change in systolic blood pressure distribution with each intervention is then derived from the intervention change in sodium intake, using regression models derived by Law et al6 for economically developed countries. We also calculate the maximum possible change in blood pressure that could occur with a reduction in dietary salt intake (figure 1). This is derived from the difference between the recommended limit of 100 mmol/day11 and Australian sodium excretion levels of 170 mmol/day (SD 52) for men and 118 mmol/day (SD 42) for women.38
The effects of blood pressure changes on the incidence of ischaemic heart disease and stroke are quantified by the potential impact fraction39 using relative risks of ischaemic heart disease and stroke from the Prospective Studies Collaboration4 and assuming a theoretical minimum systolic blood pressure of 115 mm Hg.40 We assume that all of the preventive effects of blood pressure reduction are fully realised within one year, based on the results of a recent review of randomised trials evaluating the effects of blood pressure-lowering drugs on heart disease and stroke events.41 For comparison with WHO salt analyses,34 however, we also evaluate cost-effectiveness ratios with a lag in the reversal of stroke and heart disease risks, based on WHO assumptions of full reversal of stroke risk after 3 years, and two-thirds reversal of heart disease risk after 3 years, with the remaining heart disease risk reversed over seven subsequent years.
All cardiovascular disease data and measures of disease severity are sourced from Australia's burden of disease and injury study.42 We assume a gradual decline in incidence and case fatality rates for the first 20 years of analysis, based on Australian projections,42 but assume rates remain constant thereafter.
Proportional multistate life-table analysis43 is used to determine changes in cardiovascular disease mortality and morbidity over the lifetime of the population (in 5-year age group cohorts). From these changes in morbidity and mortality, we determine the total disability-adjusted life years (DALYs) averted by each intervention and the total DALYs that could be averted if mean population salt intake were reduced to recommended levels.
The cost of each intervention is offset by changes in treatment costs for all modelled diseases. Costs of healthcare for all other diseases also accrue in added years of life. We derive these costs from the Australian Institute of Health and Welfare Disease Costs and Impacts Studies data.44 All costs are adjusted to Australian dollars for the year 2003 using Australian health price deflators,45 consumer price index46 and/or purchasing power parities47 where relevant (A$1 in 2003 ≅ US$0.74 in 2009 ≅ €0.63 in 2009).48
We model all costs and population health outcomes over the lifetime of the population aged 30 years and older from a baseline year of 2003, discounting all future costs and health outcomes at 3%.49 All costs and health outcomes are then summed over the lifetime to determine the cost-effectiveness ratio, in dollars per DALY averted, for each intervention. Interventions are compared with an Australian cost-effectiveness threshold of A$50 000 per DALY.50–52 To quantify the uncertainty in cost-effectiveness outputs, we use Monte Carlo analysis to derive 95% uncertainty intervals for all outcome measures and to determine probabilities of intervention cost-effectiveness, using @Risk (Palisade, version 4.5). Modelling input parameter values, including uncertainty distributions, sources and assumptions, are presented in the accompanying appendix (table 5), available online only.
The interventions are evaluated from the perspective of the Australian health sector, but because interpretations of perspective can vary, we evaluate cost-effectiveness ratios for a number of different costing scenarios. The first scenario includes only costs to the government or food industry to run or participate in an intervention programme, a second scenario also includes costs to patients for time and travel in participating (eg, for attending dietary advice counselling sessions), and a third scenario further includes costs of health care for unrelated diseases in added years of life.
A total of 610 000 DALY (95% uncertainty interval 480 000 to 740 000 DALY) could be averted over the lifetime of the population if everyone reduced their salt intake to recommended limits (table 1). Providing dietary advice to the population with hypertension (systolic blood pressure >140 mm Hg) or everyone at risk (systolic blood pressure >115 mm Hg) would avert less than 0.5% of this disease burden, and the Tick programme (as modelled for breads, margarines and cereals) less than 1%. Making Tick limits mandatory for all bread, margarine and cereal products, however, could avert 18% of the disease burden, which is 20 times the health gain achieved with the voluntary approach.
Targeting salt reduction intervention at the food industry is less expensive for the health sector than targeting dietary advice at individuals, an approach that also involves a substantial cost contribution from participating patients (table 2). Potential healthcare costs for treating diseases other than ischaemic heart disease and stroke in added years of life are large for all interventions, but they are in all cases offset by even larger reductions in the costs of treating ischaemic heart disease and stroke.
Both the voluntary Tick programme and mandatory salt reduction have a 100% probability of being dominant (ie, cost-saving to the health sector) under all modelled scenarios of discounting, costing and reversal of risk (table 3). In contrast, there is zero probability of dietary advice being cost-saving or cost-effective against the A$50 000 per DALY threshold, even when targeted at the population that is most at risk (systolic blood pressure >140 mm Hg).
For the prevention of cardiovascular disease, providing dietary advice to everyone with elevated blood pressure (or even just those at highest risk) will not achieve large improvements in population health and is not a cost-effective strategy for health sector investment. Programmes to encourage the food industry to reduce salt in processed foods, on the other hand, are an excellent investment; they improve population health and will reduce health sector spending in the long term. However, although the current incentive-based programme is effective in encouraging Australian manufacturers to reduce salt in processed foods voluntarily, government intervention to make moderate salt limits mandatory for all manufacturers could achieve 20 times the health benefits for the Australian population.
There are some uncertainties about the effectiveness of food industry interventions in removing salt from the diet, which stem from the difficulties in evaluating population-targeted interventions in randomised controlled trials.53 In this analysis, we have derived effectiveness from a study comparing salt content in reformulated products with the average salt content of products in the same category, along with recorded product sales over 12 months.19 We adjust for food that is purchased but not consumed, but the behavioural response of individuals to changes in product saltiness (eg, adding salt or eating more of other salty foods) was not measured in the intervention study. However, the modelled interventions only involve reducing salt to moderate levels rather than the unpalatable low levels recommended for those with pre-existing disease. There is evidence that consumer tastes gradually adapt to lower salt products as salt intake lessens.54 55 It has also been shown that reductions of up to 20% of sodium in bread cannot be detected56 and do not affect consumer acceptance of the product.57 In addition, when reformulating cereal products to meet the Tick criteria, the Kellogg company did not allow a product to be released on to the market unless consumer testing showed that it matched the existing product on overall liking and purchase intent.58 Overall, the evidence suggests that consumers are unlikely to compensate for the moderate reduction in dietary salt intake and, therefore, that targeting the food supply chain is an effective way to reduce population salt levels.
The greater cost-effectiveness of strategies targeting the whole population through the food supply, compared with strategies targeting individual behaviour, is not surprising. We have observed similar outcomes for interventions targeting physical activity (eg, mass media-based community campaign vs exercise physiologist advice)50 and alcohol misuse (eg, taxation vs general practitioner brief advice),59 and similar trends have been noted in WHO analysis of risk factor interventions globally.21 As Geoffrey Rose advocated, shifting the whole distribution of risk by targeting the whole population may achieve better long-term outcomes for public health than truncating the risk distribution by targeting individuals at high risk.60 There is growing evidence that targeting whole populations may also be more cost-effective.
Previous studies of food industry interventions have been hypothetical, based on the assumption that salt reduction targets will be achievable.20–23 These targets are typically between a 15%21 and 50%23 reduction in salt intake, which is substantially higher than our 0.2% calculation of benefit from the current Tick programme in Australia and higher than the potential 5% reduction from mandatory salt reduction. It is not surprising that the interventions do not meet salt reduction target levels, given that evidence of effectiveness is available only for bread, margarine and cereal products. Nevertheless, these are staple products of the Australian diet; 99% of the population eat bread, 83% use a spread on their bread such as margarine, and more than 90% eat cereal.61 It highlights the need for salt limits on a wide range of foods that are regularly consumed, including restaurant and take-away foods (which account for 30% of Australian food expenditure),62 if we are to meet global targets for dietary salt intake.
A recent study from the USA investigated the potential for either taxation or food industry collaboration strategies to reduce salt intake in the American diet and thus reduce healthcare costs.63 Changing taxation can be a very cost-effective strategy for reducing the consumption of alcohol and cigarettes.64 65 It is not, however, a feasible strategy for reducing the consumption of salt, which is a very low-cost ingredient for manufacturers. However, the study by Smith-Spangler et al63 did find that the current UK programme to encourage voluntary salt reductions by the food industry, if delivered in the USA, could lead to substantial improvements in cardiovascular health, averting as much as US$32 billion in US healthcare costs. Our analyses suggest that these figures could be many times higher if the limits were made mandatory for food manufacturers. In an important step, the US Institute of Medicine, in their recent report on strategies to reduce sodium intake in the USA, recommended that voluntary strategies be considered only as an interim measure, with a primary recommendation to set mandatory national standards for salt content in both processed foods and foods prepared outside the home.66
This touches on an important ethical issue about the responsibility of the state to safeguard health compared with the rights of the individual to make their own health choices. In our analyses, increasing paternalism is associated with increasing population health benefits. Leaving it to individuals to make their own choices about dietary behaviour (even with intensive advice) leads to limited health benefits. The current Australian programme of food industry incentives, with Tick symbol guidance to consumers in picking healthy products, leads to somewhat more benefit, but making salt levels mandatory clearly leads to substantially better health for the population. But is this degree of paternalism justified? It can be argued that there are many who could benefit; 94% of Australian men and 64% of Australian women exceed the recommended salt intake,38 and by reducing salt at the manufacturing stage, those who benefit will include children and those most disadvantaged in accessing healthcare services. In addition, unlike the regulation of other food ingredients, such as adding folic acid to bread or iodine to salt, the intervention does not seek to add something to the food supply but merely to reduce a commonly used ingredient to more moderate levels. There is no evidence of harm from small and gradual salt reductions. Salt is added to processed foods for many reasons; to entice further consumption, to bulk a product up cheaply by increasing water-holding capacity, or to promote drink sales by increasing thirst.67 Food manufacturers have a responsibility to make money for their shareholders, but they also have a responsibility to society. If corporate responsibility fails, maybe there is an ethical justification for government to step in and legislate.
The authors would like to acknowledge Jan Barendregt's assistance with modelling techniques.
Funding This study was funded by the Australian National Health and Medical Research Council (project ID no 351558) http://www.nhmrc.gov.au/.
Competing interests None.
Ethics approval This study was approved by the Behavioural and Social Sciences Ethical Review Committee of the University of Queensland in accordance with the National Health and Medical Research Council guidelines (clearance no. 2004000796).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.