Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Health and Economic Impacts of Eight Different Dietary Salt Reduction Interventions

  • Nhung Nghiem,

    Affiliation Department of Public Health, University of Otago, Wellington, Wellington South, New Zealand

  • Tony Blakely,

    Affiliation Department of Public Health, University of Otago, Wellington, Wellington South, New Zealand

  • Linda J. Cobiac,

    Current Address: British Heart Foundation Centre on Population Approaches to NCD Prevention, Oxford University, Oxford, United Kingdom

    Affiliations Department of Public Health, University of Otago, Wellington, Wellington South, New Zealand, School of Population Health, University of Queensland, Brisbane, Australia

  • Amber L. Pearson,

    Current Address: Department of Geography, Michigan State University, East Lansing, Michigan, United States of America

    Affiliation Department of Public Health, University of Otago, Wellington, Wellington South, New Zealand

  • Nick Wilson

    nick.wilson@otago.ac.nz

    Affiliation Department of Public Health, University of Otago, Wellington, Wellington South, New Zealand

Abstract

Background

Given the high importance of dietary sodium (salt) as a global disease risk factor, our objective was to compare the impact of eight sodium reduction interventions, including feasible and more theoretical ones, to assist prioritisation.

Methods

Epidemiological modelling and cost-utility analysis were performed using a Markov macro-simulation model. The setting was New Zealand (NZ) (2.3 million citizens, aged 35+ years) which has detailed individual-level administrative cost data.

Results

Of the most feasible interventions, the largest health gains were from (in descending order): (i) mandatory 25% reduction in sodium levels in all processed foods; (ii) the package of interventions performed in the United Kingdom (UK); (iii) mandatory 25% reduction in sodium levels in bread, processed meats and sauces; (iv) media campaign (as per a previous UK one); (v) voluntary food labelling as currently used in NZ; (vi) dietary counselling as currently used in NZ. Even larger health gains came from the more theoretical options of a “sinking lid” on the amount of food salt released to the national market to achieve an average adult intake of 2300 mg sodium/day (211,000 QALYs gained, 95% uncertainty interval: 170,000 – 255,000), and from a salt tax. All the interventions produced net cost savings (except counselling – albeit still cost-effective). Cost savings were especially large with the sinking lid (NZ$ 1.1 billion, US$ 0.7 billion). Also the salt tax would raise revenue (up to NZ$ 452 million/year). Health gain per person was greater for Māori (indigenous population) men and women compared to non-Māori.

Conclusions

This study substantially expands on the range of previously modelled salt reduction interventions and suggests that some of these might achieve major health gains and major cost savings (particularly the regulatory interventions). They could also reduce ethnic inequalities in health.

Introduction

The risk factor of a “diet high in sodium” is one of the top two dietary risk factors for disease burden identified in the Global Burden of Disease Study 2010 [1]. Indeed, this sodium risk factor alone was ranked 11th globally out of all risk factors considered (for a counterfactual of 1000 mg/day sodium intake). The scale of this problem has resulted in calls for “salt reduction” to be considered a public health priority, with it included in the top five priority actions for non-communicable disease (NCD) control internationally [2] and for reducing NCD inequalities [3]. Furthermore, in 2012 the World Health Organization (WHO) recommended a “reduction to <2 g/day sodium (5 g/day salt) in adults (strong recommendation)” [4].

Despite the above, the evidence relating to sodium and health is still perceived by some as controversial [5]. To some extent this relates to the uncertainty around the health benefits and risks of reducing sodium intakes below the 2300 mg level which is relatively low compared to current consumption (e.g., an Institute of Medicine Report [6]). But some commentators blame the perception of controversy on inadequate consideration of study limitations by authors, poor reporting by the media, and vested commercial interests exploiting the situation [7]. A recent example of a study casting doubt on aspects of the sodium and health relationship was published in 2014—the PURE Study [8]. But this observational study may have had limitations around the reliability of spot urine tests [9] and other issues [10], including reverse causation which could not be ruled out (as noted by the authors themselves). Indeed, what is more critical to consider is the totality of the evidence from published systematic reviews [11,12] and specifically: (i) the long-term trial data suggesting reduced sodium intake reduces cardiovascular (CVD) risk [13,14]; (ii) trial data showing this benefit for CVD mortality (albeit reducing sodium while raising potassium [15]; and (ii) data on reduced CVD risk in a nested observational study where participants had been randomised to a Mediterranean diet [16].

There is also a growing base of modelling studies which have considered the health gain and/or the economic aspects of dietary sodium reduction (see S1 file). Most of the published health economic evaluations indicate that sodium reduction interventions are likely to result in health gains while actually being cost-saving (e.g., by averting future health system costs). But there remains scope for methodological improvements in many of these studies, particularly around the need for more robust cost data and for better definitions around the interventions (Table A in S1 File).

Reducing health inequalities is also a common goal of health policy. Salt reduction has the potential to reduce inequalities in health, in absolute terms at least, due to the usually higher CVD rates among disadvantaged populations within countries—but this issue has not been well studied in modelling work to date (we identified just one study [17]). Also certain types of interventions to reduce sodium intakes have rarely been studied if at all. For example, there have been no studies on restricting the supply as per a cap-and-trade system on the food salt supply (as suggested elsewhere [18,19], and modelled with regard to sugar [20]), only one economic modelling study on a specific salt tax [21]; and rarely around labelling interventions [22]).

Given this picture, we aimed to advance the understanding of the effectiveness and cost-effectiveness of sodium reduction interventions by new modelling work. Our work benefited from a range of methods refinements and use of relatively good country-level disease data, which captures existing health inequalities by ethnicity (in a country, New Zealand, where these are prominent). We also used detailed individual-level health system cost data, which has recently become available for use in New Zealand. A total of eight different interventions were considered, which allowed comparisons between those used previously in New Zealand and the United Kingdom (UK), with more hypothetical ones.

Methods

Model structure and perspective

A model developed for studying CVD interventions in Australia [23], provided the base model. This model was built in Excel but, for our study as part of the BODE3 Programme of work (http://otago.ac.nz/bode3), we converted it to a Markov macro-simulation model in TreeAge Pro version 2013. Results identical to the Excel model were obtained in TreeAge, prior to populating the model with New Zealand data. The simulated population was a closed cohort of the New Zealand population aged 35 years and older (2.3 million people), modelled from the baseline year (2011) to death or age 100 years.

The Markov model has four primary health states, with annual transition rates capturing incidence and case-fatality for coronary heart disease (CHD) and stroke events (see the diagram in an online Technical Report [24] on the BODE3 website). Essentially, proportions of each age/sex/ethnicity cohort occupy the states of: being “healthy” (i.e., not having CVD), having a form of CVD (CHD or a type of stroke), or death, in each annual cycle.

In terms of modelling background disease trends we took the same approach as the New Zealand Burden of Disease Study (NZBDS) [25], and assumed a continued decline in incidence rates for both CHD and stroke of 2.0% annually, and also a 2.0% reduction in case-fatality annually (i.e., reflecting improved treatment and management). We extended this projection from 2016 (NZBDS end estimate) to the year 2026 and then held the incidence and case fatality rates constant.

Background population mortality was assumed to decline at a somewhat lower rate than for CVD with a 1.75% annual reduction for non-Māori, and 2.25% for the indigenous population of Māori (also out to the year 2026), then 0% per annum decline for both ethnic groupings thereafter. The justification for these trends is detailed in our Protocol [26].

A health system perspective was used. Costs and benefits beyond the health system (e.g., productivity gains from preventing premature deaths of workers) were considered out of scope as these are more relevant to a societal perspective. However, additional health system costs arising from extra life expectancy in the future attributable to the impact of the modelled interventions were included in the baseline analyses. Costs were calculated in 2011 New Zealand dollars and a 3% discount rate was applied to costs and future health gain (with the discount rate varied in scenario analyses: 0% and 6%). OECD 2011 purchasing power parties [27] were used for calculating results in US$ for international comparisons.

Our approach to cost-effectiveness analysis was that of the “generalised cost-effectiveness analysis” as developed for the WHO [28]. In this approach, all interventions (including current practice) are evaluated against a theoretical “do nothing” comparator (i.e., doing none of the interventions of interest in the analysis). This approach allows explicit estimation of the cost-effectiveness of current practice (if included as intervention), and so it avoids artificially making an intervention look more favourable if compared against inefficient current practice. Therefore, we back-calculated disease rates under the “do nothing” scenario using the same parameters of intervention effectiveness, adherence and costs that are used in the cost-effectiveness analyses (in this case for the Dietary Counselling and the Endorsement Label Programme interventions—as detailed below).

Input parameters

Input parameters shown in Table 1 are summarised in the text below, and also explained in further detail in an online Technical Report.[24].

thumbnail
Table 1. Input parameters to the modelling: selected baseline and epidemiological parameters

https://doi.org/10.1371/journal.pone.0123915.t001

Incidence, prevalence and case-fatality.

The estimated incidence, prevalence and case-fatality rates of CHD and stroke (ischaemic and haemorrhagic) were calculated across all combinations of sex, age-group (35–39, 40–44, … 95+ years) and ethnicity (Māori; and non-Māori). Data came from Ministry of Health data, called ‘HealthTracker’ [29], which is a collection of linked administrative datasets of publically-funded health system events. This includes hospitalisations, mortality, cancer registrations, mental health and addiction service use, pharmaceutical and laboratory claims, primary health care enrolment, and outpatient/emergency department visits for the entire New Zealand population with costs attached. But gaps in HealthTracker data exist in specific areas (e.g., some private sector expenditure and the health-related aspects of residential care) and so we scaled up both the CVD disease costs (CHD and stroke) and the annual health system costs for the non-diseased population. For the disease costs we scaled up HealthTracker costs across all age groups by 1.2. For the non-diseased population, costs were multiplied by 1.1, 1.2, 1.3 for the 65–74, 75–84 and 85+ age groups respectively to capture the estimated missing data of funding residential ‘disability support services’ care funded by the government (Vote:Health), but not yet captured in available data. All costs include the costs in the last six months of life.

Validation of model parameters and the final model outputs (relative to two official data sources) are detailed in an online Validation Report [30]. This additional work also involved parameter coherence checking, using the epidemiological software program DisMod II [31]. Of note, because of a 10 year look back period we could use in data analyses, our empiric estimates of prevalent disease are probably low (as we do not capture earlier incidence cases with no subsequent health event), and therefore our estimated case-fatality rates may be too high (as the ‘prevalent’ denominator is too low) when applied to our Markov model that projects out multiple decades. The DisMod checks possibly supported this concern for stroke. Therefore, we include reduced case-fatality rates as a scenario analysis (see below).

Morbidity and disability weights.

Overall morbidity, by sex, age and ethnicity, was quantified in the model using the years of life lived with disability (YLDs) from the NZBDS [25], divided by the population count to give ‘prevalent’ YLDs. Disease-specific morbidity was assigned in each disease state (e.g., CHD and stroke), as the total comorbidity-adjusted YLDs for that disease divided by the prevalent population. The health status valuation used to calculate these YLDs were disability weights derived from the Global Burden of Disease study (GBD2010) using pair-wise comparisons from multi-country surveys [32], as opposed to, say, disutilities from the EuroQol. These disability weights are on a scale from 0 (full health) to 1.0 (death)—and included uncertainty (for details see the online Technical Report [24]). As per other BODE3 work we assumed no future underlying trend in morbidity burdens (i.e., both the size of the weights and the background level of non-CVD morbidity were assumed constant into the future). Of note is that the use of these weights limited the maximum QALYs that would be gained with increasing age. For example, an average Māori woman aged 60–64 has an expected level of disability of 0.288, meaning a year of life gained in this population group has a maximum value of 0.712. QALYs were cumulatively tallied for the life-span of the modelled cohort.

Intervention specification and parameters.

We considered eight different interventions of which some were voluntary (e.g., dietary counselling, a labelling programme and a campaign run in the UK) and others were mandatory (requiring national laws for: legal limits on sodium in processed foods, a salt tax, and a sinking lid on the supply of salt to the New Zealand market). The details of these are in Table 2 and in Table B in S1 File.

In brief, for each of these interventions a reduction in sodium intake was linked to a reduction in systolic BP based on values derived from the regressions models developed by Law et al [33]. A reduction in systolic BP was then linked to a reduced probability of adverse health outcomes as per a meta-analysis of 61 prospective studies by Lewington et al [34] (i.e., though we also used the results of another meta-analysis in a scenario analysis [35]).

thumbnail
Table 2. Input parameters relating to the interventions effects (for further details see S1 File)

https://doi.org/10.1371/journal.pone.0123915.t002

Costing of intervention scenarios and health system costs.

We considered the net cost, which is the intervention costs plus health system costs throughout the lifespan of the modelled cohort (i.e., we captured additional health costs associated with any extra lifespan generated by the interventions). Specific details for the costing of the interventions are provided in Table 3. For health system costs, the ‘business as usual’ ones were determined by strata of sex and age using HealthTracker data, which links cost estimates to all health events. From this dataset it was possible to calculate the 2011 costs for the first year of CHD and stroke, and then the average annual cost for the second and subsequent years (S2 file). Furthermore, given that CVD is a relatively important part of baseline health system costs, we adjusted the baseline health system costs experienced by the “healthy” component of the modelled population, to remove the CVD-attributable cost component (to avoid double-counting).

thumbnail
Table 3. Input parameters relating to the interventions costs.

https://doi.org/10.1371/journal.pone.0123915.t003

As further context, New Zealand is a fairly typical OECD country in terms of health spending (at 10.0% of GDP—slightly more than the OECD average of 9.3%) [36]. But 83% of health spending was funded by public sources in 2011 (which is well above the average of 72% in OECD countries). Residential care for the elderly in New Zealand is largely funded from social welfare budget (and so is excluded from our analysis—given the health system perspective). Nevertheless, the residential care costs that relate specifically to health (i.e., residential care hospital facilities) is captured in our analysis (via scaling up from HealthTracker costs—see above).

Future trends in health costs were not modelled as these are considered very uncertain due to reasons around the New Zealand economy’s dependency on commodity prices, recent expansion in the role of the government’s pharmaceutical purchasing agency, and potential future trade agreements that might limit the government’s capacity to constrain health costs.

Scenario and uncertainty analyses

We reran models (usually for expected values only) for a wide range of scenarios to assess the impact of components of the interventions and other structural assumptions (e.g., the discount rate). We also varied the background case-fatality rates for CVD (see above) and undertook a range of one-way uncertainty analyses and derived Tornado plots, using the 2.5th and 97.5th percentile values of input parameters. This was to assess which input parameter uncertainty contributed the most to uncertainty in the model outputs i.e., QALYs, net cost and incremental cost-effectiveness ratio (ICER).

A scenario analysis relating to equity considerations involved using for the Māori population, the lower background morbidity and mortality of the non-Māori population. This approach meant that Māori are not considered to be “penalised” in terms of the scope for future health gain due to poorer background health status relative to the non-Māori population. Further justification of such an approach has been detailed previously [37].

Interpretation of cost-effectiveness.

There is no universally accepted threshold in the New Zealand setting for describing an ICER as being “cost-effective” or not. So we relied on WHO recommendations relating to GDP per capita [38], and used a nominal GDP per capita of NZ$45,000 in 2011 (US$29,600).

Results

The largest health gain was from the potential intervention of a Sinking Lid in food salt released to the market to achieve an average adult intake of 2300 mg sodium/day (Figs 1 and 2 and Table 4). It achieved 211,000 QALYs gained (95% uncertainty interval [UI]: 170,000–255,000). This QALY benefit was followed in descending order by that from a: (i) Salt Tax (195,000 QALYs gained); (ii) mandatory 25% reduction of sodium levels in processed food (“Mandatory-All”), (110,000); (iii) the package of interventions performed in the UK (85,100); (iv) mandatory 25% reduction in sodium levels in bread, processed meats and sauces (“Mandatory-3G”), (61,700); (v) Media Campaign as per the UK one (25,200); (vi) the voluntary Endorsement Label Programme as currently used in New Zealand (7900); and (vii) Dietary Counselling as currently used in New Zealand (200 QALYs gained).

thumbnail
Fig 1. Cost-effectiveness plane with the eight salt-reduction interventions for the New Zealand adult population.

https://doi.org/10.1371/journal.pone.0123915.g001

thumbnail
Fig 2. Cost-effectiveness plane with further detail on four of the salt-reduction interventions on the New Zealand adult population (for comparisons with the other interventions—see Fig 1).

https://doi.org/10.1371/journal.pone.0123915.g002

thumbnail
Table 4. Population level results for the cost, health gain and cost-effectiveness of the interventions (95% uncertainty intervals in parentheses)a.

https://doi.org/10.1371/journal.pone.0123915.t004

The Sinking Lid produced the highest discounted net savings of NZ$ 1.1 billion (US$ 0.7 billion) over the lifetime of the population. However, the Salt Tax was both cost-saving (NZ$ 1.0 billion) and would also actually raise NZ$ 452 million in revenue per annum by 2021. The only intervention not found to be cost-saving was Dietary Counselling. Nevertheless, it was still typically cost-effective with a mean ICER of NZ$ 36,900 (US$ 22,300) per QALY gained (95% UI: NZ$ 22,400–62,500).

Table 5 shows the overall cost results were largely driven by averted disease treatment costs for CVD, followed by the increased health system costs from extra life lived (as a result of the interventions).

thumbnail
Table 5. Types of costs (NZ$) by intervention (expressed per adult in 2011).

https://doi.org/10.1371/journal.pone.0123915.t005

Heterogeneity by socio-demographics

The QALYs gained were higher and the cost-savings greater for younger age groups (<65 years) for all interventions except for Counselling (Table C in S1 File). The same pattern existed for the greater health benefit and greater cost-saving for men compared to women. In contrast to the other population groups, Counselling was not cost-effective for older ages (65+ years) and for women (Table C in S1 File). For all the interventions there was greater health benefit for Māori compared to non-Māori (e.g., 1.3 times more QALYs gained for the Mandatory-All intervention) (Table C in S1 File).

To facilitate a more detailed ethnic inequalities analysis, we present both model-estimated CVD mortality rates in 2021 (i.e., once the interventions have been operational for 10 years and have had a chance for their impact to play out and are fairly stable) and QALYs gained per person, by strata of sex, age and ethnicity in Table 6. This analysis used both the Counselling intervention and one of the more plausible of the hypothetical interventions: the Mandatory-All intervention. As per Table 6, the Counselling intervention had a negligible impact. The Mandatory-All intervention reduced CVD mortality rates more in absolute terms among Māori, but less in relative terms, resulting in estimated decreases in rate differences between Māori and non-Māori—but increases in the rate ratio. For example, among 50–54 year old men the Māori:non-Māori mortality rate difference decreased from 122 to 115 per 100,000 with the Mandatory-All intervention—but the rate ratio increased slightly from 4.46 to 4.48. Similar patterns were evident for women and older age groups.

thumbnail
Table 6. Ethnic inequality impacts after 10 years from two sodium reduction interventions (CVD mortality rates, rate ratios and rate differences, and QALYs gained for individuals given model structure assumptions and parameter inputs for selected age-groups).

https://doi.org/10.1371/journal.pone.0123915.t006

Considering the QALYs gained, there appear to be larger absolute gains for Māori at younger ages—consistent with the larger absolute reduction in CVD rates. However, for older 75–79 year olds there was little difference in QALYs gains between Māori and non-Māori, due to the higher background (competing) mortality and morbidity among Māori, limiting potential health gains from a CVD-only intervention. In the additional “equity analysis” that we performed (i.e., to avoid “penalisation” of Māori we applied the lower non-Māori mortality and morbidity rates to Māori—see Methods), QALYs gained by Māori were substantially greater than for non-Māori.

Scenario analyses around the interventions

For some of the scenario analyses considered, the Counselling intervention was not always cost-effective (e.g., at the 6% discount rate) (Table D in S1 File and Table E in S1 File). This was also the case when we used the results of another meta-analysis for the relationship between sodium intake and blood pressure [11] (i.e., this pushed up the ICER to NZ$ 64,800 per QALY gained). Nevertheless, for all other seven interventions these remained cost-saving in the scenario analyses considered (including adjustments to the case-fatality rates for CVD). Further comments on these analyses are in the S1 file.

Uncertainty analyses

Tornado plots show how input parameters had an impact on the model’s incremental costs, QALYs gained, and ICERs for the Counselling and Mandatory-All interventions (Fig A in S1 File). For both interventions, uncertainty for QALYs and costs was particularly driven by the uncertainty in the level of sodium reduction from the intervention (i.e., those parameters described in Tables 2 and 3), followed by the uncertainty in the relative risk for the BP-stroke association. The impact of uncertainty in the relative risks associating stroke with BP was much greater than the parallel impact for CHD due to health gains being mediated more by stroke than CHD.

Discussion

Main findings and interpretation

This study adds to the existing literature to provide additional modelling-level evidence that legislation-based interventions for reducing sodium in the food supply would provide large health gains and also large cost-savings for a health system. The relatively greater health benefit from mandatory (vs voluntary) interventions is consistent with previous modelling work (e.g., [39,40]) and is not surprising given the strong scientific basis for the effectiveness of public health laws in general [41,42], and simply because interventions that change the food environment tend to have large reach and don’t involve individual-level behaviour changes and the costs of health professional time (e.g., as required for counselling interventions).

The net cost savings achieved for most interventions are consistent with much of the previous modelling literature around salt interventions and can be attributed to the relatively low cost of passing legislation (especially in the New Zealand setting), the cost savings from preventing CVD disease in the short and medium-term, and the fact that discounting partly erodes the impact of the more temporally distant extra health costs from increased lifespan.

Given the quality of New Zealand data by ethnicity, we were able to examine likely ethnic inequality impacts of these interventions. Assuming that the effect sizes (i.e., association of BP with stroke/CHD, association of changing salt intake with BP, and association of intervention with salt reduction) are similar across ethnic groups, but allowing for the higher age-specific CVD incidence rates among Māori, our modelling results suggest that CVD mortality rate differences will decrease in the future with a mandatory salt reduction strategy and that QALY gains will be greater for Māori. That is, a mandatory salt reduction intervention appears to be an inequality reducing intervention in absolute terms, consistent with research elsewhere on population-wide interventions on CVD risk factors [43,44]. However, the future CVD mortality rate ratios may increase; it is not uncommon for absolute difference to decrease, but relative differences to increase over time, with respect to inequalities [45]. This inequality reducing benefit of sodium reduction interventions appears to have only been detailed once before in modelling work—for African American men and women in the US [17]. Furthermore, our modelling of the inequality reduction effects might actually underestimate the benefit for Māori since we did not consider the slightly higher baseline BP for Māori (albeit only 3 mm Hg systolic BP [46]).

The results in this study probably have a reasonable level of applicability to other countries—given that high sodium intakes are a risk to health in virtually every country and most governments have the potential powers to legislate around sodium in the food supply. But for some of the interventions there will of course be specific considerations around feasibility and effectiveness (see S1 file). Furthermore, there can be tremendous variations in the net health system cost depending on how diseases are costed, model structure assumptions, and how the timing of costs is dealt with (see van Baal et al for an analysis of many of these aspects [47]). For example, the full inclusion of residential care costs as per analyses for the Netherlands may contribute to making some preventive interventions relating to tobacco control [48], and obesity control [49], less likely to be net cost saving.

Study strengths and limitations

This study included various improvements compared to previous disease and health economic models around sodium reduction (particularly in terms of cost data, but also in terms of considering such issues as ethnicity—see Introduction). Some of the interventions had not been subjected to health economic modelling before (e.g., the Sinking Lid) or else only modelled rarely (e.g., a media campaign to lower sodium, the Endorsement Label Programme [22] and the Salt Tax [21]). Although NICE reported that the UK interventions around sodium would be cost-saving [50], no economic modelling details have been published on the UK Package.

Nevertheless, as per other such modelling work there are many limitations. These are expanded on in the S1 file, but to summarise they include limitations around: (i) model structure and indeed our uncertainty estimates do not capture uncertainty arising from “model structure uncertainty”; (ii) limitations around input parameters (e.g., particularly relating to limitations with current HealthTracker costs and some epidemiological data (e.g. prevalence of CVD), and for the more theoretical interventions such as the Sinking Lid); (iii) unknowns in public and industry responses (e.g., compensatory behaviours in response to perceived reduced saltiness of processed foods); and (iv) just taking a health system perspective (e.g., ignoring the economic benefits of preventing premature deaths in workers).

Potential research and policy implications

Given the limitations with such modelling work as this, additional research is clearly desirable, particularly around an expanded set of plausible interventions and for using additional real-world data on the impact of down-regulating sodium in processed foods (e.g., as per recent laws in South Africa [51] and various European countries [52]). Other potential interventions that could be modelled further (for health gain and cost-effectiveness) include more general “junk food taxes”, and/or subsidising fruit and vegetables.

Nevertheless, waiting for such additional research is not critical if policy-makers are seeking to address the NCD epidemic in their countries and to achieve large financial savings. An optimal strategy might be to introduce a salt tax and then to use the tax revenue gained for additional health-promoting interventions (e.g., subsidising fruit and vegetables or providing healthier school lunches for children). But some policy-makers might be more interested in using the Sinking Lid approach, and this could be argued for on the grounds of the growing international experience with administering cap-and-trade systems for greenhouse gases and other air pollutants (such as sulphur and nitrogen oxides in the USA).

Other policy-makers may wish to start by regulating just the top sources of sodium by food category (as per the Mandatory-3G intervention). This would probably be easier to implement and evaluate than the more comprehensive, but possibly fairer, Mandatory-All approach.

Conclusions

In modelling work that had a range of improvements on previous models (particularly in terms of cost data) it was found that the use of mandatory controls on sodium in the food supply delivered both major health gains and major cost savings. Absolute health gain per person was greater for Māori men and women compared to non-Māori. Therefore such interventions could also reduce ethnic inequalities in health.

Supporting Information

S1 File. Supporting Information—Main file.

https://doi.org/10.1371/journal.pone.0123915.s001

(DOCX)

S2 File. Supporting Information—Costs for the salt model.

https://doi.org/10.1371/journal.pone.0123915.s002

(DOCX)

Acknowledgments

We thank June Atkinson and Giorgi Kvizhinadze for work on the disease costs, and Cristina Cleghorn for comments on a draft manuscript.

Author Contributions

Conceived and designed the experiments: NW TB LC. Performed the experiments: NN. Analyzed the data: NN. Contributed reagents/materials/analysis tools: NN NW LC AP. Wrote the paper: NW TB NN LC AP.

References

  1. 1. Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H, et al. (2012) A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380: 2224–2260. pmid:23245609
  2. 2. Beaglehole R, Bonita R, Horton R, Adams C, Alleyne G, Asaria P, et al. (2011) Priority actions for the non-communicable disease crisis. Lancet 377: 1438–1447. pmid:21474174
  3. 3. Di Cesare M, Khang YH, Asaria P, Blakely T, Cowan MJ, Farzadfar F, et al. (2013) Inequalities in non-communicable diseases and effective responses. Lancet 381: 585–597. pmid:23410608
  4. 4. WHO (2012) Guideline: Sodium intake for adults and children. Geneva, World Health Organization (WHO). http://www.who.int/nutrition/publications/guidelines/sodium_intake_printversion.pdf
  5. 5. Bayer R, Johns DM, Galea S (2012) Salt and public health: contested science and the challenge of evidence-based decision making. Health Aff (Millwood) 31: 2738–2746.
  6. 6. Institute of Medicine (2013) Sodium Intake in Populations: Assessment of Evidence. Washington, DC: The National Academies Press, http://www.nap.edu/catalog.php?record_id=18311. pmid:24851297
  7. 7. Neal B, Land MA, Woodward M (2013) An update on the salt wars-genuine controversy, poor science, or vested interest? Curr Hypertens Rep 15: 687–693. pmid:24170199
  8. 8. O'Donnell M, Mente A, Rangarajan S, McQueen MJ, Wang X, Liu L, et al. (2014) Urinary sodium and potassium excretion, mortality, and cardiovascular events. N Engl J Med 371: 612–623. pmid:25119607
  9. 9. Cook NR (2014) Sodium and cardiovascular disease. N Engl J Med 371: 2134. pmid:25427119
  10. 10. Batuman V (2014) Sodium and cardiovascular disease. N Engl J Med 371: 2134–2135. pmid:25427119
  11. 11. He FJ, Li J, Macgregor GA (2013) Effect of longer-term modest salt reduction on blood pressure. Cochrane Database Syst Rev 4: CD004937.
  12. 12. Aburto NJ, Ziolkovska A, Hooper L, Elliott P, Cappuccio FP, Meerpohl JJ. (2013) Effect of lower sodium intake on health: systematic review and meta-analyses. BMJ 346: f1326. pmid:23558163
  13. 13. Cook NR, Cutler JA, Obarzanek E, Buring JE, Rexrode KM, Kumanyika SK, et al. (2007) Long term effects of dietary sodium reduction on cardiovascular disease outcomes: observational follow-up of the trials of hypertension prevention (TOHP). BMJ 334: 885–888. pmid:17449506
  14. 14. Cook NR, Appel LJ, Whelton PK (2014) Lower levels of sodium intake and reduced cardiovascular risk. Circulation 129: 981–989. pmid:24415713
  15. 15. Chang HY, Hu YW, Yue CS, Wen YW, Yeh WT, Hsu LS, et al. (2006) Effect of potassium-enriched salt on cardiovascular mortality and medical expenses of elderly men. Am J Clin Nutr 83: 1289–1296. pmid:16762939
  16. 16. Merino J, Guasch-Ferré M, Martínez-González M, Corella D, Estruch R, Fitó M, et al. (2015) Is complying with the recommendations of sodium intake beneficial for health in individuals at high cardiovascular risk? Findings from the PREDIMED study. Am J Clin Nutr 101: 440–448. pmid:25733627
  17. 17. Bibbins-Domingo K, Chertow GM, Coxson PG, Moran A, Lightwood JM, Pletcher MJ, et al. (2010) Projected effect of dietary salt reductions on future cardiovascular disease. N Engl J Med 362: 590–599. pmid:20089957
  18. 18. Lewis KH, Rosenthal MB (2011) Individual responsibility or a policy solution—cap and trade for the U.S. diet? N Engl J Med 365: 1561–1563. pmid:22029976
  19. 19. Forshee RA (2008) Innovative regulatory approaches to reduce sodium consumption: could a cap-and-trade system work? Nutr Rev 66: 280–285. pmid:18454814
  20. 20. Basu S, Lewis K (2014) Reducing added sugars in the food supply through a cap-and-trade approach. Am J Public Health 104: 2432–2438. pmid:25365146
  21. 21. Smith-Spangler CM, Juusola JL, Enns EA, Owens DK, Garber AM (2010) Population strategies to decrease sodium intake and the burden of cardiovascular disease: a cost-effectiveness analysis. Ann Intern Med 152: 481–487, W170-483. pmid:20194225
  22. 22. Cobiac LJ, Vos T, Veerman JL (2010) Cost-effectiveness of interventions to reduce dietary salt intake. Heart 96: 1920–1925. pmid:21041840
  23. 23. Cobiac LJ, Magnus A, Lim S, Barendregt JJ, Carter R, Vos T. (2012) Which interventions offer best value for money in primary prevention of cardiovascular disease? PLoS One 7: e41842. pmid:22844529
  24. 24. Nghiem N, Wilson N, Blakely T (2014) Technical Background to the Cardiovascular Disease Model used in the BODE³ Programme. Wellington: Department of Public Health, University of Otago. http://www.otago.ac.nz/wellington/otago070188.pdf.
  25. 25. Ministry of Health (2013) Ways and Means: A report on methodology from the New Zealand Burden of Disease, Injury and Risk Study, 2006–2016. Wellington: Ministry of Health. http://www.health.govt.nz/publication/ways-and-means-report-methodology-new-zealand-burden-disease-injury-and-risk-study-2006-2016.
  26. 26. Blakely T, Foster R, Wilson N, BODE³ Team (2012) Burden of Disease Epidemiology, Equity and Cost-Effectiveness (BODE3) Study Protocol. Version 2.1. Technical Report No.3. Wellington: Department of Public Health, University of Otago, Wellington, December 2012. http://www.otago.ac.nz/wellington/otago042986.pdf.
  27. 27. OECD (2013) New international comparisons of GDP and consumption based on purchasing power parities for the year 2011. Paris.
  28. 28. Baltussen R, Adam T, Tan-Torres Edejer T, Hutubessy R, Acharya Aea (2003) Methods for generalized cost-effectiveness analysis. In: Tan-Torres Edejer T, Baltussen R, Adam T, Hutubessy R, Acharya A, Evans DB, et al., editors. Making choices in health: WHO guide to cost-effectiveness analysis. Geneva: World Health Organization.
  29. 29. Blakely T, Atkinson J, Kvizhinadze G, Nghiem N, McLeod H, Wilson N. (2014) Health system costs by sex, age and proximity to death, and implications for estimation of future expenditure. N Z Med J 127: 1–14. pmid:24851314
  30. 30. Nghiem N, Wilson N, Blakely T (2014) Validation Issues Relating to the Cardiovascular Disease Model Developed in the BODE³ Programme. Wellington: Department of Public Health, University of Otago. http://www.otago.ac.nz/wellington/otago070189.pdf.
  31. 31. Barendregt J, Oortmarssen GJ, Vos T, Murray CJL (2003) A generic model for the assessment of disease epidemiology: the computational basis of DisMod II. Popul Health Metr 1: 4. pmid:12773212
  32. 32. Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A, et al. (2012) Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010. The Lancet 380: 2129–2143. pmid:23245605
  33. 33. Law MR, Frost CD, Wald NJ (1991) By how much does dietary salt reduction lower blood-pressure? 1. Analysis of observational data among populations. British Medical Journal 302: 811–815. pmid:2025703
  34. 34. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R (2002) Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 360: 1903–1913. pmid:12493255
  35. 35. Law MR, Morris JK, Wald NJ (2009) Use of blood pressure lowering drugs in the prevention of cardiovascular disease: meta-analysis of 147 randomised trials in the context of expectations from prospective epidemiological studies. British Medical Journal 338: b1665. pmid:19454737
  36. 36. OECD OECD Health Statistics 2014: How does New Zealand compare? Paris: OECD, 2014. http://www.oecd.org/els/health-systems/Briefing-Note-NEW-ZEALAND-2014.pdf.
  37. 37. McLeod M, Blakely T, Kvizhinadze G, Harris R (2014) Why equal treatment is not always equitable: the impact of existing ethnic health inequalities in cost-effectiveness modeling. Popul Health Metr 12: 15. pmid:24910540
  38. 38. WHO (2012) Choosing interventions that are cost effective (WHOCHOICE). Geneva: World Health Organization.
  39. 39. Cobiac LJ, Vos T, Veerman JL (2010) Cost-effectiveness of interventions to reduce dietary salt intake. Heart 96: 1920–1925. pmid:21041840
  40. 40. Collins M, Mason H, O'Flaherty M, Guzman-Castillo M, Critchley J, Capewell S. (2014) An economic evaluation of salt reduction policies to reduce coronary heart disease in England: a policy modeling study. Value Health 17: 517–524. pmid:25128044
  41. 41. Moulton AD, Mercer SL, Popovic T, Briss PA, Goodman RA, Thombley ML, et al. (2009) The scientific basis for law as a public health tool. Am J Public Health 99: 17–24. pmid:19008510
  42. 42. Goodman RA, Moulton A, Matthews G, Shaw F, Kocher P, Mensah G, et al. (2006) Law and public health at CDC. MMWR Morb Mortal Wkly Rep 55 Suppl 2: 29–33. pmid:17183242
  43. 43. Kivimäki M, Shipley MJ, Ferrie JE, Singh-Manoux A, Batty GD, Chandola T, et al. (2008) Best-practice interventions to reduce socioeconomic inequalities of coronary heart disease mortality in UK: a prospective occupational cohort study. The Lancet 372: 1648–1654. pmid:18994662
  44. 44. Capewell S, Graham H (2010) Will Cardiovascular Disease Prevention Widen Health Inequalities? PLoS Med 7: e1000320. pmid:20811492
  45. 45. Blakely T, Tobias M, Atkinson J (2008) Inequalities in mortality during and after restructuring of the New Zealand economy: repeated cohort studies. BMJ 336: 371–375. pmid:18218998
  46. 46. McLean RM, Williams S, Mann JI, Miller JC, Parnell WR (2013) Blood pressure and hypertension in New Zealand: results from the 2008/09 Adult Nutrition Survey. N Z Med J 126: 1–14. pmid:24308076
  47. 47. van Baal PH, Feenstra TL, Polder JJ, Hoogenveen RT, Brouwer WB (2011) Economic evaluation and the postponement of health care costs. Health Econ 20: 432–445. pmid:21210494
  48. 48. van Baal PH, Brouwer WB, Hoogenveen RT, Feenstra TL (2007) Increasing tobacco taxes: a cheap tool to increase public health. Health Policy 82: 142–152. pmid:17050031
  49. 49. van Baal PH, Polder JJ, de Wit GA, Hoogenveen RT, Feenstra TL, Boshuizen HC, et al. (2008) Lifetime medical costs of obesity: prevention no cure for increasing health expenditure. PLoS Med 5: e29. pmid:18254654
  50. 50. NICE (National Institute for Health and Care Excellence) (2010) Prevention of cardiovascular disease. NICE public health guidance 25. London: NICE.
  51. 51. Hofman K, Tollman S (2013) Population health in South Africa: a view from the salt mines. Lancet Global Health 1: e66–67. pmid:25104152
  52. 52. European Commission Survey on members states implementation of the EU salt reduction framework: Directorate-General Health and Consumers 2012. http://ec.europa.eu/health/nutrition_physical_activity/docs/salt_report1_en.pdf
  53. 53. McLean R, Williams S, Mann J, Parnell W (2011) How much salt are we eating? Estimates of New Zealand population sodium from the 2008/2009 Adult Nutrition Survey [Presentation on 2 December 2011]. Joint Annual Scientific Meeting of the Australian and New Zealand Nutrition Societies. Queenstown, New Zealand (29 November—2 December).
  54. 54. McLean RM, Mann JI, Hoek J (2011) World Salt Awareness Week: more action needed in New Zealand. N Z Med J 124: 68–76. pmid:21946964
  55. 55. Wilson N (2014) Technical Report: Estimating the modelling parameters around dietary counselling for preventing cardiovascular disease in New Zealand. Wellington: University of Otago, Wellington. http://www.otago.ac.nz/wellington/otago071960.pdf.
  56. 56. Rees K, Dyakova M, Ward K, Thorogood M, Brunner E (2013) Dietary advice for reducing cardiovascular risk. Cochrane Database Syst Rev 3: CD002128.
  57. 57. Wilson N, Nghiem N (2014) Background Report for BODE3 Modelling on Estimating the Impact of the Tick Programme in New Zealand (a Heart Health Food Endorsement Programme). Wellington: University of Otago, Wellington. http://www.otago.ac.nz/wellington/otago071961.pdf.
  58. 58. Wilson N, Nghiem N, Eyles H, Ni Mhurchu C, Cobiac LJ, Pearson AL, et al. (2014) Possible impact of the Tick Programme in New Zealand on selected nutrient intakes: Tentative estimates and methodological complexities. N Z Med J 127(1399): 85–88. pmid:25145310
  59. 59. He FJ, Brinsden HC, Macgregor GA (2013) Salt reduction in the United Kingdom: a successful experiment in public health. J Hum Hypertens [E-publication 31 October].
  60. 60. Pietinen P, Mannisto S, Valsta LM, Sarlio-Lahteenkorva S (2010) Nutrition policy in Finland. Public Health Nutr 13: 901–906. pmid:20513258
  61. 61. NHMRC/MoH (2006) Nutrient Reference Values for Australia and New Zealand. Canberra, ACT: National Health and Medical Research Council (NHMRC); New Zealand Ministry of Health (MoH). http://www.nhmrc.gov.au; http://www.moh.govt.nz/publications
  62. 62. Landsburg S (2010) Price Theory and Applications (with Economic Applications): Joe Sabatino.
  63. 63. Wilson N, Nghiem N, Foster R, Cobiac L, Blakely T (2012) Estimating the cost of new public health legislation. Bull World Health Organ 90: 532–539. pmid:22807599
  64. 64. BODE3 Programme (2012) Results for the cost of making a new law, all in $NZ. http://www.otago.ac.nz/wellington/otago033080.pdf.
  65. 65. Asaria P, Chisholm D, Mathers C, Ezzati M, Beaglehole R (2007) Chronic disease prevention: health effects and financial costs of strategies to reduce salt intake and control tobacco use. Lancet 370: 2044–2053. pmid:18063027
  66. 66. Eatwell (2012) D2.3 Evaluation of cost utility of policy interventions. Reading: University of Reading.