Abstract
Background Poor diets contribute significantly and increasingly to the burden of chronic diseases in the United Kingdom, impacting both health and the economy. The introduction of fiscal measures that target unhealthy foods can provide a near-unique opportunity to achieve shifts towards healthier diets while raising ring-fenced revenue for spending on healthcare.
Objective To estimate the expected health and economic benefits from the reduction in consumption of salt and sugar that could be expected from the introduction of a proposed £3/kg tax on sugar and a £6/kg tax on salt.
Methods Life-table modelling was used to estimate the expected health and economic benefits from the reduction in consumption of salt and sugar for four scenarios, each reflecting different manufacturer and consumer responses the proposed tax. Relative risks for 24 different disease-risk pairs were applied, exploring direct and indirect pathways between salt and sugar consumption, and mortality and morbidity.
Results The results show that life expectancy in the UK could be increased by 1.7 and nearly 5 months, depending on the degree of industry and consumer response to the tax. The tax could also lead to almost 2 million fewer cases of preventable chronic diseases with additional economic benefits of approximately £27 to £78 billion from avoided ill-health over 25 years. The largest health benefits would accrue from reduced mortality and morbidity from cardiovascular diseases.
Conclusions Significant benefits to both population health and the economy could be expected from extending the current tax on sugar sweetened beverages to other sugary foods and from adding a tax on foods high in salt. The proposed dietary changes are likely to be insufficient to reach national public health targets for obesity and chronic disease prevention; hence, additional measures to reduce the burden of chronic disease in the UK will be equally critical to consider.
Introduction
Suboptimal diet, characterised by factors such as low intakes of fruits and vegetables and high intakes of salt and added sugars, contributes to an estimated 90,000 deaths every year in the UK [1]. More than half of over-45s are living with health conditions for which diet is a widely recognised contributing risk factor, including cardiovascular disease, type-2 diabetes, and obesity. As things stand, obesity is expected to continue increasing [2]. By 2035, type 2 diabetes is projected to cost the UK National Health Services (NHS) £15 billion a year, equivalent to one and a half times as much as cancer does today [3]. Halting this trajectory will be crucial for protecting the future of the UK’s health service.
Limiting intakes of salt and added sugar is one important action to prevent diet-related disease [1]. Current intakes of added sugar in the UK are, on average, 55g/day (9.9 E%) in men and 44g (9.9E%) in women [4]. This represents nearly double the recommendation of no more than 30g/5E% from added sugars [5]. At the same time current average intakes of salt are 9.2g/ day for men and 7.6g/day for women [4], exceeding the daily recommendation of no more than 6g/day in adults [5].
Significant behaviour change would be needed to meet the recommendations for salt and added sugars in the UK. The limited success of the “5-a-day” campaign since 2003 suggests that information measures alone are not enough to change behaviour [6]. However, evidence from the implementation of the UK Soft Drinks Industry Levy (SDIL) [7] suggests that targeted fiscal measures could be effective in steering the consumption of specific foods among consumers. Since its introduction in 2018, this levy has reduced intake of sugar from sugar sweetened beverages (SSB) by an average of 10%, with greater reductions amongst people on lower incomes [8]. Recent poll-data shows that more than two thirds (68%) of the UK public support an extension of the SDIL to other food categories [9].
In 2021, the National Food Strategy: The Plan (NFS) [10] suggested extending the SDIL and proposed that the UK Government should introduce a £3/kg tax on sugar (and some ingredients used for sweetening, but not non-nutritive sweeteners) and a £6/kg tax on salt. This would impact all manufactured food categories in which sugar or salt is used as an ingredient. The report argued that this would generate an incentive for manufacturers to lower the sugar and salt content of their products, through reformulation of recipes or by reduction of portion sizes.
This study estimates the expected health benefits from the reduction in consumption of salt and sugar, and the associated reduction in calories, for four scenarios developed through previous work by the Institute for Fiscal Studies [11], based on the proposed tax [10], each reflecting different manufacturer and consumer responses to a proposed extension of the SDIL to foods high in sugar and/or salt. We also quantify the potential economic impacts resulting from changes in salt and sugar consumption.
Material and methods
Scenarios
In this work, we adopted a methodology that combined elements of the Department of Health and Social Care (DHSC) calorie model [12] (i.e. the calculation of changes in body-mass index (BMI) across the population due to a hypothetical reduction in energy intake from sugar) with life table modelling to quantify health impacts from a reduced intake of sugar and salt in the UK. The potential sugar and salt reduction from the application of the levy proposed in the NFS has previously been calculated by the Institute for Fiscal Studies [10, 11]. Briefly, purchase panel data from the 2019 Kantar Fast Moving Consumer Goods l (Take Home) and 2016-2019 Kantar Out of Home Purchase Panel data were used to estimate changes in intakes which depend on the degrees of reformulation by the industry and the consumer response. There is considerable uncertainty about both of these aspects. Hence, four different scenarios based on this previous work were tested, each reflecting different manufacturer and consumer responses to a proposed tax. Reformulation targets for different food groups set by Public Health England(PHE)—i.e. a 20% reduction in sugar across all food groups [13], and food-group specific targets for salt [14]—were used to define the four scenarios:
The “Low-Low” scenario: Low industry change (firms reformulate to 30% of PHE targets), and low consumer change (consumers substitute away from products by one third of the price increase)
The “High-No” scenario: High industry change (full reformulation to PHE targets), but no consumer change (consumers do not respond to price increases)
The “High-Moderate” scenario: High industry change (full reformulation to PHE targets), and moderate consumer change (consumers substitute away from products by 70% of the price increase)
The “High-High” scenario: High industry change (full reformulation to PHE targets), and high consumer change (consumers substitute away from products by the same amount as the price increase)
According to these calculations a reduction in sugar intake of 4.5-12.6g per day (a 10%-29% reduction [4]) in women and 5.3-15.0g (a 10-27% reduction [4]) in men could be achieved (Table 1). A reduction in salt intake of 0.2-0.7g per day (a 3-9% reduction [4]) in women and 0.3-0.9g per day (a 3-10% reduction[4]) in men could be achieved by introducing the levy.
Based on previous evidence, changes in salt consumption were hypothesised to impact health outcomes directly (stomach cancer, [15]) and indirectly (ischemic heart disease (IHD) and stroke via hypertension [16]) (Figure 1). Changes in sugar consumption were hypothesised to impact health outcomes both directly (IHD,[15, 17]), and indirectly (through BMI [17]).
Hence, like the DHSC model, our model also considered health impacts from changes in BMI resulting from a decreased intake of calories from sugar (Figure 1).
All the scenarios used the following input parameters for the health impact modelling:
Policy lifetime and evaluation period = 25 years. The levy is implemented immediately and is active for 25 years, thus the health-related economic impacts of the levy are also evaluated over 25 years. This timeline was chosen to mirror that used for the NFS [10], but also because this is long enough to show significant health impacts from the policies in question. However, impacts are likely to continue beyond 25 years, so the net present value of these policies could increase further as more years are added to the model.
The health impacts are quantified for the entire adult population in the UK.
Scenarios have age and gender specific calorie reductions, based on previous work by the Institute for Fiscal Studies [11] and given in Table 2.
The primary outcomes of the modelling were: i) changes in life expectancy and Years of Life Gained (YLG) from changes in mortality accumulated over 25 years, ii) changes in morbidity (new cases of different diseases) accumulated over 25 years, iii) changes in Quality Adjusted Life Years (QALYs) from changes in YLG and morbidity over 25 years, and iv) associated monetary changes (£) of the changes in QALYs.
Quantifying reductions in kcal and BMI
In accordance with the DHSC calorie model, 1 gram of reduced sugar intake per day was translated to 4 kcal reduced energy intake from sugar, which in turn was assumed to correspond to a 0.042kg reduction in body weight using the steady-state models originally developed by Kevin Hall et al [18]. Based on the four scenarios, and these assumptions, the energy intake from sugar was translated to a change in body weight ranging from -0.76 to -2.12kg in men, and -0.89 to -2.52kg in women (Table 2) with half of the weight change being achieved in about 1 year and 95% of the weight change in about 3 years [18].
The UK National Diet and Nutrition Survey 2019 [4] was used to provide a nationally representative cohort to which changes in body weight (and BMI) were applied. Each adult in the survey cohort was first assigned their (baseline) BMI based on their height and weight. Similar to the DHSC approach, new body weights (and thus new BMIs) resulting from a reduced energy intake from sugar were calculated for adults considered overweight (defined as BMI ≥ 25 kg/m2) only. The average BMI of the entire baseline adult population was then compared to the average BMI of the adult population after modifications to weights/BMIs were computed under each of the four scenarios. The difference between the average BMI of the total adult population at baseline and the average BMI of the total adult population when the BMI of overweight/obese adults was changed (Table 2), were used as inputs in the life table model. BMI changes ranged from -0.17 to -0.48 amongst men and -0.19 to -0.53 amongst women. Changes in the distribution of BMI across the adult population in the UK after the reduced consumption in sugar are displayed in Supplementary Figure 1. The prevalence of overweight and obesity dropped by 10.9 and 10.5 percent among males and females, respectively (no data shown).
Disease outcomes and relative risks
Dose-response relationships (i.e. relative risks per unit change in consumption) between dietary intake and chronic disease morbidity and mortality were obtained from the 2019 Global Burden of Disease (GBD) study [15]. The relative risk (RR) for a given dietary risk-disease pair estimates how much the risk of mortality (or morbidity) would change when the dietary risk changes. For example, the risk of IHD is reduced by 49% for each 6 g reduction in salt intake (Table 3). A total of 24 disease outcomes (Table 3) were considered for the modelling as they have been shown to be linked to sugar/salt consumption and/or BMI [15]. Further details regarding the relative risks may be found in the Supplementary Methods.
Health impact modelling
Health impacts were the main outcome of this study, and calculated on the basis of a previously applied methodology [19], using the life table model IOMLIFET [20], implemented in R [21]. Life tables were separately generated for males and females, due to their different underlying mortality and morbidity rates. The latest (2019) data on age-specific and sex-specific population-size estimates, and on age-specific and sex-specific all-cause mortality were obtained from the Office for National Statistics [22]. Data on disease-specific mortality (mortality rates) and morbidity (incidence rates) for relevant outcomes (Table 3) were downloaded from the GBD results tool [23]. Together, these data provided the baseline input data for the UK.
Life table models were used to quantify changes in life expectancy and Years of Life Lost (YLL) from changes in dietary risk exposures according to the four different reformulation scenarios (Table 3). Briefly, the IOMLIFET model estimates survival patterns in the population over time based on age-specific mortality rates. Based on the information of a hypothetical change in diet (risk-exposure) and a known exposure-response function (i.e. relative risk), changes in survival rates can be quantified as e.g., YLL or changes to life expectancy. YLL can be explained as the years of life lost for an individual (or a population) as a result of premature avertable mortality, considering the age at which deaths occurred. Since the dietary modifications were expected to reduce mortality rates, YLL were translated to years of life gained (YLG).
Changes in salt and sugar consumption were assumed to be adopted instantly while underlying mortality rates remained constant for the duration of follow-up. In cases where several dietary exposures affected the same disease, the risks were multiplied together as done previously [19]. Changes in YLG were quantified as the additional number of years of life that individuals in the UK adult population would live as a result of reducing their risk of dying prematurely from disease outcomes relating to salt and sugar consumption. Changes in morbidity (new cases of disease) were quantified using the same principles. Morbidity calculations were performed using the output population from the life table as the baseline population, to which changes in risk exposures were applied. Each morbidity calculation was performed separately based on incidence rates, and summed over the UK adult population. Changes in life expectancy at birth were calculated as the difference between the baseline life expectancy (the expected life years divided by the starting population) and the impacted (modelled) life expectancy (the impacted expected life years divided by the impacted starting population). In addition to the central RR estimates, upper and lower 95% CIs for the RRs provided by the GBD were used in the modelling to provide a range for plausible health impacts (Supplementary Table 1). More information regarding data management and assumptions for the health impact modelling may be found in the Supplementary Methods.
Quality Adjusted Life Years and monetary gains
Quality Adjusted Life Years (QALYs) are a measure of disease burden, and consider both the quality and quantity of life lived. Changes in QALYs were modelled by taking into account both changes in YLG from the introduction of a salt and sugar tax, as well as the reduction in disease cases.
The YLG from the health impact modelling were translated to QALYs based on the assumption that an increased duration of life of one year would result in a gain of one QALY [25]. To calculate QALYs based on changes in morbidity, we used a previously developed measure from the DHSC model [12] that represents the health-related quality of life that can be associated with different disease outcomes. Accordingly, a reduction of one case in heart diseases, strokes, diabetes, and cancers was assigned a value of 0.16, 0.18, 0.11, and 0.16 of a QALY respectively. Respiratory disease (asthma) was not part of the DSHC model and thus had not previously been given a health-related quality of life value. We therefore gave it the lowest value in the existent range (0.11 of a QALY). The value assigned to each disease was multiplied by the number of annual reduced cases of that disease in order to provide the number of QALYs resulting from changes in disease outcomes.
Changes in QALYs can be converted into monetised QALYs using a representation of how much society values a QALY. This model assumes the monetary value for a QALY to be £25,000 [26]. Therefore, the change in QALYs was multiplied by £25,000 to produce the monetised QALY from the implementation of the salt and sugar tax. However, monetary values occurring in the future need to be discounted in order to reflect both pure time preference (a preference for something to come at one point in time rather than another merely because of when it occurs in time), and the diminished marginal utility of income (the change in human satisfaction resulting from an increase or decrease in an individual’s income). These are combined with the assumption that real incomes rise over time. Hence, similar to the DHSC model and that recommended by the UK Treasury [27], a discount rate of 1.5% to total monetary gains was applied as a cumulative multiplier to model outputs.
Results
Health impacts and associated monetary gains
The results of the health impact modelling show that average life expectancy in the UK could increase by up to 4.9 months if there was both a high manufacturer and high consumer response to the proposed tax on salt and sugar (Table 4). Even with a minimal response from both manufacturers and consumers, the estimated gain in life expectancy would be 1.7 months. Total YLG accumulated over 25 years from avoided deaths ranged from 1.2 million in the Low-Low scenario to 3.5 million in the High-High scenario, with similar numbers of QALYs being accumulated through reduced morbidity. The largest reductions in new cases of disease came from cardiovascular disease, followed by type 2 diabetes, with smaller reductions in cancers and respiratory disease. The associated monetary gains ranged from £27-£78 billion depending on the scenario.
Differences in YLG between the four scenarios show that the High-Moderate and High-High scenarios are relatively similar in terms of mortality impacts, and that even the scenario assuming no consumer response to the taxes would achieve around 75% of the benefit of the High-High scenario that assumes a high degree of consumer response (Figure 2). The Low-Low scenario assuming a low degree of manufacturer as well as consumer response would achieve less than half of the benefits of the other scenarios. The 95% CI for the health and economic impacts are displayed in Supplementary Table 2.
Discussion
The results of this modelling show that extending the current soft drinks levy to other foods high in sugar and adding a tax on salt could lead to significant health and economic benefits for the UK. Life expectancy in the UK could be increased by between 1.7 months and nearly 5 months, depending on the degree of industry and consumer response to the tax. In the best-case scenario, with a high response from both consumers and industry, the total number of preventable cases of chronic disease could be reduced almost 2 million over 25 years, including reductions in CVD, type 2 diabetes, respiratory disease and cancer. In addition, economic benefits for society of approximately £27 to £78 billion from avoided ill-health could be achieved by introducing the proposed tax. The main part of these gains can be attributed to reduced mortality and morbidity from cardiovascular diseases, which remain the commonest causes of death and disability in the UK and are strongly associated with consumption of salt and sugar.
Our results show that a 29 and 27 percent reduction in sugar intake resulting from the tax among males and females, respectively, could reduce overweight and obesity by as much as 10.9 and 10.5 percent in males and females, respectively. This is similar to other research where the prevalence of obesity was reported to reduce by 1.5 to 10 percent when added sugar intake reduced by 10 to 20 percent [28, 29]. Our results are also consistent with previous modelling studies showing improved health outcomes from diets that better align with dietary recommendations [30–32], and tally with the findings of previous modelling studies that have shown both health and economic gains resulting from a reduced intake of added sugars specifically [33–35]. Health and economic gains have also been reported as a result of the taxation of salt or foods high in salt in other countries [36–39]. For example, taxation of sodium in the United States has been estimated to potentially reduce sodium intake by 9.5% in the population, and could avert nearly a million cardiovascular events over the lifetime of adults aged 40 to 85, increasing QALYs by 2.1 million [37]. Our findings add to this previous body of research, and thus reinforce arguments for why currently high intakes of salt and sugar should be targeted using fiscal policies.
To the best of our knowledge, this is the first study to explore the health effects of the salt and sugar tax proposed by the NFS. This analysis brings together different methods and data sources to explore co-benefits of various pathways of reduced salt and sugar consumption. An important strength is the ability to assess both direct and indirect pathways through which these consumption changes would lead to changes across time and dimensions of health and cost. Compared to looking at indirect vs. direct pathways in isolation, these analyses are therefore likely to capture a more complete picture of the possible health and economic benefits of fiscal measures that reduce intakes of salt and of added sugars. UK dietary recommendations (the Eatwell Guide) recommend that adults limit intakes of salt and added sugars to no more than 6g/d and 30g/d, respectively [5]. Given the UK current average consumption being considerably higher than that, our modelled scenarios, leading to reductions of up to 10% in salt and 29% in sugar, while ambitious, are metabolically rational and concordant with governmental goals.
A limitation to our modelling is that we made no assumptions about possible dietary substitution in our analyses, although removal of salt and added sugars in the diet could result in subsequent substitution of other foods that are instead high in e.g. saturated fats to maintain caloric balance. Our analysis could therefore have over-or under-estimated the health gains. To date, evidence from intervention and modelling studies concerning what happens to consumption of other foods if intakes of salty and sugary foods reduce is sparse. However, recent analyses of food consumption trends in the UK show that intakes of sugar have declined while intakes of fruits, and vegetables in particular, have increased over the last 40 years [40]. Furthermore, several studies indicate that fruit consumption can replace consumption of sweet snacks [41, 42]. Therefore, our models are more likely to have underestimated the health impacts of the proposed tax due to substitution with healthy foods as opposed to overestimating them due to substitution with other unhealthy foods. Measures such as subsidies to reduce the cost of fruits and vegetables could be combined with a salt and sugar tax to facilitate and encourage health-promoting substitutions. In fact, previous studies that explored the combined effects of taxes for unhealthy foods/food components with subsidies (vs. taxes only) reported larger health benefits [36, 38, 43]. Combining taxes with subsidies would also help avoid constraining the resources of the lowest-income households that already dedicate >15% of total budgets to food purchases [44]. The funds raised by the tax could be spent on complementary measures addressing current inequalities around food, such as e.g., an expansion of the free school meal scheme, public health initiatives, and subsidisation of, and increasing access to, health promoting foods.
Despite being adjusted for age, sex, smoking, and physical activity, using RRs in our modelling of health impacts is a limitation because of the possibility for residual confounding. In our case, we applied 95% CIs for the RRs to provide a measure of variation in the selected health outcomes, but these may overestimate the CI widths due to aggregation across disease endpoints. There are also uncertainties related to the survey data used to develop the scenarios. These data are based on purchases and do thus not necessarily accurately capture people’s actual consumption. Moreover, we only modelled the adult population. Considering the rise in childhood obesity in the UK [45], it is likely that the inclusion of children with type-2 diabetes would have generated additional health benefits. Our model also does not take into consideration the effects of socioeconomic status or ethnic origin. Because dietary patterns and disease burdens vary across sociodemographic/economic groups in the UK [46], this could have influenced our final outputs.
This study highlights the significant health and economic benefits that could be achieved by a fiscal measure that reduces consumption of salt and added sugars in UK adults. Despite the potential benefits, food taxes are complex and sometimes controversial, and it is important to consider potential barriers to implementation/effective implementation of any proposed tax. For example, a sugar tax may lead manufacturers to substitute fat for sugar which may have negative consequences for health. Furthermore, industry acceptability is likely to be crucial for avoiding increased costs to consumers, especially to low-income families. Combining taxes with subsidies could be a way to help avoid business losses and thus increase industry acceptability [47]. Equally important is consumer acceptability. Previous research shows that taxes are generally unpopular and seen as intrusive [48, 49]. However, it is promising that substantial public support seems to exist in the UK for an extension of the SDIL to other food categories [9]. Lastly, the success of the proposed tax is also likely to depend on the degree of the management, accountability and public resources available to enable its implementation [50].
Conclusions
By investigating the effects of a combined salt and sugar tax applied to a wide range of food categories, and through a range of different plausible scenarios of manufacturer and consumer response, our work adds new evidence to support the introduction of targeted fiscal measures to reduce both salt and sugar consumption in the UK. These benefit estimates are not definitive and carry significant uncertainty. Nevertheless, they should help determine the order of magnitude of health benefits that might be expected, and thus lend important weight to support health policy makers in their review of options to implement salt and sugar reduction.
Abbreviations
- AF
- Atrial Fibrillation
- BMI
- Body mass index
- DHSC
- Department of Health and Social Care
- GBD
- Global Burden of Disease
- IHD
- Ischemic heart disease
- NFS
- National Food Strategy
- NHS
- National Health Services
- QALY
- Quality Adjusted Life Year
- RR
- Relative Risk
- SDIL
- Sugary drinks industry levy
- SSB
- Sugar sweetened beverages
- YLG
- Years of Life Gained
- YLL
- Years of Life Lost
Data Availability
All data produced in the present study are available upon reasonable request to the authors.
Author disclosures
All authors report no conflicts of interest.
Acknowledgements
We would like to thank Hannah Brinsden and Ria Saha at the Food Foundation for their important contributions to designing this research. PEC designed and conducted the research, analysed the data, wrote the paper, and had primary responsibility for the final content. JM designed the research, provided essential materials and wrote the paper. SP designed the research, provided essential materials and wrote the paper. RG designed and conducted the research, wrote the paper and maintained study oversight. All authors have read and approved the final manuscript.
Footnotes
↵# Rosemary Green is an Editor for Current Developments in Nutrition and played no role in the Journal’s evaluation of the manuscript.
Funding: This project received funding from Impact on Urban Health. PEC had financial support from the Swedish Research Council (VR, grant nr. 2022-00344) for the submitted work. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
We have updated mainly the introduction and discussion.