Research Gaps from a Disciplinary Perspective
Key Messages Noted by Participants
Discussions of needed research often use the words transdisciplinary, multidisciplinary, and interdisciplinary, said Christine Olson, moderator of the session on research gaps. She admitted to finding it challenging to plan and implement the research implied by these terms, largely because researchers almost inevitably have discipline-based perspectives that include potentially differing basic assumptions about how the world works. This session was planned to facilitate understanding some of the assumptions of common disciplinary perspectives. The four speakers at the session therefore were asked how their individual discipline would frame impor-
tant issues, questions, and concerns about the relationship between food insecurity and obesity.
A NUTRITIONIST’S PERSPECTIVE
Athens, Ohio, is a rural Appalachian community that is searching for solutions to food insecurity, said David Holben, professor in the School of Applied Health Sciences and Wellness at Ohio University and registered dietitian. More Supplemental Nutrition Assistance Program (SNAP) benefits have been used at its farmers’ markets than in all the other farmers’ markets in the state of Ohio combined. The U.S. Department of Agriculture (USDA) has funded an incubator kitchen that has spawned many new food-based businesses. Also, gleaning of produce grown in the Ohio River valley, including tomatoes, peppers, pumpkins, and potatoes, is an important part of coping with community food insecurity in the region.
Obesity is a multifactorial disease, Holben observed. A person may have a genetic susceptibility to becoming overweight or obese, but that person also has been exposed to an environment that encourages obesity through diet, chemical exposures, lack of physical activity, and so on. Furthermore, these gene-environment interactions take place in a natural environment, a policy environment, and a social environment. The complexity of these circumstances points toward many of the research gaps that exist in the extremely heterogeneous field of nutrition.
Food insecurity can lead to poor dietary intake, malnutrition, and negative health outcomes. Various risk factors can contribute to food insecurity, including limited resources, functional constraints, and poor management strategies. Again, the complexity of even this simplified framework points toward a variety of research gaps.
Measurement of Food Security and Obesity
One of the most basic gaps involves the measurement of food security and obesity. Does the current measure of food insecurity fully capture the experience among children? Is body mass index (BMI) enough to measure obesity, or should a measure such as waist circumference be added?
Causality and Directionality
The relationship between food insecurity and obesity is especially trouble some, as this workshop made clear. If there is a causal relationship, in which direction does causation extend, or does it extend in both directions?
Controlled experiments have been lacking. Prospective and retrospective studies all need to be done, Holben said, with a special focus on critical and sensitive periods of development.
If there is a causal relationship between food insecurity and obesity, then what are the mediating factors? It will be important to test each of the hypotheses that have been proposed and incorporated into conceptual frameworks. Holben expressed the opinion that one of the most important factors is diet quality, in that some diets have adequate calories but lack the qualities needed to promote health. This hypothesis leads to questions such as: What is the diet of people on a low food budget? What kinds of foods do they eat and what are the components of those foods? Do their foods have high levels of dietary contaminants such as bisphenol A?
In this respect, a particularly exciting experiment is being conducted with SNAP in Massachusetts to look at a form of cash transfer that would encourage the consumption of healthful foods. As Holben said, perhaps SNAP should be reoriented to resemble the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) in which specifying healthful foods is acceptable.
Another interesting set of questions surrounds the concept of weight cycling. Does weight gain among individuals from food-insecure households result from consuming inexpensive foods with high energy density when money is less available? Or does inconsistent access to food lead to periods of underconsumption followed by compensatory overconsumption? The answers to these questions could have implications for food benefit distribution schedules.
Lifestyle and daily schedules also can affect dietary patterns. Does working at night disrupt eating? How does smoking affect eating patterns and obesity? What role does a lack of physical activity play?
Answering many of these questions will require multidisciplinary collaborations, said Holben. Economists can look at questions of resource utilization. Psychologists can look at indicators of stress, dietary quality, and hormonal responses. Sociologists can examine socioeconomic deprivation
and early life stressors. Urban planners and geographers can contribute to analyses of physical activity. “It’s vital that we understand the other disciplines and develop policies so we can be more effective in what we do,” Holben concluded.
A SOCIOLOGIST’S PERSPECTIVE
Sociology was launched in many ways by Emile Durkheim’s study of suicide, which is generally considered a very individual choice, noted Janet Poppendieck, professor of sociology at Hunter College in New York City. However, Durkheim was able to show that suicide rates varied tremendously depending on social factors. “That’s very much the task that the research community has taken on with thinking about obesity,” said Poppendieck. “Every time somebody raises the question of environmental as opposed to individual determinants, you are joining the sociologists.”
Sociology has three basic theoretical perspectives—the functionalist, the conflict-oriented, and the interactionist—each with a set of core questions. Most sociologists are associated with one or another of these perspectives, but it takes all three to get a complete picture of most social phenomena.
Functionalists are concerned with order and stability in society, with social structure, and with the ways in which society’s parts fit together. Looking at any social arrangement they ask: What functions (or dysfunctions) does it perform for society? They are particularly interested in finding “latent functions”—impacts that are not obvious. Almost everyone who engages in program evaluation research becomes a kind of functionalist, like it or not, Poppendieck said.
Conflict theorists are especially concerned with the ways in which the benefits of human social arrangements are distributed. Who benefits and who does not benefit from any particular social arrangement? Conflict theorists see society as constantly changing, and they see conflict as the primary engine of change.
Social interactionists focus more on the microsociological level—on interactions between individuals and among small groups, and they ask what meanings are conveyed by these interactions and how such meanings are conveyed. It is especially important to ensure that the interactionist perspective is included in evaluation studies, said Poppendieck. Programs need to be closely observed “on the ground.” Policy intent is one thing, but program outcomes may
sometimes be very different. If students in a particular school characterize school lunch as “welfare food,” no amount of dedicated menu planning may induce them to eat it. If a rumor spreads that a new entrée is “health food,” students may avoid it without giving it a chance.
Sources of Information
One of the best sources of information about what is actually taking place on the ground is the staff—the workforce. Too many program evaluations focus solely on the clients, consumers, or customers, treating the people who actually fix and serve meals or lead exercise classes as mere conduits. “I believe that this is a mistake,” said Poppendieck. “In my experience, workers are often keen observers and have the benefits of a sort of built-in longitudinal perspective. We should listen harder and more frequently.”
Microsociological approaches take a somewhat different approach. For example, geographic information system (GIS) software has made it possible to identify food deserts without ever going there, but it is important to go there and to understand the dynamics taking place in these locations, said Poppendieck.
A focus on place can also permit wider use of “action research” or “community-engaged research” techniques. As an example, Poppendieck cited research on diabetes and other health matters conducted in Humboldt Park in Chicago with the assistance of local community leaders, who not only were consulted in the survey design but also were hired to conduct the interviews. One result was that when the data were compiled, the community was receptive to learning the results and acting on them. The intersection of obesity and food insecurity may be particularly well suited to community-based research interventions, Poppendieck said.
Another line of research that Poppendieck said is urgently in need of attention involves the differential impact of programs in areas with widely differing costs of living. For example, with the National School Lunch Program and School Breakfast Program, school systems in relatively low cost-of-living areas benefit far more than those in high cost-of-living areas. In the former, more of the children in need are likely to be covered by the 185 percent of poverty income threshold for reduced-price meals. This raises
participation rates and, by embracing a larger spectrum of the community, changes the atmosphere and “reputation” of the program, Poppendieck explained. At the same time, lower labor costs in a low cost-of-living area mean that the federal reimbursement goes further; more of the $2.72 is available for food, and the meals are often better (which can further increase participation). According to Poppendieck, careful research should be conducted on the effects of cost-of-living differentials on food assistance programs, especially where food quality is important to the capacity of programs to address obesity and related health concerns.
Speaking to the larger social context, Poppendieck suggested that rather than prohibiting SNAP recipients from purchasing soda with their benefits, the federal government should consider how its subsidization policies affect the nation’s increase in weight. She suggested that subsidized corn, soy, and wheat are “artificially cheap” and are often processed into sweet and salty snack products, concluding that “the structure of our agricultural subsidies is obesogenic for us all.” She pointed to potential research gaps in the realm of history and political science.
Poppendieck asserted that the way we frame an issue makes a difference in ways that we do not intend and probably do not anticipate. She explained her feeling that focusing on “obesity” may be a mistake. “The more we talk about obesity as a problem, the more we’re apparently contributing to a movement in society to stigmatize a distinction between the fat and the fit—a divide.” Poppendieck expressed a belief that a broader, more inclusive concept of diet-related health—healthful eating, healthful lifestyles, healthful weight, a food system that produces nutritious, palatable food—is the direction in which we need to move. She cautioned that the choice of how we frame our research questions affects the discourse that occurs.
A PSYCHOLOGIST’S PERSPECTIVE
Food insecurity has a variety of psychological effects, said Maureen Black, professor of pediatrics at the University of Maryland School of Medicine. It creates stress and anxiety about a lack of food or an inability to meet basic needs. Food insecurity also has been shown in several studies to be associated with maternal depression or depressive symptoms. For example, using childhood longitudinal data, Bronte-Tinkew et al. (2007) showed an association between food insecurity, maternal depression, and
insecure attachment among children. Zaslow et al. (2009) related food insecurity at 9 months to maternal depression, nonresponsive parenting, and insecure attachments and to low mental proficiency at 24 months. Also, the link between maternal depression and children’s growth is particularly apparent in low- and middle-income countries (Black et al., 2009).
Depressive Disorder and Depressive Symptoms
Depressive disorder is a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) diagnosis made through a structured interview. The prevalence of depressive disorder is low compared with the presence of depressive symptoms. The latter is recognized through validated symptom checklists, readily available in the published literature. In a recent paper looking at variations in depressive symptoms among adolescent mothers with very young children, Black and her colleagues found three patterns of depression (Ramos-Marcuse et al., 2010). The first category consisted of women who have never been depressed (41 percent of the sample). The second consisted of women who have multiple depressive symptoms consistently over time (14 percent). The third consisted of women in the middle range who sometimes were above the cut point for depressive symptoms and sometimes were below (45 percent). This is the group Black focused on in her talk.
Depression and Weight
These women in the middle are particularly sensitive to environmental stresses such as food insecurity. However, the response to stress among mothers and children varies according to context. In low- and middle-income countries, the children of depressed mothers are at risk not for obesity but for stunting and for undernutrition (Black et al., 2009; Surkan et al., 2011). In high-income countries, there is some evidence for a link to obesity ( McConley et al., 2011).
If caregivers are highly stressed, they may not be able to see the outcomes of their condition. Caretakers may rely on emotional coping strategies rather than problem-solving coping strategies. “If you find that your best friends are named ‘Ben and Jerry,’ you may feel great today, but you’ll pay a price tomorrow—that’s emotional coping.” Caregivers who rely on emotional coping strategies may be unresponsive to their children during meals. They also may use food to control behavior. “Food is not [an effective] way to manage behavior, but if your child is crying, what is the quickest way to have [the child] stop? Have a cookie.”
Do these behaviors lead to overweight? Not necessarily, said Black. In some contexts, they may lead to overweight. For most children, the snack is
given in addition to the regular meal, and as a result, children consume excess calories and gain weight. In some cases, however, the behaviors may lead to underweight or growth faltering when the snack food replaces the more healthful food or meals. For depressive symptoms, this context can be critical. For example, some evidence from Children’s HealthWatch suggests that the alleviation of stress among mothers through WIC Food Package supplements can have positive effects on children’s health outcomes (Black et al., 2011).
Adult and Child Affect
The impact of caregiver mood on children can best be understood from a transactional perspective, considering both the caregiver and the child (Sameroff, 2009). When a depressed caregiver is matched with a child who is sick or fussy, negative outcomes are much more likely than in other circumstances. When a caregiver is stressed or depressed and a child is happy, that child may be vulnerable to negative outcomes. When the caregiver has a positive affect but a child is fussy or unhappy, the outcome is variable, but a positive caregiving environment can often attenuate the negative effects of a fussy child. If both the caregiver and the child are happy, positive outcomes are most likely. “This paradigm, to my knowledge, has not been applied to food insecurity,” noted Black.
The same sort of analysis can be applied to feeding interactions. A parent can be responsive or nonresponsive to a child. In the latter case, a parent may impose pressure to eat or not to eat, may use food to manage behavior, or may have very limited monitoring of a child’s eating. Parental behavior is guided in part by parents’ perceptions of the child’s size and behavior. Although findings vary, there are suggestions that children of nonresponsive parents are more likely to exhibit poor growth (under- or overweight) than children of responsive parents.
Black recommended that future research include integrated longitudinal investigations of the effects of food insecurity on children’s growth and development. Such studies could look at the psychological functioning of caregivers and children and the food-related interactions between them. She also acknowledged that psychologists do not spend much time thinking about nutrition. Black encouraged integrated research to work through relations linking food insecurity with caregiver mental health, caregiver-child feeding interactions, and children’s growth and development.
AN ECONOMIST’S PERSPECTIVE
It is difficult to describe the perspective of economics on food insecurity and obesity because economists have so many different perspectives,
said Margaret Andrews, an economist in the Food Assistance Branch of USDA’s Economic Research Service. In her presentation, she focused on two: a traditional neoclassical approach that she called the allocative efficiency approach and a more recent school of thought known as behavioral economics.
Allocative efficiency refers to a situation in which resources are allocated in a way that maximizes the net benefit attained through their use. In the context of food insecurity and obesity, this allocation of resources refers to households, which are endowed with such resources as human capital, financial capital, and physical capital. The household maximizes its welfare—based on its preferences—by allocating those resources to participation in the labor market, home production, or other income-generating activities such as participation in public assistance programs. Given the rewards of that resource allocation, the household distributes those gains to both consumption and savings.
The allocative efficiency approach recognizes that resources are not necessarily equitably distributed in society. As a result, the most disadvantaged may allocate their resources efficiently but still not be able to achieve economic security and food security. In that case, said Andrews, food insecurity and obesity are problems that may require interventions that transcend the household level.
Behavioral economics recognizes that rationality is bounded, which is the main difference between it and allocative efficiency, said Andrews. The behavioral economics approach uses concepts from psychology and other social sciences to explain irrational decisions while acknowledging that there are limitations on humans’ ability to optimize resources and behave in what would be considered a rational manner.
One insight from this approach that has been applied to food insecurity and obesity is that consumers tend to be biased toward the present. When they are making decisions, they tend to favor today’s needs relative to the future. If asked about trade-offs in some future situation, they are likely to behave in an economically rational manner, but if the costs are up front, they are likely to have a pronounced present bias. For example, households facing deprivation have a very strong present-based focus. This can result in cyclical eating behavior, particularly in situations where resources are received on a cyclical basis, resulting in food shortages and possible obesity.
From the allocative efficiency perspective, the most important research gap involves the direction of causality, as often mentioned at the workshop. Most researchers have looked at food insecurity as a risk factor for obesity, but to what extent is obesity a risk factor for food insecurity? It can be argued that obesity degrades household productivity. Not only does it result in labor market discrimination, but also it increases the need for nutrients to sustain activity—because heavier people require more calories to maintain a given expenditure of energy—which in turn can require reliance on cheaper calories.
Another research gap involves the incompatibility of variables in studying food insecurity and obesity. Food insecurity is a household condition, while obesity is an individual condition. Obesity is measured at single points in time, while food insecurity is measured across periods of time.
Another major question involves the extent to which food assistance programs should be reformed to improve efficiency. Should more thought be given to how to improve the efficiency of such programs so as to alleviate obesity? Can benefits be modified to increase the choice of individuals or households over how to alleviate resource constraints?
Finally, from the behavioral economics gap, is there sufficient physiological evidence to document the tendency for cyclical eating to cause obesity? Why do women seem to be more susceptible to this problem than men? Would more frequent issuance of SNAP or food stamp benefits alleviate the problem? What would be needed to implement a pilot project?
Moderator: Christine Olson
During the group discussion period, points raised by participants included the following:
Elizabeth Dowler briefly mentioned some of the other disciplinary perspectives that can be brought to bear on food insecurity and obesity. For example, she mentioned a community horticultural project1 in the United Kingdom that has led to some participants being able to reduce their medications for hypertension or depression. Learning to grow foods, teaching others how to grow foods, and becoming engaged in foods have turned
See http://www.sandwellfoodnetwork.org/viewProject.php?id=1 (accessed February 22, 2011).
people’s lives around. Such initiatives are hard to evaluate on a large scale, but evaluations are being done on smaller-scale projects.
Dowler also mentioned a project to map the sources and prices of foods in selected neighborhoods. As part of this work, the researchers had long conversations with shopkeepers about the foods they stocked and their costs. These conversations in turn have helped some shopkeepers change their practices. Holben added that gardening projects provide an opportunity to examine physical activity as well as changes in the food environment.
Diego Rose pointed out that people who live a long distance from a supermarket may gain efficiency by shopping just once a month because of transportation and time costs. For people who rely on food stamps, a transportation deduction or supplement, combined with semimonthly distribution of benefits, would give them more money with which to buy food. Andrews responded that an opt-in program giving households more flexibility in how to receive and use benefits could allow households to behave in ways in which they would like to behave.
Black, M. M., A. H. Baqui, K. Zaman, S. E. Arifeen, and R. E. Black. 2009. Maternal depressive symptoms and infant growth in rural Bangladesh. American Journal of Clinical Nutrition 89(3).
Black, M. M., A. Quigg, P. Casey, D. Cutts, S. Coleman, T. Heeren, J. Cook, M. Chilton, R. Rose-Jacobs, A. Meyers, S. Ettinger de Cuba, E. March, and D. Frank. 2011. WIC attenuates the association between caregiver depression and negative child health indicators. Presented at the Pediatric Academic Societies and Asian Society for Pediatric Research, April 30-May 3, Denver, CO.
Bronte-Tinkew, J., M. Zaslow, R. Capps, A. Horowitz, and M. McNamara. 2007. Food insecurity works through depression, parenting, and infant feeding to influence overweight and health in toddlers. Journal of Nutrition 137(9):2160-2165.
McConley, R. L., S. Mrug, M. J. Gilliland, R. Lowry, M. N. Elliott, M. A. Schuster, L. M. Bogart, L. Franzini, S. L. Escobar-Chaves, and F. A. Franklin. 2011. Mediators of maternal depression and family structure on child BMI: Parenting quality and risk factors for child overweight. Obesity 19(2):345-352.
Ramos-Marcuse, F., S. E. Oberlander, M. A. Papas, S. W. McNary, K. M. Hurley, and M. M. Black. 2010. Stability of maternal depressive symptoms among urban, low-income, African American adolescent mothers. Journal of Affective Disorders 122(1-2):68-75.
Sameroff, A. J. (Ed.). 2009. The transactional model of development: How children and contexts shape each other. Washington, DC: American Psychological Association.
Surkan, P. J., C. E. Kennedy, K. M. Hurley, and M. M. Black. 2011. Maternal depression and early childhood growth in developing countries: a systematic review and meta-analysis. Bulletin of the World Health Organization 89.
Zaslow, M., J. Bronte-Tinkew, R. Capps, A. Horowitz, K. A. Moore, and D. Weinstein. 2009. Food security during infancy: Implications for attachment and mental proficiency in toddlerhood. Maternal and Child Health Journal 13(1):66-80.