Toward Evidence-based Nutrition Policy

Journal Articles

Angell SY, De Cock KM, Frieden TR. A public health approach to global management of hypertension. Lancet. 2015;385:825-7.

Angell SY, Farley TA. Can we finally make progress on sodium intake? Am J Public Health. 2012;102:1625-7.

Angell SY, Yi S, Eisenhower D, et al. Sodium intake in a cross-sectional, representative sample of New York City adults. Am J Public Health. 2014;104:2409-16.

Basu S, Andrews J. Complexity in mathematical models of public health policies: a guide for consumers of models. PLoS Med. 2013;10:e1001540.

Basu S, Lewis K. Reducing added sugars in the food supply through a cap-and-trade approach. Am J Public Health. 2014;104:2432-8.

Basu S, Seligman H, Bhattacharya J. Nutritional policy changes in the supplemental nutrition assistance program: a microsimulation and cost-effectiveness analysis. Med Decis Making. 2013;33:937-48.

Basu S, Seligman HK, Gardner C, et al. Ending SNAP subsidies for sugar-sweetened beverages could reduce obesity and type 2 diabetes. Health Aff (Millwood). 2014;33:1032-9.

Bleich SN, Wolfson JA, Jarlenski MP, et al. Restaurants with calories displayed on menus had lower calorie counts compared to restaurants without such labels. Health Aff (Millwood). 2015;34:1877-84.

Block JP, Chandra A, McManus KD, et al. Point-of-purchase price and education intervention to reduce consumption of sugary soft drinks. Am J Public Health. 2010;100:1427-33.

Block JP, Condon SK, Kleinman K, et al. Consumers' estimation of calorie content at fast food restaurants: cross sectional observational study. BMJ. 2013;346:f2907.

Block JP, Roberto CA. Potential benefits of calorie labeling in restaurants. JAMA. 2014;312:887-8.

Brittin J, Sorensen D, Trowbridge M, et al. Physical activity design guidelines for school architecture. PLoS One. 2015;10:e0132597.

Brownell KD, Kersh R, Ludwig DS, et al. Personal responsibility and obesity: a constructive approach to a controversial issue. Health Aff (Millwood). 2010;29:379-87.

Caro JJ, Briggs AH, Siebert U, et al. Modeling good research practices—overview: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-1. Med Decis Making. 2012;32:667-77.

Choi SE, Brandeau ML, Basu S. Expansion of the National Salt Reduction Initiative: a mathematical model of benefits and risks of population-level sodium reduction. Med Decis Making. 2016;36(1):72-85.

Elbel B, Gyamfi J, Kersh R. Child and adolescent fast-food choice and the influence of calorie labeling: a natural experiment. Int J Obes (Lond). 2011;35:493-500.

Elbel B, Mijanovich T, Dixon LB, et al. Calorie labeling, fast food purchasing and restaurant visits. Obesity (Silver Spring). 2013;21:172-9.

Frerichs L, Brittin J, Sorensen D, et al. Influence of school architecture and design on healthy eating: a review of the evidence. Am J Public Health. 2015;105:e46-57.

Huang TT, Cawley JH, Ashe M, et al. Mobilisation of public support for policy actions to prevent obesity. Lancet. 2015;385:2422-31.

Huang TT, Ferris E, Crossley R, et al. A protocol for developing an evaluation framework for an academic and private-sector partnership to assess the impact of major food and beverage companies' investments in community health in the United States. BMC Obes. 2015;2:36.

Kersh R. The politics of obesity: a current assessment and look ahead. Milbank Q. 2009;87:295-316.

Kersh R. Health reform: the politics of implementation. J Health Polit Policy Law. 2011;36:613-23.

Kersh R. Of nannies and nudges: the current state of U.S. obesity policymaking. Public Health. 2015;129:1083-91.

Laraia BA. Carrots, sticks, or carrot sticks?: using federal food policy to engineer dietary change. Am J Prev Med. 2012;43:456-7.

Laraia BA. Food insecurity and chronic disease. Adv Nutr. 2013;4:203-12.

Lederer A, Curtis CJ, Silver LD, et al. Toward a healthier city: nutrition standards for New York City government. Am J Prev Med. 2014;46:423-8.

Levings J, Cogswell M, Curtis CJ, et al. Progress toward sodium reduction in the United States. Rev Panam Salud Publica. 2012;32:301-6.

Lewis KH, Roblin DW, Leo M, et al. The personal shopper—a pilot randomized trial of grocery store-based dietary advice. Clin Obes. 2015;5:154-61.

Oreskovic NM, Huang TT, Moon J. Integrating mHealth and systems science: a combination approach to prevent and treat chronic health conditions. JMIR Mhealth Uhealth. 2015;3:e62.

Pachucki MC, Lovenheim MF, Harding M. Within-family obesity associations: evaluation of parent, child, and sibling relationships. Am J Prev Med. 2014;47:382-91.

Skouteris H, Huang T, Millar L, et al. A systems approach to reducing maternal obesity: The Health in Preconception, Pregnancy and Postbirth (HIPPP) Collaborative. Aust NZJ Obstet Gynaecol. 2015;55:397-400.

Wang Y, Xue H, Esposito L, et al. Applications of complex systems science in obesity and noncommunicable chronic disease research. Adv Nutr. 2014;5:574-7.

Websites

Access to Nutrition Index
Nutrition scoring system for large global food corporations.

NYC Department of Health and Mental Hygiene. National Salt Reduction Initiative.

NYC Department of Health and Mental Hygiene. National Salt Reduction Initiative Factsheet.

Menustat
A publicly available database of nutritional data for chain restaurant foods.

Pushcart
A website that collects food purchase data and shows how these purchases work for or against health goals.

Prevention Together
An organization created to find evidence-based primary and secondary prevention solutions for childhood obesity operating at scale.

We Move Schools Forward
A website describing systems approaches to promote healthy activity levels in schools.

SNAP to Health
A community for researchers interested in the possible uses of SNAP to promote healthy dietary changes.

Sponsors

Presented by

Obesity, with its concomitant increased risk of cardiovascular disease, diabetes, and cancer, has become an epidemic in the U.S. and many places around the world. Although some believe that dealing with obesity is largely a matter of personal responsibility, public health researchers and many others advocate for local, state, and federal governments to step in with public policies that will assist people with making healthier food choices. Obesity and nutrition researchers are at the forefront of these efforts, charged with gathering evidence to support public policies and with determining which ones are most likely to be effective. On October 16, 2015, researchers gathered to discuss the current state of evidence in nutrition policy making, as well as the status of recent policies implemented with the goal of reducing obesity.

Four speakers summarized research on health-related interventions proposed or implemented in recent years. Sonia Angell described policy initiatives in New York City designed to reduce residents' sodium intake, a primary contributor to hypertension and thus to premature death from cardiovascular disease and stroke. Juan Angel Rivera Dommarco discussed Mexico's national strategy to reduce extremely high levels of obesity, focusing on the recent passage of a sugar-sweetened beverage tax. Barbara Laraia provided an analysis of whether and how the U.S. Supplemental Nutrition Assistance Program (SNAP), a primary safety net for thousands of households at risk of food insecurity, could be used to promote healthier diets. Jason Block discussed the U.S. Patient Protection and Affordable Care Act, explaining a provision of the act that will require comprehensive calorie labeling for all restaurant-type foods, including those sold at outlets such as grocery stores, bakeries, and entertainment venues. The provision is scheduled to be implemented in 2016.

Three speakers discussed new methods of research that are becoming increasingly important in the nutrition and obesity field. Terry Huang discussed systems science, a research approach that is particularly useful for studying the complex interplay of factors that lead to obesity. Matthew Harding described how big data approaches can be used to observe and understand behavior, suggesting ideas for new strategies to change consumer, food industry, and retail store practices that lead to obesity. Sanjay Basu discussed mathematical modeling in nutrition and obesity research, showing how to interpret these studies and explain their results to colleagues who are sometimes more familiar with traditional public health research methods.

Rogan Kersh, a political scientist, provided a reality check by describing the problems inherent in translating public health research into U.S. public policy, particularly in today's political climate. To bring about changes in public policy that will prevent or reduce obesity, nutrition and obesity researchers will need to become more strategic in their choices of research topics and to play a greater role in the dissemination of research results to policy makers and the public.

Highlights

The NYC Department of Health is attempting to reduce premature death from cardiovascular disease and stroke by working with manufacturers to reduce sodium content in prepared foods and restaurant meals and by requiring sodium content labeling for such foods, including special warning labels for very high sodium foods.

Systems approaches have yielded insights into the multifaceted problem of obesity and suggested new policy approaches.

The reality of using evidence to inform nutrition policy

Keynote speaker Sonia Angell of the NYC Department of Health and Mental Hygiene began the conference with a discussion of efforts to reduce sodium consumption, and thus high blood pressure, heart attack, and stroke, among NYC residents. The city's public health plan, OneNYC, focuses on reducing race disparities and inequities, with an overall goal of reducing premature mortality by 25% by 2040. Some regions of the city have much higher levels of premature death, which track closely with areas of high poverty and areas where non-Hispanic black or Hispanic residents predominate, suggesting that premature death is an indicator of health disparities. Heart disease is the second-leading cause of premature death in NYC and 1.6 times more prevalent among non-Hispanic blacks than among non-Hispanic whites. Stroke is 2.7 times higher and diabetes 2.8 times higher in non-Hispanic black populations as well.

Geographic differences in premature death rates in NYC. (Image courtesy of Sonia Angell)

Reduced sodium intake is one of several nutrition interventions that can lower rates of cardiovascular disease. A large body of clinical and epidemiological evidence shows that high sodium intake increases blood pressure, interferes with blood pressure control by medication, and increases the risk of heart disease and stroke. Racial and ethnic disparities in sodium intake have been documented, with higher intake among non-Hispanic blacks and Hispanics. Most sodium intake comes in the form of salt added to processed foods and restaurant meals. Reducing salt in these food sources would thus reduce an important environmental source of dietary sodium.

The NYC Department of Health created the National Salt Reduction Initiative to set targets for reducing salt in processed and restaurant foods. First it was necessary to investigate how much salt foods contained. The department collected nutritional data from public and proprietary sources, building one database of processed food by brand and another of restaurant food, and collecting data on the quantities of each food consumed to find the most problematic sources of extra sodium.

Angell and her colleagues identified specific foods, manufacturers, and restaurants likely to be the most productive targets for intervention. Their conversations with food and restaurant industry representatives after sharing these data led to commitments to reduce sodium levels in key foods from both manufacturers and chain restaurants. The databases will now be used to track changes in sodium levels in food with time. To measure the effects of sodium reductions, the researchers also established baseline sodium intake levels in 2010, using 24-hour urine collection test values from participants in the NYC Community Health Survey.

This work identified many restaurant foods that were intended to be eaten at one sitting but contained more than the recommended sodium intake for a full day. These findings led to the successful passage of a regulation, taking effect in December 2015, that requires NYC chain restaurants to put warning labels on high-sodium foods. The labels include information about the link between high sodium intake and increased risk of heart disease and stroke. Another outcome of this research is a public database called MenuStat that includes full nutritional information for over 150 000 foods and beverages from more than 150 chain restaurants. The database is both a resource for nutrition researchers and a source of nutritional information for the public.

Racial and ethnic differences in daily sodium intake in NYC. (Image courtesy of Sonia Angell)

Angell highlighted the role of public health departments in collecting data to inform nutrition-related public policies. She also described how the departments can expand data collection to new sources, such as the clinic. One way of collecting such data is to use mandatory reporting registries for chronic disease risk factors, similar to the registries for infectious disease reporting that have been in place for many years. The city now requires, for example, that results of A1c tests, a measure of blood-sugar control in people with diabetes, be reported to the Department of Health. Efforts such as these can help identify geographic hotspots where public health interventions would affect the greatest number of people.

Research approaches for a complex system

Many public health problems are multifactorial; obesity, for example, involves a complex web of dietary choices, food environments, and economic, political, and social issues. Terry Huang of the CUNY School of Public Health explained why systems science is in some ways better suited than traditional research methods to deal with this complexity, particularly to capture the interdependent factors and feedback loops characteristic of obesity.

Systems science is the interdisciplinary study of simple or complex systems. It seeks to understand the relationship between the big picture and the details of a system, identifying the details that are most important and recognizing how they fit into the whole. Systems approaches can also capture emergence, the term for how the properties of the details give rise to the properties of the system.

The heterogeneity of obesity lends itself well to systems approaches. Obesity involves diverse entities, such as consumers, food companies, schools, and government bodies, interacting with a wide range of contributing factors—economic, social, political, and psychological. This dynamic, interactive network evolves in response to change, and systems research tracks this evolution. Systems science facilitates a long-term approach that accounts for the types of delays and changing situations that are common in public health research, which frequently features circumstances that are outside researchers' control. Unintended consequences are common in obesity interventions and difficult to anticipate using traditional research tools. An example is the rapid increase in sugar consumption that accompanied the dissemination of low-fat foods in the U.S. in recent years, resulting in the substitution of one unhealthy nutrient for another rather than the nutritional improvements researchers might have expected.

Systems approaches have yielded multipronged intervention strategies. The experimental website Pushcart, for example, gathers food-purchase information from online grocers and makes it available to individual consumers in the context of their health goals. Consumers can see whether their purchases support their stated goals for better nutrition. Another is the Access to Nutrition Index, an accountability tool that evaluates large food and beverage manufacturers for how their products and policies affect undernutrition and obesity worldwide, encouraging more socially responsible behavior.

In addition to informing the design of interventions, systems science can be used to develop or reframe the goals of nutrition research. Huang described the need to create public demand for healthy foods through citizen and political engagement. Focusing on this goal leads to a new set of potential interventions, including nutrition education for families and media involvement to promote such education as a form of civic engagement.

Highlights

Big data approaches can be used to investigate and model consumer behaviors and the effects of interventions, such as taxes on unhealthy foods or specific nutrients.

Public health policy researchers should know how to evaluate research that involves mathematical simulation modeling as these methods become more prevalent.

Building evidence for nutrition policy

Matthew Harding of Duke University described another trend in nutrition research: the application of big data approaches to study consumer, food industry, and retail store behavior, developing information that can be used to influence food choices. Nearly every aspect of food buying is bar-coded and quantified, generating large amounts of data on consumer buying habits. Other sources of big data include geographic information on the types and locations of stores, prescription drug information, clinical data, and genetic data on the people making the purchases. The "Internet of things" may soon add to these sources of big data, generating information on how people make choices in grocery stores: what they look at, how they compare products, and which aisles they frequent. Cell phone data can show the locations of consumers in a store, which raises the possibility of changing behavior by notifying consumers of the locations of healthier food choices as they shop.

The big data that makes this kind of research possible is expanding in variety with the inclusion of new sources such as social media, and in volume from gigabytes to petabytes. Data can also now be collected in real time, not just periodically. However, methods are needed to link the different types of data that are relevant to obesity. Nutritional data, based on the 1.1 million available universal product codes listing nutritional information and ingredients, could be linked with commercial data on purchasing environments, to identify not just the locations of stores but also the quality of the food available in them. Such data could be used to study purchasing behaviors and to identify food deserts where healthy food is not readily available.

Big data can be used to model how public policies, such as taxes on unhealthy foods or specific nutrients like sugar or fat, are likely to affect consumer behavior. Harding described a highly detailed model of U.S. food purchases developed using proprietary transaction databases, data on food purchasing environments, and household demographic information. Because taxes change food prices, the taxes were modeled as variations in food prices, to identify food product groups more sensitive to price changes and to find out if changing the price of one product might increase consumption of other products. Harding found that some taxes, such as those on sugar-sweetened beverages or soda, reduced the amount of calories and sugar purchased. However, a tax on packaged meals increased calories and fat purchased, because these meals are purchased for the sake of convenience and were replaced with similarly convenient and equally unhealthy packaged snacks. In general, compared to specific food-based taxes, nutrient-based taxes were more successful in increasing healthy food consumption in the model, because they were less likely to be circumvented by unhealthy substitutions.

Researchers have also used big data approaches to investigate the relationships between medications and diet; for example, asking whether people starting on statins eat a healthier diet because they are more aware of their blood cholesterol levels or eat a less healthy diet because they think the statin will make up for poor nutritional choices. The study found that people starting a statin prescription purchased fewer servings of cholesterol-containing foods and reduced their saturated fat and sugar consumption. This finding suggests that it might be a good time when patients begin statin prescriptions for physicians to deliver a behavioral nudge, reminding patients of the importance of making healthy food choices or eating less overall.

How to evaluate simulation studies

Mathematical simulation models such as the one Harding described are frequently used to analyze the potential effects of nutrition and obesity policies, and are increasingly finding publication in public health and health policy journals. Sanjay Basu of Stanford University provided an overview of how to choose such models and interpret their results for colleagues who may be more familiar with traditional public health research methods such as case cohort studies and randomized controlled trials. These models are particularly useful in circumstances where randomized trials would be difficult or unethical, or in cases where population-level rather than individual-level analyses are desired.

Basu explained what to look for in papers on simulation models, both from a general public health standpoint and specifically for nutrition and obesity studies. He recommended that readers consider the appropriateness of the model and data inputs chosen, as well as the assumptions made by the study authors, which should be made explicit in the paper.

An example of Markov simulation modeling. (Image courtesy of Sanjay Basu)

Markov modeling and microsimulation methods are used most often in simulation studies. Markov modeling simulates the effects of jumping between discrete states. In a model of the cardiovascular effects of a tax on sugar-sweetened beverages, for example, the participants could be defined as existing in states that might include: being healthy, having a stroke, or having a myocardial infarction. This form of modeling is simple to implement. The states are relatively easy to define, and software to run the model is available for use with common spreadsheet programs. However, these models do not deal well with multiple visits to each state; for example, the person who has more than one myocardial infarction. More states can be added to make up for this deficiency, but the model can quickly become overly complicated, and input data, which comes from the medical literature, may be scarce on individuals who have had multiple myocardial infarctions.

The most common alternative method is microsimulation. These models simulate the behavior of many individual entities to draw conclusions about behavior in the aggregate. In a public health model, the entities would be individual people, represented in a table by characteristics such as age and disease risk factors. Known risk values are used to calculate how outcomes are affected when different aspects of the data inputs change. This type of model does not deal well with situations that involve characteristics that interact with each other. In obesity studies, for example, reduced consumption of sugar-sweetened beverages may lead to increased consumption of a different unhealthy product, confounding results. These models can become overly complex, or there may not be enough specific data on individuals to populate the data table.

An example of microsimulation modeling. (Image courtesy of Sanjay Basu)

Basu explained that it is important to consider whether the level of complexity in a model is appropriate to the question being asked. Sometimes it is more useful to boil a model down to the most important variables and then characterize these variables and their interrelationships in detail, rather than including too many variables that increase complexity and reduce the quality of available input data.

Basu also provided a checklist of three types of analyses that should be present in a publication to allow readers to evaluate whether the results of the model are believable: external validation, sensitivity analysis, and uncertainty analysis. For external validation, a model should be applied to a new, independent set of data to see if the results are the same as with the original dataset. If another dataset is not available, internal validation can be also be used, often by taking a random subset of the data used in the original simulation.

In sensitivity analysis, the value ranges of the input data are changed to see whether and how the results change. Sensitivity analysis can also be used to question the underlying assumptions about which data inputs to a model are most important. Basu showed an example from his own research: simulating a cap-and-trade policy for added sugars, which changes the amount of sugar entering the system from manufacturers, resulted in larger reductions in obesity compared to extra taxes on foods with added sugars, which instead affect consumer behavior. In nutrition and obesity research, sensitivity analysis is sometimes complicated by cross-substitution of foods from different categories, making the effects of price variation smaller and more difficult to detect. In uncertainty analysis, the degree of uncertainty inherent in the model is assessed by looking at the uncertainty of the individual variables used for data input.

Basu described two common weaknesses of model-based studies of nutrition and obesity. One is failure to account for food substitution patterns that can affect measures of price elasticity, the term economists use to describe the relationship between the demand for goods and their prices. Another is use of unequal time horizons when measuring the costs of interventions. One intervention may have immediate results and another take much longer to manifest. To fairly compare immediate and long-term results, Basu recommended modeling on the basis of the full life course of the person being simulated, even though doing so could add additional uncertainties to a simulation.

Highlights

The current political climate, which emphasizes personal responsibility, makes passage and implementation of anti-obesity policies difficult in the U.S.

Mexico has instituted taxes on sugar-sweetened beverages and low-nutrient foods to tackle widespread levels of obesity in its population.

Research-rich, policy-poor

Rogan Kersh of Wake Forest University provided a political scientist's perspective on the problems inherent in translating public health research into policy changes at the local, state, and national levels. He noted that despite thousands of experiments on nutrition and obesity policy and a wealth of data on the links between empty caloric foods and poor health outcomes, meaningful policy changes have only occurred in a handful of places.

Much of the resistance to anti-obesity policies in the U.S. can be attributed to the current political climate, which includes high resistance to "big government" and the so-called nanny state. This resistance reduces support for strong regulatory action and for subsidies that reward healthy behavior. Many policy makers invoke individual responsibility for health eating, as well as the concept of consumer sovereignty, which asserts that government doesn't necessarily know what's good for us. Kersh asserted that these concepts have become dominant because of the concerted efforts of leading right-wing think tanks. He showed examples of how these ideas are promulgated not just in the right-wing media but also by major independent media outlets.

Contradictions inherent in the American approach to nutrition and obesity. (Image courtesy of Rogan Kersh)

At the same time, however, awareness of the toxic food environment has risen. Recent maps produced by the Centers for Disease Control and Prevention make more apparent the differences in the geography of health and nutrition, such as the increased prevalence of obesity, diabetes, and stroke in the Southern states. Media outlets have disseminated the idea that some foods are obesogenic and raised awareness of the substantial increases in serving sizes that have occurred since the 1960s. The business and economic implications of obesity have also come to the forefront, as business journals have started reporting on lost work days and productivity costs associated with obesity-related diseases.

Kersh described several current approaches that fall between the extremes of sole personal responsibility and big government. Some major government nutrition policies are submerged; that is, not readily visible to most citizens. These policies can involve changes in the tax code or actions by the judiciary. One example is tax incentives that have been offered to food manufacturers, such as Pepsi and Lays, to overcome their resistance to reformulating popular products. Such changes are not usually announced by the food companies because of fear that sales will fall. Although overt public policies have larger effects and more staying power, Kersh said, submerged policies are often the best outcome that can be attained.

Kersh described subtle behavioral nudges that government can employ, including psychological approaches such as public service advertising, soda serving-size caps like those in NYC, and behavioral incentives such as fruit subsidies for school lunches. He described the Obama White House, under the leadership of Michelle Obama, as particularly adept at promoting small behavioral changes such as these.

Nutrition policy has been neglected legislatively in recent years. The only piece of obesity-related legislation that passed in the last decade was the Personal Responsibility Act of 2005, which limits fast-food companies' liability in suits by people claiming that fast food has had negative effects on their health. The Affordable Care Act also contains broad requirements for calorie labeling by makers of restaurant-type food (as described more fully later in the program by Jason Block), although this provision has not yet been implemented because of political issues.

Legislators are predisposed to inaction because action changes the status quo and might threaten their reelection. Kersh said that the anti-obesity movement is in need of more political activists to spur legislators. Very few anti-obesity advocates are professional lobbyists. The void is currently filled by an eclectic mix of academics, chefs, filmmakers, and other private citizens. Very few public advocacy groups are involved in this issue either. Nutrition researchers should produce more feasibility and impact studies to encourage lawmakers and should engage the media so that these research results are visible to entities that can influence policy changes, including legislators, courts, nongovernmental organizations, industry, and the public at large.

Hitting the sweet spot

Mexico is one of the few countries to have tackled the obesity problem head-on. Juan Angel Rivera Dommarco of the National Institute of Public Health of Mexico summarized the country's national obesity strategy, which includes recent passage of a tax on sugar-sweetened beverages. Mexico has the highest prevalence of obesity of any country in the world. Its National Nutrition Survey documents a sharp rise in obesity between 1988 and 2012, when the prevalence of overweight and obesity in adult women, for example, rose from 34.5% to 70.8%. Diabetes is also an epidemic in Mexico, with 14% of adults affected.

These findings prompted the development of a national strategy that recognized the need for interventions at the individual and environmental levels. Many of its multifaceted policies have already been implemented. Sodas and unhealthy foods are banned from schools, food marketing to children is regulated, and a consumer-friendly front-of-the-pack nutritional labeling system is in place. Mexico has instituted taxes on sugar-sweetened beverages and energy-dense nutrient-poor foods. Further interventions that are just getting underway include efforts to promote public awareness of diet and physical activity and to support healthy infant feeding, including breast feeding.

The sugar-sweetened beverage tax was supported by compelling evidence for the role of these beverages in the obesity epidemic. Children and adults living in Mexico consume about 12.5% of their total energy intake in the form of sugar-sweetened beverages, and discretionary foods such as beverages, desserts, and snacks make up 24.5% of total energy intake. The World Health Organization recommends that no more than 10% of dietary calories should come from added sugars, with the goal of getting below 5% for additional health benefits. Depending on age and gender, between 58% and 85% of Mexican residents get more than 10% of total energy from added sugars. About 70% of these added sugars come from sugar-sweetened beverages. Dommarco also reviewed considerable research linking the consumption of sugar-sweetened beverages to increased risk of overweight or obesity, as well as to increased risk of diabetes.

Public campaigns in Mexico raised awareness of the sugar content of sugar-sweetened beverages. (Image courtesy of Juan Angel Rivera Dommarco)

The Mexican sugar-sweetened beverage tax was approved in a relatively short period of time, about 3 years. Arguments for the tax focused on its public health benefits rather than on economic considerations. Using data from the Mexican National Income and Expenditure Surveys and the Mexican Family Life Survey, the Institute of Public Health estimated that a price increase of 10% would reduce sugar-sweetened beverage consumption by 10%–12%. Microsimulation studies also showed varying reductions in average bodyweight and rates of diabetes by 2030 and 2050, depending on the tax percentage applied.

Civil society organizations in Mexico participated in the effort to pass the tax with public information campaigns that supported the legislation and educated the public on the health detriments associated with added sugars. Public support was also garnered by proposing to use proceeds from the new tax to install water fountains in Mexican schools, a sorely needed convenience. Nongovernmental organizations lobbied the Mexican Congress and identified legislators willing to present the tax proposal. After considerable debate, the Congress passed an excise tax of $1 peso per liter on all sugar-sweetened beverages, which represents about a 10% tax, as well as an 8% tax on energy-dense, nutrient-poor food products. The tax was approved in 2013 and implemented in 2014.

Dommarco reported that preliminary results indicate that the tax is working: intake of taxed products is down, intake of untaxed healthy substitutes is up, and the tax has generated revenue to fund obesity prevention. The Institute of Public Health will continue to monitor the long-term effects of the tax to keep policy makers informed.

Highlights

The Supplemental Nutrition Assistance Program (SNAP), an important safety net for households at risk of food insecurity, could also be employed to promote healthier diets in recipients.

The Affordable Care Act includes a provision that will require calorie labeling on all restaurant-type foods in the U.S. beginning in December 2016.

Could the U.S. influence dietary change through SNAP?

The focus on federal-level nutrition policies next moved to the U.S., as Barbara Laraia of the University of California, Berkeley, provided insight into whether and how the Supplemental Nutrition Assistance Program (SNAP) could be used to influence dietary changes. Laraia remarked that, given the poor dietary intake of most Americans, a comprehensive national nutrition policy is badly needed. Currently, our only national food policy is the Dietary Guidelines, last updated by the Department of Health and Human Services and the U.S. Department of Agriculture in 2010. All federal government programs are required to comply with these guidelines. Compliance among the U.S. population is measured by the Healthy Eating Index, which scores intake of nine nutritious foods and three foods and nutrients that should be avoided. According to the most recent National Health and Nutrition Examination Survey, the U.S. population scored only 52.7 out of a possible 100 points on this index. Less than 9% of the U.S. population received a score of 70 or higher.

The goal of SNAP, as the largest federal food-assistance program, is to reduce poverty and food insecurity by improving poor individuals' purchasing power through normal retail channels. SNAP represented 75% of the $70 billion spent by the government on food-assistance programs in 2014, when it assisted 23 million U.S. households. SNAP participation has more than doubled since the 2008 recession to about 47 million people, most of whom live below the poverty line. SNAP is an important safety-net program because it responds to changes in unemployment levels, and it has been shown to reduce food insecurity and childhood poverty in the U.S. However, a recent meta-analysis of studies comparing SNAP participants with non-participants who would be eligible for SNAP found that, as currently configured, SNAP has no significant effect on dietary quality.

Laraia has studied the factors that influence whether SNAP benefits are an adequate supplement for households at risk of food insecurity. Household income and individual and household characteristics, as well as the availability of other food-assistance programs, the time available for food preparation, and the food environment in terms of prices, location, and available transportation, all influence food purchasing and consumption patterns, and thus the ability of SNAP to reduce poverty and influence nutrition.

Lariara described several program characteristics that play into benefit adequacy and nutritional impact. These include how the benefit amount is calculated, whether nutrition education is provided, and which retail outlets can be used by program participants. Increases in benefits or in the budget for SNAP Education, a nutrition counseling program, could improve SNAP's effectiveness. The program could also change the food environment for participants by strengthening criteria for vendors, providing guidance to vendors on healthy food product placement, and requiring that vending machines be removed from grocery store premises.

A range of "levers" could potentially be used to encourage healthier food choices in SNAP recipients. (Image courtesy of Barbara Laraia)

It has been proposed that SNAP could be more successful in improving diet quality if it provided incentives for the purchase of fruits, vegetables, and whole grains and/or restricted the purchase of sugar-sweetened beverages and low-nutrition foods. In Massachusetts, the Healthy Incentive Pilot provided SNAP participants with an incentive for purchasing targeted fruits and vegetables. This program increased fruit and vegetable purchases by about a quarter of a cup per participant at a cost of $4.4 million. It would cost around $90 billion over 5 years to implement nationwide. In addition to its cost, the program required a long time for set up and to educate participants. Adding to these deterrents is the concern that tampering with SNAP can reduce participation rates, which might compromise its role as a primary safety net for people living in poverty.

A supportive environment can help people make healthy dietary changes. Laraia suggested that projects to improve the food environment should be pursued before making changes to SNAP. She envisions an integrated, multilevel approach that would start at the local and state levels. Interventions would include taking vending machines out of county buildings, running media campaigns that promote healthy eating, and encouraging retailers to stock more healthy options. These types of changes would benefit everyone, not just SNAP recipients, while protecting SNAP from changes that might reduce enrollment.

Calorie labeling: knowledge is power

Jason Block of Harvard Medical School reviewed a provision of the Affordable Care Act (ACA) that will require calorie labeling for restaurant-type foods. These foods, defined as foods that are prepared away from home and consumed on the premises or shortly after purchase, now make up 43% of the food consumed in the U.S. These types of foods tend to be higher in calories, saturated fat, and salt, and lower in fiber compared to foods prepared at home. Studies show that people underestimate the number of calories in restaurant food by an average of around 250 calories per meal. Underestimation varies by restaurant chain, with foods from restaurants with a healthy halo, such as Subway, subject to larger underestimation.

Underestimation of calories in restaurant meals varies by restaurant chain. (Image courtesy of Jason Block)

A calorie-labeling law for restaurant-type foods was first passed and implemented in 2008 in New York. This law started a policy epidemic as other states picked up the idea, but local and state laws were superseded in 2010 by the ACA, which includes a provision requiring all chain-restaurant establishments with 20 or more sites nationally to post calorie counts on menus. The final regulations for this law were completed in December 2014, and compliance was to begin in December 2015 but has been delayed to 2016. The law requires a succinct anchoring statement on menus and food tags, in addition to calorie counts, that describes daily calorie requirements for adults (or for children and adolescents on food intended to be consumed by them).

The law extends beyond restaurants to cover all establishments selling restaurant-type prepared food, such as grocery stores, coffee shops, bakeries, convenience stores, and entertainment venues. Only schools and institutional food service are exempted. Block explained that this comprehensiveness is important because many non-restaurant outlets, such as grocery stores, sell prepared foods and many of these food are not healthy. About two-thirds of American consumers buy restaurant-type foods in grocery stores, and about two-thirds of grocery stores offer such foods.

Studies of calorie labeling, based largely on the New York experience, unfortunately show no clear effects on food consumption. Block and other have studied food choices at fast-food restaurants before and after implementation of the calorie labeling law, or by comparing New York with similar areas where no such laws are in effect. One study involved collecting receipts from and interviewing about 1600 people leaving restaurants before and after the law was implemented. Although calorie consumption declined at a few chains, the overall average showed no effect. At Subway, calorie counts even went up markedly, possibly because the marketing campaign for the $5 foot-long sandwich occurred at the same time calorie counts were being implemented, illustrating the types of factors that can confound this research. Another study compared demographically matched restaurants in NYC and in Newark, NJ, where calorie labeling was not implemented. This study found that although about half of consumers noticed the calorie labels that were present at the NYC restaurants, only 10%–15% of them said that the labels had changed their food choices or that they had purchased fewer calories. This lack of effect has been shown by several meta-analyses as well.

Block noted that restaurants responded to the New York law by introducing new menu items with fewer calories. He described a few other potential issues for calorie labeling: innumeracy may make it difficult for some people to assess the calorie numbers; others may not understand the concept of calories at all. Better framing of the concepts and ongoing education will be needed to help consumers put the calorie numbers on menus into the context of their overall health and food consumption.

Panel: nutrition policy and the media

The final session of the conference was a panel discussion devoted to the influence of the media on public understanding of nutrition policy. The panelists included Eliza Barclay, a food writer and reporter for National Public Radio; Julia Belluz, a journalist with Vox.com; Helena Bottemiller Evich, a senior food and agriculture reporter for Politico; and Nancy Huehnergarth, a consultant and advocate for nutrition policy. The panelists discussed how they choose stories to pursue, how they vet sources and evaluate the validity of published studies, and what measures they take to maintain a fair balance among opposing points of view. Some of their challenges include the current rapid news cycle, the plethora of inaccurate and unsubstantiated information available about nutrition issues, and the frequent over-interpretation of nutrition study results by both journalists and university public information officers. When asked what researchers could do to improve the accuracy of nutrition reporting, the panelists suggested providing more comprehensive explanations of the background, limitations, and nuances of studies during interviews. They also recommend notifying journalists promptly when errors are found in published reports, particularly because many news stories are published in digital form and thus are easy to correct.