CCBYNC Open access
Research

# Cost effectiveness of a government supported policy strategy to decrease sodium intake: global analysis across 183 nations

BMJ 2017; 356 (Published 10 January 2017) Cite this as: BMJ 2017;356:i6699
1. Michael Webb, doctoral candidate1,
2. Saman Fahimi2,
3. Gitanjali M Singh, assistant professor3,
4. Shahab Khatibzadeh, postdoctoral fellow2,
5. Renata Micha, research assistant professor3,
6. John Powles, honorary senior visiting fellow4,
7. Dariush Mozaffarian, dean3
1. 1Stanford University, Stanford, CA, USA, and Institute for Fiscal Studies, London, UK
2. 2Harvard TH Chan School of Public Health, Boston, MA, USA
3. 3Tufts Friedman School of Nutrition Science & Policy, 150 Harrison Ave, Boston, MA 02111, USA
4. 4Cambridge Institute of Public Health, Cambridge, UK
1. Correspondence to: D Mozaffarian dariush.mozaffarian{at}tufts.edu
• Accepted 1 December 2016

## Abstract

Objective To quantify the cost effectiveness of a government policy combining targeted industry agreements and public education to reduce sodium intake in 183 countries worldwide.

Design Global modeling study.

Setting 183 countries.

Population Full adult population in each country.

Intervention A “soft regulation” national policy that combines targeted industry agreements, government monitoring, and public education to reduce population sodium intake, modeled on the recent successful UK program. To account for heterogeneity in efficacy across countries, a range of scenarios were evaluated, including 10%, 30%, 0.5 g/day, and 1.5 g/day sodium reductions achieved over 10 years. We characterized global sodium intakes, blood pressure levels, effects of sodium on blood pressure and of blood pressure on cardiovascular disease, and cardiovascular disease rates in 2010, each by age and sex, in 183 countries. Country specific costs of a sodium reduction policy were estimated using the World Health Organization Noncommunicable Disease Costing Tool. Country specific impacts on mortality and disability adjusted life years (DALYs) were modeled using comparative risk assessment. We only evaluated program costs, without incorporating potential healthcare savings from prevented events, to provide conservative estimates of cost effectiveness

Main outcome measure Cost effectiveness ratio, evaluated as purchasing power parity adjusted international dollars (equivalent to the country specific purchasing power of US$) per DALY saved over 10 years. Results Worldwide, a 10% reduction in sodium consumption over 10 years within each country was projected to avert approximately 5.8 million DALYs/year related to cardiovascular diseases, at a population weighted mean cost of I$1.13 per capita over the 10 year intervention. The population weighted mean cost effectiveness ratio was approximately I$204/DALY. Across nine world regions, estimated cost effectiveness of sodium reduction was best in South Asia (I$116/DALY); across the world’s 30 most populous countries, best in Uzbekistan (I$26.08/DALY) and Myanmar (I$33.30/DALY). Cost effectiveness was lowest in Australia/New Zealand (I$880/DALY, or 0.02×gross domestic product (GDP) per capita), although still substantially better than standard thresholds for cost effective (<3.0×GDP per capita) or highly cost effective (<1.0×GDP per capita) interventions. Most (96.0%) of the world’s adult population lived in countries in which this intervention had a cost effectiveness ratio <0.1×GDP per capita, and 99.6% in countries with a cost effectiveness ratio <1.0×GDP per capita. Conclusion A government “soft regulation” strategy combining targeted industry agreements and public education to reduce dietary sodium is projected to be highly cost effective worldwide, even without accounting for potential healthcare savings. ## Introduction Excessive sodium consumption is common and linked to cardiovascular burdens in most countries. Overall, 181 of 187 countries, representing 99.2% of the global adult population, have mean sodium intakes exceeding the World Health Organization recommended maximum of 2 g/day.1 Based on this threshold, an estimated 1 648 000 annual deaths from cardiovascular diseases worldwide were attributable to excess dietary sodium in 2010.2 Accordingly, the 2013 United Nations’ Global Action Plan for the Prevention and Control of Noncommunicable Diseases has prioritized sodium reduction as one of nine key targets for all member nations in 2013-20.3 A potential barrier for implementation of this recommendation is cost. Many countries have limited resources for health interventions, requiring careful assessment of costs and cost effectiveness. Several countries now have national programs that include a specific aim of reducing population sodium intake; for instance, as of 2012, 29 European nations, consisting of all EU Member States as well as Norway and Switzerland, had salt reduction initiatives in place.4 Yet the cost effectiveness of such efforts globally is uncertain. While prior studies have estimated sodium reduction policies to be highly cost effective, or even cost saving, in specific countries, the potential cost effectiveness of such strategies has been analyzed for only a handful of nations and regions, mostly focused on high income nations, and in ways that are not generally comparable.5 6 7 8 9 10 11 12 13 14 To address this key gap in knowledge, we assessed the cost effectiveness of sodium reduction strategies in 183 nations, based on the most up to date available data on age specific and sex specific sodium intakes, blood pressure levels, and cardiovascular disease burdens worldwide, the dose-response effects of sodium on blood pressure and of blood pressure on cardiovascular disease, and nation specific costs for each component of the intervention. Together, these allowed us to model and estimate, using comparable and consistent methods, the cost effectiveness of sodium reduction strategies for every country. ## Methods ### Sodium reduction intervention We modeled the effects and costs of a 10 year government “soft regulation” policy to reduce population sodium consumption (see supplementary eTable 1 for details of the model assumptions). The intervention program was modeled on recent experience in the UK15 and included government supported industry agreements to reduce sodium in processed foods, government monitoring of industry compliance, and a public health campaign targeting consumer choices. In the UK, for example, this intervention was based on collaboration between national government offices focused on nutrition (Food Standards Agency) and health (ministers of public health), together with non-governmental advocacy organizations (Consensus Action on Salt & Health). The program applied sustained pressure on food manufacturers to pursue progressive reformulation, reinforced by food group specific targets, independent monitoring, and a sustained media campaign against excess salt intake. The program we modeled was thus more robust and costly than simple “voluntary reformulation.” We assumed the intervention would scale up linearly over 10 years, with one 10th of the total sodium reduction in the first year, two 10ths in the second, and so on, reaching full efficacy in the final year. We recognized that alternative programs, such as mandatory regulation, would likely have larger effects, reduce sodium consumption more quickly, and at lower cost, but may be less politically feasible in many countries. ### Intervention costs Country specific resource needs and costs were derived using the WHO-CHOICE database,16 which includes detailed component specific estimates of inputs (ingredients) required for each intervention stage for each country’s government and the estimated unit price for each input in that country including for example costs of human resources, training, meetings, supplies, equipment, and mass media (see supplementary eMethods). To facilitate comparisons between countries, we converted all costs to international dollars (I$) (see supplementary eMethods), which account for each nation’s currency as well as purchasing power parity.17 One I$in any given country can be interpreted as the funds needed to purchase the same amounts of goods or services in that country as one US$ would purchase in the US. For countries with lower incomes than the US, conversion of our findings from I$to US$ would substantially increase the apparent cost effectiveness (ie, the cost in US$per disability adjusted life year (DALY) saved would be much lower). We summed costs by year to calculate the total cost of the 10 year intervention for each country, with 3% annual discounting. We modeled only governmental intervention costs, for several reasons. First, this cost is most relevant to budget constrained governments, since the program cost must be borne directly and immediately. Second, net industry sector costs for product reformulation in each country would be difficult to determine because once the relevant reformulation has been undertaken in any single country, the knowledge of that reformulation can be extended with much less additional cost to other countries. For example, multinational companies transfer improved recipes and reformulation strategies across borders with no cost, as do food scientists moving between firms, and so on. Third, in contrast to recent US models,10 11 we did not include estimated healthcare savings or increased productivity from prevented cardiovascular disease events because such savings could, in theory, be partly offset by new downstream health events resulting from enhanced survival18 19 and because comparable healthcare and productivity costing data are available for a minority of countries globally. Because including such cost savings would be optimal according to many cost effectiveness guidelines, our results for overall cost effectiveness should be considered a conservative estimate. ### Heterogeneity in intervention costs and effectiveness Though the WHO costing framework already accounted for some sources of variation by country in terms of resources required and nation specific costs, we recognized that details of planning, development, and implementation might further vary from country to country beyond what is captured by the costing tool. We also recognized that achieved effectiveness would vary from country to country. Our base model assumed an average cost of this framework (already adjusted for in-country differences in resource use and costs, according to the WHO costing tool), and an average effectiveness. To understand the robustness of our findings to these assumptions, we tested widely varying costs—including variations in resource use and cost of between 0.25-fold and fivefold the base—and varying intervention effectiveness, including 10% and 30% proportional reductions and 0.5 g/day and 1.5 g/day absolute reductions in sodium intake over 10 years. Plausible intervention effectiveness was informed by experiences in the UK, which achieved a 14.7% (0.6 g/day) reduction in population sodium intake over 10 years,20 and Turkey, which reported a more rapid 16% (1.2 g/day) reduction over four years.21 Together, these findings provided a broad range of possible scenarios against which to evaluate the cost effectiveness of the intervention. ### Intervention impact on DALYs Using data on population demographics, sodium consumption, blood pressure levels, and rates of cardiovascular disease, each in 26 strata by age and sex within each country,2 we estimated the number of disability adjusted life years that would be averted by the intervention in each country for each year between 2011 and 2020. Risk reduction in each age-sex-country stratum was calculated from the effect of sodium reduction on systolic blood pressure, including variation in this effect by age, race, and hypertensive status; and the effect of blood pressure reduction on cardiovascular disease, including variation in this effect by age.2 The final comparative risk assessment model incorporated each of these sources of heterogeneity, as well as their uncertainty. Stratum specific effects, accounting for underlying demographics and baseline cardiovascular disease rates, were summed to derive national (or regional) effects (see supplementary eMethods for details on these inputs and their modeling). While some prior observational studies suggest a J-shaped relation between sodium intake and cardiovascular disease,22 this could be explained by potential biases of sodium assessment in observational studies (see supplementary eMethods).23 In extended follow-up of sodium reduction trials that overcame many of these limitations, linear risk reductions were seen, including lower risk with intakes less than 2.3 g/day.24 We recognized that while the precise optimal level of sodium intake remains controversial, every major national and international organization that has reviewed all the evidence has concluded that high sodium intake increases cardiovascular disease risk and that lowering sodium intake reduces such risk, with optimal identified intakes ranging from less than 1.2 g/day to less than 2.4 g/day.2 We used an optimal intake of 2.0 g/day (WHO) for our main analysis. For any sodium reductions below this level, we modeled neither additional benefit nor risk, consistent with recent Institute of Medicine conclusions.25 In sensitivity analyses, we also evaluated lower (1.0 g/day) and higher (3.0 g/day) thresholds for optimal intake. Our modeling further utilized known strengths of blood pressure as “an exemplar surrogate endpoint for cardiovascular mortality and morbidity.”26 Prospective cohort studies suggest log-linear associations between systolic blood pressure and cardiovascular disease events, down to around 110 mm Hg27; and randomized controlled trials indicate that benefits of blood pressure lowering interventions are largely proportional to the magnitude of blood pressure reduction, rather than the specific intervention, with similar proportional reductions in cardiovascular disease events down to pretreatment blood pressures of around 110 mm Hg.27 28 29 In our model, we assumed a log-linear dose-response between blood pressure and cardiovascular disease until a systolic blood pressure level of 115 mm Hg, below which we assumed no further lowering of risk. Given the relatively rapid reductions in cardiovascular disease events in randomized trials of blood pressure lowering therapies, and the prolonged period of our intervention (10 years), we did not model any lag and assumed concurrent gradual benefits in both blood pressure reduction and cardiovascular disease. ### Cost effectiveness ratios To calculate the cost effectiveness ratio for each country, we divided the total effect on DALYs by the total cost of the intervention over 10 years. We compared these cost effectiveness ratios to WHO benchmarks, which define a cost effectiveness ratio <3×gross domestic product (GDP) per capita as cost effective, and <1×GDP per capita as highly cost effective.30 We appreciated the potential limitations of these WHO benchmarks31 yet also their practicality for multinational studies such as this. To quantify statistical uncertainty, we used probabilistic sensitivity analyses based on 1000 Monte Carlo simulations to derive 95% uncertainty intervals, with varying inputs for sodium use, blood pressure levels, effects of sodium on blood pressure, and effects of blood pressure on cardiovascular disease (see supplementary eMethods). ### Patient involvement No patients were involved in setting the research question or the outcome measures, nor were they involved in developing plans for design or implementation of the study. No patients were asked to advise on interpretation or writing up of results. There are no plans to disseminate the results of the research to study participants or the relevant patient community. ## Results ### Cost effectiveness of sodium reduction by national income level and region Taking into account local prices, currencies, and purchasing power, the relative contributions of each intervention component to the total 10 year cost differed appreciably between countries (see supplementary eFigure 1). For instance, costs of supplies and equipment, meetings, and training were uniformly low (averaging I$0.01 per capita, I$0.01 per capita, and I$0.04 per capita, respectively), whereas costs of human resources and mass media were much higher and more variable across countries. Globally, average purchasing power parity adjusted costs for human resources (personnel salaries) were I$0.27 per capita, but with a ninefold range comparing high income (I$0.93) with low income (I$0.10) countries. Human resources were most costly in Australia/New Zealand (I$1.26 per capita), Western Europe (I$1.03), and Canada/US (I$0.82); and lowest in South Asia (I$0.06). Mass media costs were generally the most expensive component of the intervention: I$0.80 per capita globally, I$1.07 for high income nations, and I$0.44 for low income nations. They represented the most costly component of the intervention in every region except for Australia/New Zealand, Canada/US, and Western Europe, where human resources was the most costly component.

Globally, the estimated average cost effectiveness ratio of the 10 year intervention was approximately I$204 per DALY saved (95% uncertainty interval I$149 to I$322) (table 1). This did not include potential savings from lower healthcare costs or higher productivity owing to averted cardiovascular disease events, which would each further improve the estimated cost effectiveness. The estimated cost effectiveness ratio was lowest (best) in lower middle income (I$111, I$81 to I$175) and upper middle income countries (I$146, I$109 to I$223), higher in low income countries (I$215, I$139 to I$400), and highest in high income countries (I$465, I$341 to I$724). By region, the lowest cost effectiveness ratios were in South Asia and East/Southeast Asia (I$116 and I$123, respectively). In Central Asia/Eastern and Central Europe, high intervention efficacy partly offset its higher projected cost, generating the next best cost effectiveness ratio (I$211, I$157 to I$324).

Table 1

Cost effectiveness by income and geographic region of a national government supported policy intervention to reduce sodium consumption by 10% over 10 years*

View this table:

### Effectiveness, cost, and cost effectiveness by country

Across individual countries, the estimated intervention efficacy, in terms of DALYs averted per 1000 people, was highest in Kazakhstan (23.0, 95% uncertainty interval 15.6 to 29.8), Georgia (21.6, 14.3 to 28.3), Belarus (19.8, 12.8 to 26.9), Ukraine (19.0, 12.3 to 25.9), Mongolia (18.9, 12.1 to 25.0), and Russia (18.8, 12.2 to 25.5) (see supplementary eTable 3). The relative rankings of these nations should be considered in the context of the uncertainty in the estimates that preclude, for example, confirming statistically significant differences in efficacy between Kazakhstan and Russia. Yet, the range of estimated efficacy across the 183 nations was large—for example, compared with the countries above, much lower in Jamaica (1.9, 1.1 to 2.7), Qatar (1.4, 0.8 to 1.9), Rwanda (1.3, 0.6 to 2.3), and Kenya (0.4, 0.2 to 0.7).

Per capita, estimated 10 year intervention cost was lowest in Myanmar, Vietnam, Democratic People’s Republic of Korea (each I$0.31), Thailand (I$0.33), Nepal (I$0.40), and Uzbekistan (I$0.41) (see supplementary eTable 3). A total of 68 countries had estimated 10 year intervention costs of less than I$1.00 per capita. For 84 countries, estimated costs were between I$1.00 and I$9.99, for 19 countries, between I$10 and I$29.99, and for 12, greater than I$30.

Estimated national cost effectiveness ratios were correspondingly variable (fig 1). Uzbekistan’s was lowest (best) at I$26.08/DALY (95% uncertainty interval 20.08 to 39.02), followed by Myanmar (I$33.30, 25.10 to 50.46). Twenty eight countries had estimated cost effectiveness ratios below I$100/DALY, and 112 more, below I$1000/DALY. Eleven nations, all small, had estimated cost effectiveness ratios between I$10 000 and I$30 000/DALY (see supplementary eTable 3).

Fig 1 Cost effectiveness (purchasing power adjusted I$/disability adjusted life year) by country of a national policy intervention to reduce sodium consumption by 10% ### WHO benchmarks for cost effectiveness In comparison with WHO benchmarks (cost effectiveness ratio <3×GDP per capita is cost effective, <1×GDP per capita, highly cost effective),30 the 10 year sodium reduction intervention was estimated to be highly cost effective globally. Across all 183 countries, the estimated cost effectiveness ratio of this policy intervention was >3×GDP per capita in only one nation (Marshall Islands: 4.7×GDP per capita), between 3×GDP per capita and 1×GDP per capita in six nations (Kenya, Tonga, Kiribati, Samoa, Micronesia, Comoros), and highly cost effective in all other nations (fig 2). Indeed, in 130 countries, representing more than 96% of the world’s population, the estimated cost effectiveness ratio was <0.1×GDP per capita, far below usual cost effectiveness thresholds. This included each of the world’s 20 most populous countries (fig 3). Fig 2 Cost effectiveness (purchasing power adjusted I$/disability adjusted life year (DALY) as a multiple of gross domestic product (GDP) per capita) by country of a national policy intervention to reduce sodium consumption by 10%

Fig 3 Affordability of a national policy intervention to reduce sodium consumption by 10% in the world’s 20 most populous countries. Each point represents the cost effectiveness of the intervention (I$/disability adjusted life year (DALY)) for a given country against that country’s gross domestic product (GDP) per capita (I$), adjusted for purchasing power. The lines represent 0.01×, 0.05×, and 0.1×GDP per capita, selected as reasonable fractions against which to compare our estimates of affordability. Notably, each of these thresholds is substantially lower than the World Health Organization benchmarks for an intervention being cost effective (<3.0×GDP per capita) or highly cost effective (<1.0×GDP per capita). For example, Nigeria and Bangladesh, being to the right of the blue line and to the left of the red dotted line, have a cost effectiveness ratio less than 0.1×GDP per capita but greater than 0.05×GDP per capita

### Potential heterogeneity of effectiveness and costs

A national policy intervention to reduce sodium intake remained highly cost effective globally and by world region when we considered alternative effectiveness (proportional reduction of 30%, absolute reduction of 0.5 g/day or 1.5 g/day); and alternative thresholds of optimal intake (the level at which further sodium reduction produces no further health benefits) of 3.0 g/day or 1.0 g/day (table 2). Generally, achieving larger sodium reduction targets (eg, 30%, 1.5 g/day) was more cost effective (see supplementary eFigure 2), but even modest achieved reductions (10% or 0.5 g/day over 10 years) were highly cost effective. Under any of these scenarios, the estimated cost effectiveness ratio was <0.05×GDP per capita in nearly every world region. In Sub-Saharan Africa, owing to generally low sodium intakes in that region, the estimated cost effectiveness ratio was <0.1×GDP per capita when the optimal intake threshold was 1.0 g/day or 2.0 g/day, but up to 6.0×GDP per capita when it was assumed to be 3.0 g/day.

Table 2

Variation in cost effectiveness depending on heterogeneity of both intervention efficacy and optimal level of sodium intake by income and geographic region.* Values are I$/disability adjusted life years (DALYs) unless stated otherwise View this table: As expected, cost effectiveness ratios were sensitive to variations in estimated intervention cost. We evaluated the proportion of the world’s adult population living in countries with a cost effectiveness ratio (I$/DALY) <0.05×GDP per capita and <0.5×GDP per capita, for varying intervention costs that were 25%, 50%, 150%, 200%, or 500% of baseline cost estimates (see supplementary eFigure 3). For a 10% reduction in sodium intake, under the base case scenario for cost estimates, 89% of the global adult population would live in countries with a cost effectiveness ratio <0.05×GDP per capita. This decreased to 23% of the global adult population if costs were fivefold higher, 68% if costs were twofold higher, and 85% if costs were 1.5-fold higher. In contrast, 96% of the global adult population would live in countries with a cost effectiveness ratio <0.05×GDP per capita if costs were half as large, and 99% if costs were one quarter as large. For a 30% reduction in sodium intake, the corresponding figures for a benchmark of <0.05×GDP per capita were 85%, 92%, 96%, 98%, 99.1%, and 99.3% of the global adult population based on intervention costs that were 500%, 200%, 150%, 50%, or 25% of the baseline cost estimates, respectively. We also made comparisons against a cost effectiveness ratio benchmark <0.5×GDP per capita, still substantially below the WHO criterion of 1×GDP per capita as highly cost effective. For a 10% reduction in sodium intake, even if the intervention costs were fivefold greater than the baseline estimate, 96% of the world’s population would live in countries with a cost effectiveness ratio <0.5×GDP per capita; and for a 30% reduction in sodium intake, 99% would.

## Discussion

We found that a government “soft regulation” policy intervention to reduce national sodium consumption by 10% over 10 years was projected to be highly cost effective in nearly every country in the world (<1×gross domestic product (GDP) per capita per disability life year (DALY) saved), and remarkably cost effective (<0.05×GDP per capita per DALY) in most countries. Hundreds of thousands of deaths, and millions of DALYs, were estimated to be potentially averted annually, at low cost.

## Footnotes

• We thank the World Health Organization for use of the Noncommunicable Disease Costing Tool, and the Institute for Health Metrics and Evaluation for use of their cardiovascular disease DALY estimates.

• Contributors: MW and DM conceptualised the study and wrote the final draft of the paper; DM also provided funding support and supervision. MW undertook the analysis and wrote the first draft of the paper. SF, GMS, SK, and RM prepared data and commented on the paper. JP commented on the paper. All authors approved the final version. MW acts as guarantor of the study.

• Funding: This research was supported by the National Heart, Lung, and Blood Institute (R01 HL115189; principal investigator DM) and the National Institute of Diabetes and Digestive and Kidney Diseases (T32 training grant in academic nutrition, DK007703; GMS), National Institutes of Health. The sponsors had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

• Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: financial support from the National Institutes of Health for the submitted work. DM reports ad hoc honorariums or consulting fees from Boston Heart Diagnostics, Haas Avocado Board, Astra Zeneca, GOED, DSM, and Life Sciences Research Organization, none of which were related to topics of dietary sodium. The other authors report no financial relationships with any organizations that might have an interest in the submitted work in the previous three years.

• Ethical approval: Not required.

• Data sharing: The global data on sodium intake may be requested from the authors for academic collaborations; see www.globaldietarydatabase.org/requesting-data.html. Global data on blood pressure is available for download at www1.imperial.ac.uk/publichealth/departments/ebs/projects/eresh/majidezzati/healthmetrics/metabolicriskfactors/metabolic_risk_factor_maps/ Global data on cardiovascular events is available for download from the Global Burden of Diseases Study at http://ghdx.healthdata.org/global-burden-disease-study-2013-gbd-2013-data-downloads.

• Transparency: The lead author (MW) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 3.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/3.0/.

View Abstract