The first session of the workshop, moderated by Christina Economos, co-founder and director of ChildObesity180 and professor and New Balance chair in childhood nutrition at the Friedman School of Nutrition Science and Policy, Tufts University, described global trends in obesity and examined its collective prevalence, costs, and drivers worldwide while also highlighting country and regional differences.
Lindsay Jaacks, assistant professor in the Department of Global Health and Population, Harvard T.H. Chan School of Public Health, and visiting professor at the Public Health Foundation of India in Delhi, set the stage for her talk by stating that obesity is increasing in all regions of the world. She began by discussing global trends in body mass index (BMI) and obesity, pointing to contrasts in these trends between adults and children around the world.
Jaacks reported that adult BMI has increased steadily in the United States and worldwide, according to measured height and weight data from the Noncommunicable Diseases Risk Factor Collaboration (NCD-RisC) for 1975 through 2016 (see Figure 2-1). These trends show “no sign of any substantial plateau, let alone a decrease,” she commented. She speculated that these trends influenced the World Health Organization’s (WHO’s) target for overweight and obesity, simply to halt the increase in prevalence rather than achieve a decline.
Jaacks again referenced data from NCD-RisC to highlight a global trend of increasing BMI among children and adolescents (aged 5–19 years). She noted, however, that the rise in BMI in this age group has not been as steep as that among adults (see Figure 2-2). She explained that despite this apparent lag in the prevalence of child and adolescent overweight and obesity, BMI for this age group is still increasing globally, and a plateau or decline has been observed only in certain sociodemographic groups in a handful of countries.
Jaacks went on to discuss regional differences in mean BMI, based on NCD-RisC data. She highlighted large increases in mean BMI in females in South and East Asia, especially among children. She noted that the dataset’s East Asia region excluded high-income countries such as Japan and South Korea, which she said have observed a plateau in overweight and obesity
prevalence among females. For males, she reported, the data show a steady increase in mean BMI in South, Southeast, and East Asia (NCD-RisC, 2017). Given these trends, she suggested, it is not surprising that there are “huge increases” in the prevalence of obesity globally when the binary indicator of BMI is used.
Jaacks illustrated those overall increases by comparing the global prevalence of obesity in 1975 with the most recent estimates for adults and children (see Figure 2-3). She reported that global obesity prevalence has risen approximately 2 percentage points per decade (NCD-RisC, 2016, 2017). When the first estimates of obesity were available in 1975, she remarked, Russia was the only country with a prevalence greater than 5 percent (NCD-RisC, 2017). Now, she continued, about half of the world’s most populous countries have a prevalence greater than 20 percent. Since 1975, she summarized, there has been a “remarkable increase” in the prevalence of obesity across almost all countries and in all regions of the world.
Jaacks then displayed world maps illustrating differences in obesity prevalence among countries for men, women, boys, and girls. These maps show that the highest prevalence of obesity globally occurs in Pacific Island countries, where it reaches upwards of 60 percent. Prevalence in this region was high initially, and its increase has been substantial, Jaacks observed, especially relative to high-income Asian countries, such as Japan and South Korea, that have an obesity prevalence of around 5 percent (NCD-RisC, 2016, 2017). She added that countries with lower gross domestic product (GDP) tend to have a higher obesity prevalence among women than among
men. As GDP rises, she said, women and men approach each other in obesity prevalence.
Next, Jaacks shared data on the global proportions of adult men and women in low (<20 kg/m2), normal (20 to <25 kg/m2), and high (≥25 kg/m2) BMI categories across several decades. For the first time in history, she remarked, the overall proportion in the high BMI category has surpassed the proportion in the low BMI category for both sexes, according to the most recent data (see Figure 2-4). She acknowledged that globally, undernutrition is still more prevalent than high BMI among children and adolescents (see Figure 2-5), but she cited a projection that the opposite will be true by 2022 if trends of increasing BMI continue (NCD-RisC, 2017).
Jaacks then offered examples of socioeconomic differences, referencing data indicating that women in the top education quartile are more likely to be overweight than those in the bottom quartile, up until a relatively high level of country GDP. Beyond that GDP threshold, she said, women with the least education have a consistently greater probability of being overweight (Goryakin and Suhrcke, 2014). She suggested that this finding helps explain national differences in the prevalence of overweight and
obesity, adding that disparities in prevalence occur both among and within countries.
Regarding geographic variation, Jaacks pointed to increases in the prevalence of overweight and obesity in both rural and urban areas of several major world regions. In approximately half of the countries analyzed, she elaborated, there were greater increases in rural than in urban areas (Jaacks et al., 2015). She focused on India as an example of a country that has a low national prevalence of obesity but substantial variation in within-state obesity prevalence. Appealing for more local data, she mentioned China, Nigeria, and the United States as other large countries with high subnational variability in obesity prevalence (CDC, 2017a; Kandala and Stranges, 2014; National Center for Chronic and Noncommunicable Disease Control and Prevention, 2016).
Jaacks went on to highlight recent estimates of the consequences of the global obesity epidemic. High BMI is among the top risk factors contributing to disability-adjusted life years in high-income countries, as well as across the world overall (GBD 2016 Risk Factor Collaborators, 2017), she reported. She cited as another consequence increasing loss of disease-free years as a person moves from normal to higher levels of BMI (Nyberg et al.,
2018). Third, she said, is a rise in diabetes mortality (IHME, 2018), and she cautioned that diabetes diagnoses will increase if the high-BMI epidemic is not controlled. “The health systems that we have been working with have no capacity to take on this high burden of diabetes,” she warned.
Jaacks closed by reiterating the increasing burden of obesity in every region of the world, with data showing that 50 million girls, 74 million boys, 390 million women, and 281 million men were estimated to have obesity in 2016 (NCD-RisC, 2017).
Vasanti Malik, research scientist in the Department of Nutrition, Harvard T.H. Chan School of Public Health, discussed obesity trends in Asian populations. She opened by explaining that while obesity prevalence increased steadily across all global regions from 1975 to 2014, it remained lower in Asia than in other regions (see Figure 2-6). She reported that in India, according to Patel and colleagues (2015), obesity prevalence is higher in urban than in rural regions, but it is increasing in rural areas as well, particularly among adults. She cited other data indicating that the prevalence of overweight among Asians living in the United States is 27.5 percent, with variation among subgroups: it is highest among Asian Indian and Filipino populations and lower in East Asian populations, such as the Chinese (Barnes et al., 2008).
Malik reminded participants that obesity is linked to a number of chronic diseases, elaborating in particular on its relationship with diabetes. Diabetes prevalence has increased across the globe, she noted, mirroring trends in obesity prevalence (see Figure 2-7). But she clarified that the relationship between obesity and diabetes is not consistent across populations, pointing to a comparison of the United States and eight Asian countries, among which India had the highest prevalence of type 2 diabetes but the lowest prevalence of obesity. This is not what one would expect to see, she observed, suggesting that it raises questions about the relationship between obesity and diabetes in Asian populations. She noted further that Asians develop diabetes at younger ages, illustrating this point with data showing that the prevalence of type 2 diabetes among 30- to 39-year-olds in Asian countries was higher than that among the same age group in the United States (Yoon et al., 2006).
The development of diabetes at younger ages may help explain its rapidly rising prevalence in Asian regions, Malik suggested. She displayed a map of worldwide adult diabetes prevalence for 2017, pointing out that prevalence is comparable in India and the United States (IDF, 2017). She also showed the projected increases in adult diabetes prevalence: the regions expected to experience the greatest rise between 2017 and 2045 are Southeast Asia (84 percent increase), the Middle East and North Africa (110 percent increase), and Africa (156 percent increase) (IDF, 2018). She characterized these projections as alarming, considering that many of these countries have the coexisting burden of undernutrition and health care systems that are not equipped to handle the increasing prevalence of diabetes.
Malik then highlighted the high prevalence of diabetes and cardiovascular risk factors in Asian populations, among whom the average BMI is below 25, the typical threshold for overweight. Yet, she added, while this is the threshold at which metabolic risk factors are observed among Western populations, the same is not necessarily the case among Asian populations. Some of these populations, especially South Asians, tend to have less muscle and more abdominal fat relative to white Europeans, she elaborated, so the same BMI may represent a higher percentage of body fat in the former than in the latter groups. According to Malik, these observations represent a limitation of BMI for assessing adiposity, but may help explain why diabetes tends to occur at lower BMIs and younger ages among Asian populations. She cited the example of two authors sharing the same BMI but differing in percentage of body fat.
Malik continued by reporting that WHO suggested lower BMI cutoff points for overweight and obesity in Asian populations (WHO Expert Consultation, 2004) because of their increased metabolic risk at lower BMIs (and younger ages) compared with Western populations. She pointed out that WHO’s traditional overweight range is 25.0 to 29.9 kg/m2, whereas for Asian populations, it is 23.0 to 27.4 kg/m2; for obesity, the ranges are ≥30 kg/m2 and for Asian populations, ≥27.5 kg/m2. Ethnic-specific cutoff points have also been suggested for waist circumference (WHO, 2011), Malik added.
Malik explained that the concept of ethnic-specific BMI cutoff points has been informed by observations such as those of Deurenberg-Yap and colleagues (2000), who observed a paradox of low BMI and high body fat percentage among Asian subgroups living in Singapore. At the same BMI, she reported, Indians had the highest percentage of body fat, while Chinese had the lowest. These researchers suggested that for the same amount of body fat as Caucasians with a BMI of 30, the BMI cutoff points for obesity would be about 27 for Chinese and Malays and 26 for Indians.
Next, Malik showed data from Gujral and colleagues (2017), who examined the relationship among cardiometabolic abnormalities with normal-weight persons from five ethnic groups in the United States (white, Chinese American, African American, Hispanic, and South Asian). She reported that the prevalence of metabolic abnormalities within the normal-weight category was much higher in South Asians and Hispanics, followed by African Americans and Chinese Americans, compared with whites. The equivalent metabolic abnormalities observed among a white population with a BMI of 25, she elaborated, were observed at lower BMI cutoff points in the other ethnic groups, including as low as 19.6 for South Asians.
“This is telling us that BMI alone is not the best indicator of cardiometabolic risk in most of these Asian populations,” she observed.
Similar findings have been observed in a different population, Malik continued, referencing a study of South Asian, Chinese, European, and
Aboriginal populations in Canada (Razak et al., 2007). For a given BMI, she said, elevated glucose and lipid factors were more likely to be present in South Asian, Chinese, and Aboriginal populations compared with Europeans. According to these researchers, she added, the cutoff point for defining obesity is lower by approximately 6 kg/m2 among non-European groups.
Malik went on to discuss a recent prospective cohort study in India in which significant ethnic differences in the prevalence of type 2 diabetes without excess weight were observed (Gujral et al., 2018). Compared with a white American population, the Asian population experienced higher diabetes prevalence at underweight and normal weight, she elaborated, commenting that diabetes is not seen among the underweight population of white men, and its prevalence is very low in underweight white women. She added that an increased risk for diabetes has also been observed at lower levels of BMI in migrant South Asian groups compared with white individuals or Europeans (Sattar and Gill, 2015). In summary, she said, South Asians develop diabetes at lower weights, at younger ages, and more rapidly (with regard to the progression from impaired glucose tolerance to diabetes) compared with their white counterparts. She stressed the implications of these findings for prevention strategies in Asia, as well as among migrant populations living in the United States and other countries.
Data on the relationship between BMI and metabolic risk among children are sparse, Malik continued. She cited one study that examined differences in body composition and metabolic status between white children in the United Kingdom and Asian Indian children in India. She reported that, despite having lower BMIs, the Indian children had greater adiposity than the white children, and they were also more insulin resistant even after adjustment for adiposity (Lakshmi et al., 2012). However, given the paucity of data, Malik suggested that more research in this area may be useful.
Malik ended by noting that diabetes costs many countries more than $10 billion annually (IDF, 2018), underscoring the importance of obesity and diabetes prevention strategies. Asia’s coexisting problem with underweight has implications for obesity policy, she added. Finally, she proposed incorporating Asian-specific BMI and waist circumference cutoff points in screening programs to help reduce the diabetes burden in Asian populations around the globe.
The world is experiencing rapid ethnic diversification due to an increase in international migration, said Karlijn Meeks, postdoctoral research fellow in the Department of Public Health, Academic Medical Center, University of Amsterdam. According to the United Nations (2017a), she
reported, there were about 173 million international migrants in 2000 and 258 million in 2017. High-income countries host about two-thirds of all migrants, she added, and one in three people is of migrant descent in large European cities such as Amsterdam and London.
Meeks outlined three methods for assessing migrant health. The first and most commonly used is to compare the migrant population with the host population, looking at ethnic differences or ethnic inequalities. The second method is to compare the same migrant group living in different countries, studying the role of national context. And the third is to compare members of a migrant group with their compatriots who have not migrated, studying the role of migration.
As an example of the first method, Meeks pointed to HELIUS (Healthy Life in an Urban Setting), a large cohort study that compared five ethnic minority groups in Amsterdam with the Dutch host population. She reported that all five ethnic minority groups were more affected by overweight and obesity compared with the Dutch, with the highest rates of overweight and obesity being seen in the populations of African descent (Snijder et al., 2017).
Migrants’ destinations matter, Meeks stressed, because the prevalence of overweight and obesity varies widely among high-income countries based on Organisation for Economic Co-operation and Development (OECD, 2017) health statistics. The national context differs in these countries in ways that can influence socioeconomic status, lifestyle, the food environment, and access to health care, she elaborated. Thus, she suggested, comparing the same migrant group across different countries can help pinpoint the contextual factors that drive increased risk for overweight and obesity, findings that can benefit both the migrant and host populations.
To illustrate the second method, which entails exploring how national context influences health behaviors, Meeks displayed data from the Netherlands and England on the prevalence of tobacco smoking. The prevalence of smoking is higher in the Netherlands among the Dutch than in England among the English, she pointed out, and the same pattern is reflected in the migrant populations living in these countries: a higher prevalence of smoking among African migrants living in the Netherlands compared with African migrants living in England, and a higher prevalence among Indian migrants living in the Netherlands compared with Indian migrants living in England (Agyemang et al., 2010).
Meeks then turned to the third method, which involves studying the role of migration, which she said is important for both migrant populations and nonmigrants. Comparing migrant populations with their compatriots who have not migrated, she explained, can reveal lifestyle changes that occurred upon migration, and help identify key predisposing factors for increased risk of overweight and obesity. She pointed out that migrant
populations take results of studies comparing them with their nonmigrant compatriots more seriously than those comparing them with host populations, which they may perceive as unfair. Results are also important to nonmigrant populations, she added, because rapid changes in lifestyle and urbanization experienced by migrants in host countries likely reflect what is happening in low- and middle-income countries and can help forecast future health threats to the nonmigrants.
To illustrate the role of migration, Meeks shared data on the prevalence of overweight and obesity for two different age groups of adult male and female Ghanaians living in three locations. In each age and sex group, prevalence was lowest in rural Ghana, higher in urban Ghana, and highest in the Netherlands, reaching 95 percent among Ghanaian women over age 40 living in Amsterdam (see Figure 2-8).
Meeks next presented insights from the Research on Obesity and Diabetes among African Migrants (RODAM) study, which examined the roles of both migration and national context (Agyemang et al., 2016). Data were collected for nearly 6,400 Ghanaians in 5 locations: residents of rural and urban Ghana and migrants living in London, Berlin, and Amsterdam. Meeks reported that, for both men and women, the prevalence of overweight and obesity (BMI ≥25), obesity only (BMI ≥30), and abdominal obesity (waist circumference >102 cm in men and >88 cm in women) was lowest in rural Ghana, higher in urban Ghana, and highest in the three European cities. According to Meeks, migration played a larger role than national context; however, context played a role as well, given that there was about a 10 percent higher prevalence of overweight and obesity among Ghanaians (both men and women) living in London than in Berlin, for example (see Figure 2-9).
Meeks also pointed to the lower burden of overweight and obesity among Ghanaians in Ghana, based on data published in 2009 (see Figure 2-8), noting that the prevalence of obesity and overweight in urban Ghanaian women was closer to the prevalence in Europe (see Figure 2-9). While the prevalence of overweight, obesity, and abdominal obesity was higher among women than among men in all five locations, she reported, the prevalence of type 2 diabetes was higher among men in every location except rural Ghana (see Figure 2-10).
Meeks echoed Malik’s observation regarding the existence of ethnic differences in the relationship between obesity and health outcomes such as diabetes. She showed data illustrating the positive relationship between BMI and the probability of diabetes among men in the five locations. Among men with the same BMI, the probability of developing diabetes varied by location, she observed, highlighting an example in which the probability was higher in Berlin than in rural Ghana. There was also a positive relationship between waist circumference and the probability of developing
diabetes, she stated, with similar patterns for women. “This clearly illustrates the role of national context beyond just BMI,” she maintained.
To unravel the contributors to overweight and obesity in the RODAM study population, Meeks briefly discussed environmental, genetic, and epigenetic factors. Environmental factors, such as the food environment and diet, physical activity, and stress, she said, differ by national context and can increase the risk of overweight and obesity in migrants compared with nonmigrants. Turning to genetics, she explained that more than 80 loci have been associated with polygenic obesity, but these loci explain only a small fraction of heritability (Locke et al., 2015). More important, she added, data on African populations are limited: only 19 percent of genetic studies are performed in non-European populations, and a majority of that segment consists of Asian populations (Popejoy and Fullerton, 2016). Finally, she observed that lifestyle can affect epigenetics, the cellular mechanisms that regulate gene expression. She cited as an example that if a person starts smoking or makes dietary or physical activity changes, that behavior can induce epigenetic changes that may increase health risk. According to Meeks, while the RODAM study described novel loci associated with obesity in its Ghanaian cohort (Meeks et al., 2017), more research could help determine whether certain environmental factors drive these epigenetic changes to increase the risk of overweight and obesity, or if overweight and obesity induce the epigenetic changes and thereby increase the risk for other diseases.
In summary, Meeks pointed to the high burden of overweight and obesity among African migrants in Europe and an increasing burden in the African region. She added that obesity is an independent risk factor for diabetes, but context matters, and she highlighted the importance of unraveling the complex interplay between genetic and environmental factors as determinants of obesity in Africans.
According to Rachel Nugent, vice president for noncommunicable diseases (NCDs), RTI International, it is important to include the double burden of malnutrition in the conversation about global obesity. Also known as the “dual burden,” the concept of the double burden is relatively new and underresearched, she observed, so a number of its important aspects are not fully described.
Nugent explained that the double burden refers to the simultaneous presence of undernutrition (one or more of stunting, wasting, and micro-nutrient deficiencies) and overweight/obesity, and can be measured at the individual, household, regional, and national levels. According to the Food and Agriculture Organization of the United Nations (FAO), undernutrition affects an estimated 800 million people worldwide, while the problem of
overweight and obesity affects 2 billion people (FAO et al., 2018; WHO, 2018b). In 2014, the Second International Congress on Nutrition framed the issue as malnutrition in all its forms, Nugent said, and she suggested that this broad vision of malnutrition is helping to motivate research on the topic. FAO and WHO are also perpetuating this framing, she noted, and it is gaining greater acceptance, although many of the relationships between undernutrition and overweight and obesity are not yet fully understood.
Nugent shared unpublished maps showing where the double burden of malnutrition is found. She explained that its extent varies depending on the selection of criteria for undernutrition and prevalence of overweight and obesity. At the countrywide level, only a few countries experience a high (40 percent) prevalence of overweight and obesity alongside high levels of undernutrition, she said. But she explained that the two conditions coexist in more countries when the definition is based on a lower prevalence (20 or 30 percent) of overweight and obesity and a broader definition of undernutrition. According to Nugent, this variability reflects the lack of a firm definition for the problem. She added that if subnational-level data were being considered instead of only national-level data, the double burden of malnutrition would appear in even more places.
Nugent went on to explore the drivers and conditions associated with the double burden. She referred to a figure originally developed to examine the drivers of cardiovascular disease (see Figure 2-11), which she said conveys that many environmental drivers are related to diet and to factors upstream from food systems and agriculture. If one examines the changes over time in dietary behaviors and intake patterns, she observed, warning signs emerge along the way. She also argued that shifts in the global food system, such as the commercial sector’s increasing influence over the nutrition conditions in many countries, contribute to the environmental factors that are associated with increases in overweight and obesity alongside continued undernutrition. In addition, she said, it is challenging to measure consumption of processed food accurately. Based on available data, she highlighted a trend of increasing volumes of retail and food service sales for sugar-sweetened beverages and “junk foods” in Chile, Malaysia, and South Africa (Euromonitor, 2018).
Nugent then transitioned to the economic costs of the double burden, first acknowledging that the economic literature on the topic is sparse. This is because the heterogeneity across economic studies of underweight and of overweight and obesity makes it challenging to combine the literature, she explained. She added that only two studies have examined the economic costs of both undernutrition and overweight and obesity, and emphasized that common measures for the two conditions do not exist. For example, she elaborated, the impact of undernutrition may be based on the risk of disease attributable to undernutrition or on educational attainment related
to undernutrition, while the cost of obesity is typically measured in medical expenditures and productivity losses.
In one of the above two studies, Nugent reported, the Economic Commission for Latin America and the Caribbean (ECLAC, 2016) considered multiple pathways to measure the projected costs of undernutrition and overweight and obesity over 65 years in three countries. The study found a total cost ranging from 0.2 percent of GDP in Chile (all from obesity) to 4.3 percent of GDP in Ecuador (2.6 percent from undernutrition and 1.7 percent from obesity). Nugent observed that the study’s method of measuring and combining two different phenomena was “a far from perfect way of measuring the double burden of malnutrition, but it’s the best that could be done.”
In the second study, Nugent continued, conducted in China and India, it was estimated that in 1993, the cost of NCDs due to under- and over-nutrition amounted to about 1.1 percent and 2.1 percent of GDP in India and China, respectively (Popkin et al., 2001). When the estimate for China was updated, she noted, the cost was 4 percent of GDP in 2000 and projected to reach 9 percent by 2025 (Popkin et al., 2006). Despite the lack of both modeling capacity and empirical data, she argued “we can feel pretty certain that there is a significant economic impact from both of these conditions.”
Nugent then cited a number of “double-duty” interventions that can address all forms of malnutrition, adding the caveat that while there is evidence for their impact on one or another form of malnutrition, evidence demonstrating their impact on the double burden is lacking (Shekar et al., 2017; WCRFI, 2018; WHO, 2017). She highlighted three interventions that were chosen for economic analysis based on the ability to get strong data (i.e., data for which there is confidence in the effect sizes, although effects are not necessarily large) for their impact on both undernutrition and overweight/obesity: breastfeeding promotion, school nutrition programs, and food advertising.
To close, Nugent recapped some of the challenges associated with the double burden: a complex set of drivers and conditions, uneven and noncomparable data sources on forms of malnutrition (with much more information available on undernutrition than on overweight and obesity, particularly in developing countries), intergenerational factors that are both epigenetic and environmental, and different outcome measures that reflect the various impacts of malnutrition across the life cycle. Most important, she asserted, is the lack of evidence from double-duty interventions and programs (Ruel and Alderman, 2013).
During a discussion period following the four presentations summarized above, speakers discussed how to better design programs and policy interventions and addressed questions from the audience on epigenetics and environments, food systems, cultural food practices, and trade issues.
Program and Policy Interventions
The four speakers shared ideas for improving the design of programs and policy strategies to address overweight and obesity. Nugent suggested designing interventions that address malnutrition in all its forms. From the perspective of the double burden, she explained, “there is some evidence that we have been causing harm in different parts of the world, where our historical knowledge base and our paradigms are to address undernutrition.” She cited that as a result of the nutrition community’s focus on the various forms of undernutrition, it has been late in realizing the harms caused by changes in environments and food systems. “So we’re really behind the eight ball on addressing the obesity and overweight problem,” she said. She cautioned that basing actions on national data could exacerbate the problem in places where a coexisting burden of underweight and overweight/obesity is apparent only after one looks at subnational data. She urged stakeholders to make programming, funding, and policy decisions that consider malnutrition in all its forms.
Jaacks concurred with Nugent, adding that “the focus has been on undernourishment for so long, and undernourishment continues to be a problem, so you can’t divert all resources to overweight and obesity.” She agreed that collecting more local-level data is critical to identify the problem in a given location—whether undernutrition, overweight and obesity, or both—so that programs and policies can be designed accordingly. Economos advocated for more frequent data collection, adding that “we can’t look at data from 10 years ago and design programs.”
To address migrant health, Meeks proposed moving beyond the common assessment method of comparing migrants with their host population. She acknowledged that “many of those studies show indeed a higher burden in the migrant population,” but, she argued, “to really pinpoint what underlying factors are driving this increased risk, we need to look at other designs as well,” such as those that examine the roles of national context and migration. Genetic, epigenetic, and other underlying factors could explain different rates of certain health outcomes in migrants, she added. Economos suggested that qualitative data could provide additional information about what is going on in a specific population and context.
Malik suggested that an important step toward prevention would be for screening programs to adopt lower BMI cutoff points for Asian populations in the United States and in other high-income countries. She reminded participants that diabetes and other metabolic risk factors develop at lower BMI levels in Asian populations, and thus early intervention is important. She also called for more research to better understand the relationship between body fat and BMI in children. Jaacks mentioned that the American Diabetes Association has adopted a lower BMI cutoff point at which Asian Americans are screened for diabetes.
Epigenetics and Environments
In response to a participant’s question about whether epigenetic changes varied across the RODAM study cohort in the three European countries, Meeks replied that the small sample size of the study cohort that included epigenetic data precluded an examination of epigenetic differences among the European contexts. She noted, however, that her group has compared the epigenetic profiles of Ghanaians in rural Ghana, urban Ghana, and Europe and found that across the genome, many loci differed among those locations. The different environmental factors in those sites is “triggering immense epigenetic changes,” she asserted, and suggested that environmental factors are likely increasing Ghanaians’ disease risk in the European settings.
A participant asked for examples of farming and food distribution practices that countries are implementing to address the double burden. Nugent replied that a field of study called “nutrition-sensitive agriculture” seeks to identify methods for changing agriculture and food systems; she also noted that agriculture programs and studies are not designed to measure nutrition and health outcomes. Notwithstanding the lack of strong evidence of causal connections, she continued, there is reasonably good evidence that improving the processes along the value chain of food production and distribution can deliver healthier food to people. Regarding “upstream” approaches to improve the food system, she suggested considering subsidy, tax, and trade policies that provide incentives to farmers and others in the food system and ultimately support healthier food production and distribution practices. She urged a more “cohesive, integrative way of thinking about these things, because it’s all connected.” Economos underscored the complexity and nonlinearity of food system processes, and Jaacks added that the conversations around agriculture policy are even more complicated in low- and middle-income countries. In India, she observed, the vast major-
ity of farmers own less than 1 hectare of land, so the large-scale, national policies that are discussed for higher-income countries are not as applicable.
Another participant asked Malik to shed light on the continued use of BMI as a screening indicator, particularly for diabetes, and wondered whether a transition to body fat percentage had been considered. Malik replied that the use of waist circumference (as a measure of abdominal adiposity) has been discussed, but a challenge is that it is more complicated to measure, while measurement of BMI is relatively straightforward. Nonetheless, she believes that “waist circumference, if measured properly, would be a better indicator of adiposity.” Jaacks agreed that measuring waist circumference in the field is difficult, and stated that the tools currently available for the purpose would make it difficult to collect such data at the population level.
Cultural Food Practices
According to Meeks, migrants may retain food practices from their home countries even if those practices are no longer suited to their new context. As an example, she recounted a national campaign advocating palm oil consumption in Ghana to help remediate the country’s high rate of vitamin A deficiency. Ghanaian migrants in Europe remembered that campaign, she said, and continued to use palm oil, which is high in saturated fat, even though vitamin A deficiency is uncommon in that region. Malik suggested that additional qualitative research could help clarify the factors influencing dietary choices and inform interventions to improve diet quality. Nugent asserted that qualitative research could help in understanding a culture’s food behaviors and values-driven attachment to certain traditional foods, even though food consumption data may suggest that these foods are no longer commonly consumed.
A participant claimed that rates of chronic disease were lower in some populations when mostly traditional foods were consumed, and asked whether there have been global efforts to help lower-income countries prevent the importation of highly processed foods. Nugent replied that overall, trade has generally had a positive effect on nutrition, although she acknowledged the complexity of the relationships among trade issues, food consumption, and health outcomes.