Table 2. Modelling studies included in the systematic review of dietary trans-fatty acid reduction policies.

Study	Study aim	Policy analysed	Geographical scope	Participants and sample size	Methods	Outcomes	Comments	Qualitya
Allen et al., 201532	To determine the health and equity benefits and cost–effectiveness of policies to reduce or eliminate TFAs from processed foods, compared with consumption remaining at most recent levels in England	(i) Total ban on TFAs in processed foods; (ii) improved labelling of TFAs; (iii) bans on TFAs in sit-down and takeaway restaurants	England	Adults aged ≥  25 years (numbers not stated)	For policies aimed at reducing consumption, health benefits and cost outcomes were calculated for 2015–2020 in England only. Government national data and health economic data from other published studies were used for the model. Adults were stratified by fifths of socioeconomic circumstances	A total ban on TFAs in processed foods: might prevent or postpone about 7200 (2.6%) of the 273 000 total deaths from coronary heart disease from 2015–2020 and reduce socioeconomic inequality in mortality from coronary heart disease by about 3000 of 20 400 deaths (14.7%). Policies to improve labelling: could save 3500 (1.3%) of 273 000 total deaths from coronary heart disease and reduce inequalities by 1500 (7.4%) of 20 400 deaths, thus making them at best half as effective as a ban. Policies to simply remove TFAs from restaurants or fast-foods: could save between 1800 (0.7%) and 2600 (1.0%) of the 273 000 total deaths from coronary heart disease and reduce inequalities by 600 (2.9%) to 1200 (5.9%) of the 20 400 deaths. A total ban would have the greatest net cost savings of about £ 265 million, excluding reformulation costs, or £ 64 million if substantial reformulation costs are incurred	The health outcomes analysis assumed continuing declines in incidence of and mortality from coronary heart disease. The study used an area-based measure of socioeconomic status. Within an area there will be individuals of higher and lower socioeconomic status. Therefore, the study could not make firm conclusions about individuals. As the effect of TFAs operated on a percentage basis (food energy from TFAs divided by total food energy), differences between surveys could only affect the results if consumption were substantially misreported in the surveys used	Good
Martin-Saborido et al., 201633	To assess the added value of EU-level action by estimating the cost–effectiveness of three possible EU-level policy measures to reduce population dietary TFA intake	(i) Status quo; (ii) impose mandatory TFA labelling of prepackaged food; (iii) seek voluntary agreements with food industry and retailers towards further reducing industrially produced TFA content in food; (iv) impose a legislative limit for industrially produced TFA in foods	EU	EU population (numbers not stated)	A computer-simulated model was developed, using effect sizes from different studies, complemented with results from a survey of EU Member States. The model considered three types of cost: (i) health care costs, (ii) non-health care costs and (iii) costs of policy-associated measures	The model estimated that imposing an EU-level legal limit would avoid the loss of 3.73 of 1073 million DALYs due to coronary artery disease over the course of a person’s lifetime (85 years), and making voluntary agreements would avoid 2.19 of 1076 million DALYs. Imposing EU-level legal limits: would save an estimated € 51 billion of € 10 723 635 million in total costs when compared with the reference situation and voluntary agreements would save € 23 billion (of € 10 752 032 million). Implementing mandatory TFA labelling: could also avoid the loss of 0.98 of 1076 million DALYs, but this option incurred greater costs (€ +95 billion) than it saved compared with the reference option	Major sources of potential errors were the estimated current TFA intake; the wide variability observed for many variables between EU countries; and the lack of data in some instances, e.g. lack of data on number of coronary artery disease events per year (coronary artery disease-related hospital discharges were used instead). The results should be interpreted as a comparison between different policy options rather than considering absolute costs, DALYs or deaths per option	Fair
Vyth et al., 201234	To investigate the potential impact on cholesterol levels of consuming a diet consisting of products that comply with the criteria for a healthier choice logo	Food labelling	Netherlands	Dutch adult population (aged 18–70 years) (n = 4336)	The healthier choices logo for food packages was implemented in 2006. National food consumption and food composition data were used to estimate the nutrient intake of the Dutch adult population before and after replacing foods that did not comply with the choices criteria. Four different scenarios were modelled: (i) reference; (ii) minimum, if 24% of the population replaced their food products which complied with the label criteria; (iii) medium, if 48% replaced their food products which complied with the label criteria; (iv) maximum, if 100% replaced their food products which complied with the label criteria. The difference in cholesterol levels in the Dutch population was assessed before and after replacement by means of equations from meta-analyses that calculated how blood lipids change when diet composition changes	Replacing non-complying products with products that complied with the label’s criteria: median intakes of TFA (as a % of total energy intake) would fall from the reference value of 0.95% (2.1 g/day) to 0.80% (1.8 g/day), 0.70% (1.6 g/day) and 0.57% (1.3 g/day) in the minimum, medium and maximum scenarios, respectively	The study was based on theoretical food replacements not people’s actual practices. The study assumed that people would eat the same amounts of replacement foods as their traditional choice, whereas people may eat high amounts of products they perceive to be healthier. The minimum scenario was based on a single study that may not be representative of the general population. The available national representative food consumption data used were based on self-reports, and were outdated	Poor
Roodenburg et al., 201335	To describe a nutrient intake modelling method to evaluate the choices programme – a nutrition profiling system with nutrition criteria for TFAs, SFAs, sodium, added sugar and product groups by investigating the potential effect on nutrient intakes	Food labelling	Netherlands	750 Young Dutch adults (aged 19–30 years)	Data from the 2003 Dutch food consumption survey in young adults and the Dutch food composition tables were combined into a Monte-Carlo risk assessment model. Three scenarios were calculated: (i) actual intakes; (ii) intakes when all foods that did not comply with the healthy choices criteria were replaced by similar foods that did comply; (iii) intakes when food replacements were adjusted for the difference in energy density between the original and replacement food. Another two scenarios were calculated where snacks were not replaced or partially replaced	An estimated reduction of –62% for TFA intake was found when foods complied with the choices labelling programme compared with the actual scenario. TFA intakes in the different scenarios were 2.2 g/day for the actual scenario; 0.8 g/day for the choices labelling programme and 1.0 g/day for the choices labelling programme, adjusted for energy. TFA intakes were 1.3 g/day and 1.4 g/day, when snacks were partially replaced or not replaced, respectively	Replacements chosen may be susceptible to some subjectivity and bias. Product acceptability was not taken into consideration. The same replacement food was used for a large number of snacks. Snacks are usually eaten for indulgence; therefore it is unrealistic to assume that consumers will replace all snacks with the same healthier alternative	Fair
De Menezes et al., 201336	To evaluate the impact of introducing products in agreement with the choices labelling criteria for TFAs, saturated fatty acids, sodium and added sugar in the typical Brazilian diet	Food labelling	Brazil	1720 food products in the Brazilian diet	Data on industrialized and packaged products available in the market in São Paulo state were collected in 2011. The sources of nutritional information were product labels or websites. To evaluate the impact of the consumption of products aligned with the choices criteria, ingestion of key compounds was estimated based on theoretical menus. Typical menus consumed by the Brazilian population were compared with the choices menu (and with the choices menu with energy adjustment). The estimated menus were based on data from a Brazilian household budget survey carried out between 2008 and 2009	Replacement of typical products by those meeting the choices criteria was estimated to cause a decrease in the intake of TFAs of 92%. Estimated TFA intakes were: 0.8 g/day (SD: 1.0) for typical menus; 0.1 g/day (SD: 0.2) for choices menus; and 0.2 g/day (SD: 0.3) for energy-adjusted choices menus, i.e. the same as choices menu, but adjusted for energy of typical menu	The study compared the typical menu with the choices criteria to see how the intake of dietary components might change. There was no specific focus on TFA	Good
Temme et al., 201137	To estimate the impact of recent reformulations of food groups in the Netherlands on median intake of TFA and saturated fatty acids	Food reformulation	Netherlands	750 young adults (aged 19–30 years): 352 men, 398 women	Intakes of TFA were estimated before reformulation (started in 2003), using national food composition data of 2001 as a reference and including most recent TFA composition of foods. Food composition of other foods and food consumption was assumed to be unchanged	Average TFA intake decreased significantly from 2.3 g/day (95% CI: 2.2 to 2.5) to 1.9 g/day (95% CI: 1.8 to 2.0) in the reformulation scenario. Pastry, cakes and biscuits, and snacks contributed most to the decrease of TFA than potato, bread, fats and margarines. Median TFA intakes were 2.3 g/day (95% CI: 2.2 to 2.5) in the reference scenario and 1.9 g/day (95% CI: 1.8 to 2.0) in the reformulation scenario. Estimated reduction in TFA intake was 0.4 g/day (–0.2 of total energy)	Composition data provided by members of the Dutch task force for the improvement of fatty acid composition was purchasing data not actual intake data. Therefore it was not always possible to link this information with food consumption data	Poor
Restrepo and Rieger, 201638	To assess whether Denmark's TFA policy reduced deaths caused by cardiovascular disease	Mandatory food reformulation	Denmark	Danish population (number not stated)	A policy restricting the content of artificial TFA in certain food ingredients was implemented in 2004. Annual mortality rates in OECD and development countries from 1990 to 2012 were used to estimate the effect of Denmark's food policy on cardiovascular disease mortality rates. A synthetic control group was composed of a weighted average of other OECD countries that did not implement the policy. Analyses were conducted in 2015	In the period before the policy (1990‒2003), the mean annual number of deaths per 100 000 people in Denmark were 441.5 and in the synthetic control group were 442.7. In the 3 years after the policy was implemented (2004–2006), mortality attributable to cardiovascular disease decreased on average by 14.2 deaths per 100 000 people per year in Denmark relative to the synthetic control group. The policy reduced male and female cardiovascular disease deaths by 24.4 per 100 000 and 14.3 per 100 000 per year over the 2004–2006 period, respectively. For coronary heart disease, the estimated reduction over the 2004–2006 period was 26.5 deaths per 100 000 people per year	The study investigated what would have happened if mandatory reformulation had not been applied in Denmark. The paper focuses on 2004–2006 before the anti-smoking law was implemented.	Good
Barton et al., 201139	To estimate the potential cost–effectiveness of a population-wide risk factor reduction programme aimed at preventing cardiovascular disease	Legislation to ban industrially produced TFA	England and Wales	Entire population aged 40–79 years (about 50 million)	A spreadsheet model was used, with a range of possible interventions to quantify the reduction in cardiovascular disease over a decade, assuming the benefits applied consistently for men and women across age and risk groups	Legislation to reduce intake of industrial TFA by approximately 0.5% (from 0.8% to 0.3%) of total energy content could prevent approximately 2700 deaths annually and thus gain 570 000 life years and generate savings to the national health service worth at least £ 230 million a year	The study made no attempt to consider recurrent events or subsequent deaths. The estimates of deaths avoided, life years gained and cost savings were thus likely to be underestimates, making the analysis conservative. The study only modelled a 10-year timeframe; reduction in cardiovascular disease would clearly be greater over a lifetime. The analysis was pragmatically limited to people aged between 40 and 79 years at the time of the intervention. This initial modelling lacked a full probabilistic sensitivity analysis	Good
O’Flaherty et al., 201240	To estimate how much more cardiovascular disease mortality could be reduced in the United Kingdom of Great Britain and Northern Ireland through more progressive nutritional targets	(i) Target of 0.5% decrease in the fraction of total energy derived from TFA by 2015; (ii) legislative ban	United Kingdom	Adults aged 25–84 years (number not stated)	Potential reductions in cardiovascular disease mortality in the United Kingdom between 2006 (baseline) and 2015 were estimated by synthesizing data on population, diet and mortality. The effect of specific dietary changes on cardiovascular disease mortality was obtained from recent meta-analyses. The potential reduction in cardiovascular disease deaths was then estimated for two dietary policy scenarios: (i) conservative scenario, with modest improvements (assuming recent trends would continue until 2015); (ii) aggressive scenario. with more substantial, but feasible reductions (already seen in several countries) in saturated fats, industrial TFAs and salt consumption, plus increased fruit and vegetable intake. A probabilistic sensitivity analysis was conducted	In the conservative scenario: reducing the TFA intake by 0.5% in total energy, approximately 3500 of the 12 500 total cardiovascular disease deaths would be prevented. In the aggressive scenario: effectively eliminating the consumption of TFA (to reach 0% of total energy) could result in approximately 4700 of the 29 900 fewer cardiovascular disease deaths (range: 2500–8800) per year	The study did not explicitly model lag times. The study assumed that the effects of food policies on dietary intake in the United Kingdom would be quantitatively similar to those in other countries, without explicitly considering political, commercial, cultural and socioeconomic differences or whether countries’ baseline dietary values were high or low. The study assumed commercial vested interests could be minimized	Good
Pearson-Stuttard et al., 201641	To quantify the potential health effects and costs and benefits of the United Kingdom-wide policies to eliminate dietary intake of TFA	(i) Elimination of industrial TFA; (ii) elimination of both industrial and natural TFA	England and Wales	England and Wales population stratified by age, sex and socioeconomic status (number not stated)	The study extended a previously validated model to estimate the potential effects on health and economic outcomes of mandatory reformulation or a complete ban on dietary TFA in manufactured products in England and Wales from 2011 to 2020. Two policy scenarios were modelled: (i) elimination of industrial TFA consumption from 0.8% to 0.4% daily energy; (ii) elimination of all TFA consumption from 0.8% to 0%	Elimination of all TFA resulted in the largest gains in mortality and life years, with slightly larger gains when modelling unequal baseline TFA by socioeconomic status. Scenario 1 (elimination of industrial TFA only) Annual deaths prevented: 1700. Annual life-years gained: 15 000. Annual hospital admissions averted: 4400. Hospital admissions averted over 10 years: 38 000 Scenario 2 (elimination of all TFA) Annual deaths prevented: 3300. Annual life-years gained: 29 000. Annual hospital admissions averted: 8400. Hospital admissions averted over 10 years: 72 000	The model assumed immediate health benefits. However, rapid improvements might reasonably be expected. The study assumed equal mortality gains from elimination of natural and industrial TFAs	Good
Rubinstein et al., 201542	To estimate the impact of policies to reduce TFA on coronary heart disease, DALYs and associated health-care costs in Argentina	Reformulation (voluntary and mandatory) and mandatory food labelling	Argentina	Adults aged 34+ years (number not stated)	Baseline intake of TFA before 2004 was estimated to be 1.5% of total energy intake. A policy model was built including baseline intake of TFA, the oils and fats used to replace artificial TFAs, the clinical effect of reducing artificial TFAs and the costs and DALYs saved due to the coronary heart disease events averted. To calculate the percentage reduction of risk, coronary heart disease risks were calculated on a population-based sample before and after implementation of the intervention. The effect of the policies was modelled in three ways, based on (i) projected changes in plasma lipid profiles; (ii) projected changes in lipid and inflammatory biomarkers; and (iii) the results of prospective cohort studies. The current economic value of DALYs and associated health-care costs of coronary heart disease averted were also estimated	Baseline number of deaths were: 24 875 for coronary heart disease and 17 942 for acute myocardial infarction. Baseline costs were: US$ 6416 per acute coronary syndrome, US$ 5765 per acute myocardial infarction, US$ 1199 per follow-up and treatment, and US$: 129 001 for programmatic costs. The proportion of CHD events averted by the modelled TFA reduction policy in 2014 ranged from 1.3% (scenario 1) to 6.4% (scenario 3) of the total. The estimated reductions in coronary heart disease were sensitive to the assumed baseline TFA intake in 2004. Based on projected changes in plasma lipid profiles: an estimated 301 coronary heart disease deaths, 572 acute myocardial infarctions, 1066 acute coronary heart disease events and 5237 DALYs would be annually averted after 2014. This is calculated compared with the expected events if the policy had not been implemented. In addition, more than US$ 17 million would be saved annually due to acute coronary heart disease events averted and lower costs of chronic treatment and follow-up. Based on projected changes in lipid and inflammatory biomarkers: using the baseline estimate of 1.5% energy intake from TFA, a total of 3109 acute coronary heart disease events, 15 271 DALYs and more than US$ 50 million in costs would be annually averted after 2014. Based on the results of prospective cohort studies (baseline): an estimated 1517 coronary heart disease deaths, 2884 acute myocardial infarctions, 5373 acute coronary heart disease events and 26 394 DALYs would be averted, resulting in estimated savings of $US 87 million	The cardiovascular risk calculator used was based on equations developed a couple decades before when the coronary heart disease incidence was higher in Argentina. The study used global percentage estimates to adjust for under-reporting of mortality from coronary heart disease. The study only looked at cost from a health system perspective and not at the cost for the industry to reformulate. The study did not have precise data on baseline TFA and the level of this would influence the results	Fair

CI: confidence interval; DALYs: disability-adjusted life years; EU: European Union; OECD: Organisation for Economic Co-operation; £: Pounds sterling; SD: standard deviation; TFA: trans-fatty acid; US$: United States dollars; WHO: World Health Organization.

^a We used adapted version of a published quality assessment tool by Fattore et al..¹⁸