Research to Advance Understanding of the Interrelationship of Poverty and Nutrition

Food Assistance in the United States 

Food insecurity has been recognized as a problem in the United States for some time, and it has been addressed on a broad and large scale at federal, state, and local levels. The USDA, in partnership with cooperating organizations, administers 15 food and nutrition assistance programs providing children and low-income individuals with access to food, a healthful diet, and nutrition education. The three largest of these programs are the Food Stamp Program, the National School Lunch Program, and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC).

The Food Stamp Program started in the 1930s as the Food Stamp Plan to help the needy purchase food to obtain a nutritious diet. It served 25.7 million individuals per month in 2005 (11). The National School Lunch Program also has its roots in Depression-era efforts to help low-income children. In 2005, the program served 4.98 billion lunches (59% of them being free or reduced in price) to 29.6 million schoolchildren, and the School Breakfast Program served 1.6 billion breakfasts (89% being free or reduced in price) to 9.4 million children (12). The WIC Program provides supplemental foods, nutrition education, and health care referrals for low-income pregnant and postpartum women, infants, and children up to age 5 years. In fiscal year 2005, WIC served 8 million women, infants, and children per month (13).

In addition to the federal programs that address food insecurity, a network of community-based food assistance providers plays a critical role in assisting the food needs of those in poverty. In a study by Ohls and colleagues (14), five major types of community-based organizations that are direct providers of food to those in need were studied, including emergency kitchens, food pantries, food banks, food rescue organizations, and emergency food organizations. Food pantries and emergency kitchens were found to be the main direct providers for emergency food assistance. An estimated 32,700 pantries and 5,300 emergency kitchens were in operation in 2000 in the United States, providing the equivalent of approximately 175 million meals (14). Although these private providers are much smaller in scale than the federal programs, the paper by Ohls and colleagues suggests that both public and private food assistance programs work in tandem providing more comprehensive food assistance than either provides alone.

In this special issue of the Journal, three studies address various aspects of poverty and nutrition in approaches that are unique as well as more traditional. Two studies focus on quite different aspects of fruit and vegetable intake by low-income populations. One study examines the impact of food price on intake and the other examines the relationship of geographic and biomarker data in assessing fruit and vegetable intake. Fruit and vegetable consumption is becoming a prominent indicator of health because of its potential role in chronic disease prevention. The third study compares dietary adequacy of low-income food-secure adults vs food-insecure adults in the Lower Mississippi Delta region, an area with one of the highest rates of poverty in the United States.

Is Cost a Barrier to Consuming Fruits and Vegetables? 

Cost is one of the factors often cited as contributing to less fruit and vegetable consumption among low-income populations. Cost, as well as convenience, has been shown to be a leading influence on food choice for low-income individuals (15). For fruits and vegetables, cost was considered a barrier to intake for the low-income (16, 17). The impact that cost may have on the intake of these foods is of particular concern because USDA’s MyPyramid recommends more fruits and vegetables than its predecessor, the Food Guide Pyramid: 9 servings (based on a reference 2,000-calorie diet) as compared with 5 servings, respectively (18). We know that Americans do not consume the recommended amounts of fruits and vegetables, with fewer than 10% meeting the MyPyramid recommendations (19). Further, those in low-income households were less likely to meet the recommendations (20).

Cassady and colleagues (21) report on a market basket survey of the price of fruits and vegetables conducted at 25 supermarkets selected to represent a variety of income levels in Sacramento, CA, and Los Angeles, CA. In that study, data were collected over three time periods to allow for seasonal variation in produce prices. The types of fruits and vegetables surveyed include 34 low-cost fruits and vegetables identified in the Thrifty Food Plan, 1999. The Thrifty Food Plan is a national standard for a nutritious diet at low cost and includes a set of market basket foods for specific ages and sexes for consumption at home; it meets dietary standards, accounting for the food consumption patterns of US households (22). The costs of these items were used in a comparison of the quantity and cost of 2 weeks of fruit and vegetable intake for a family of four based on recommendations in the Thrifty Food Plan, 1999, vs MyPyramid, and prices by neighborhood income and store type. An important note in this research is that the Thrifty Food Plan, 1999, standards included the Food Guide Pyramid recommendation of 5 servings of fruits and vegetables per day, so the study design provided the means to assess the cost of meeting the increased recommendations for fruits and vegetables.

Cassady and colleagues report that the market basket cost for MyPyramid fruit and vegetable recommendations was significantly less than that of the Thrifty Food Plan, 1999, and was significantly less expensive in low-income areas and in bulk supermarkets. The lower cost of fruit and vegetable recommendations from MyPyramid compared with the Thrifty Food Plan, 1999, was unexpected by the authors. The researchers qualify that these results may be an underestimate of the true difference in cost because the Thrifty Food Plan, 1999, used an optimal food plan, whereas MyPyramid was based on minimum calorie requirements. However, revision of the Thrifty Food Plan, 2006, conducted by the USDA to reflect the current dietary recommendations of MyPyramid, the Dietary Reference Intakes, and food costs, resulted in a plan that cost the same as that in 1999 after adjustment for inflation (22).

The fact that fruits and vegetables were found to be less expensive in low-income areas in these two cities was also contrary to what has been reported in other studies. Because of the care that Cassady and colleagues took to include grocery stores from a variety of income neighborhoods and types of stores, including chain supermarkets, small independent stores, and supermarkets that sold bulk foods, and not including club warehouses that charge a fee, it would seem that their results are strong. It is important to assure that representation of the types of stores available to low-income consumers is included in this type of research. That representation was achieved by selecting stores from very low–income neighborhoods as well as those from within a 5-mile radius of the core areas. Los Angeles and Sacramento represent the first and seventh largest cities, respectively, in California. Although the researchers caution that the results may not be generalizable to cities outside California, the results from these two large urban areas are encouraging.

These important results provide evidence contrary to the belief that cost is a barrier to consuming MyPyramid recommendations for fruits and vegetables for low-income populations. Given that low-income consumers are most influenced by cost and convenience in making food choices, knowledge and understanding of the local food environment is an important component for dietetics professionals counseling low-income clients. Further research is recommended to determine whether similar results are found in other areas of the country and whether not only the fruit and vegetable intake goals, but also all of the MyPyramid recommendations are being met. With such research, food and nutrition professionals will be able to develop the best strategies to help low-income consumers make food choices within their budgets to meet the recommendations of MyPyramid.

Does Where You Live Impact Fruit and Vegetable Intake? 

Stimpson and colleagues (23) report a study testing the hypothesis that neighborhood deprivation will be associated with lower levels of serum carotenoids in comparison with wealthy residential areas. Using data from the Third National Health and Nutrition Examination Survey (NAHNES III) linked with 1990 Census tract data, 17,002 participants age 17 years and older were studied. The levels for five serum carotenoids (β-carotene, α-carotene, lutein/zeaxanthin, β-cryptoxanthin, and lycopene), were examined in relation to neighborhood deprivation, scored using 11 indicators found in the 1990 US Census tract data.

Stimpson and colleagues report that neighborhood deprivation was negatively associated with serum carotenoid levels, with the exception of lycopene. Overall, higher levels of carotenoids were found among older adults, women, non-Hispanic whites, those who are employed, nonsmokers, former/never consumers of alcohol, those who are physically active, nonobese, those with high cholesterol, those with incomes more than $20,000, and those with 12 or more years of education. Lycopene generally followed the opposite pattern.

This study is the first to analyze neighborhood distribution of an objective indicator for fruit and vegetable intake. The research protocol used by these investigators is unique in two important ways. First, the NHANES III dataset was linked to 1990 Census tract data by the National Center for Health Statistics. This linkage allowed the matching of NHANES participants’ health and nutrition measures to characteristics of their households’ neighborhoods from the Census data. Enhancing the NHANES dataset with that of the Census tract data has been called for by an expert panel report from the National Academies of Science on how to improve available data for analyzing food and nutrition policies (24). This recommendation has now been met with these results being reported. Research must continue to explore linkages of large-scale databases, which may be particularly beneficial to dietary research because of the number and variety of factors shown to influence what people eat.

Second, the use of a biomarker to assess fruit and vegetable intake from these data cannot be overstated. The biomarker data to estimate fruit and vegetable intakes provide a more objective measure than self-reported dietary intakes, which are generally regarded as underreported at the group level (25). Given that dietary data were collected in NHANES III, it would be of value to analyze the dietary data for amount and type of fruits and vegetables reported and examine the relationship between serum carotenoid levels and intakes.

The results from Stimpson and colleagues provide a new direction of research investigation to better understand relevant characteristics of the community that may be related to food intake. The analytical capability this linkage provides should be further exploited by researchers to examine the association of other nutrition and health measures collected by NHANES III to the Census data, with indicators of low-income being the underlying theme. For example, what is the prevalence of food insecurity among individuals residing in low socioeconomic residential areas and how does it vary across areas? What is the impact of participation in food assistance programs on diets? Are diets and nutrition-related serum biochemistry values different between food-secure and -insecure participants in these areas?

Dietary Inadequacy in Rural America 

Rates of rural poverty and nutrition-related chronic diseases are among the highest in the nation in the Lower Mississippi Delta region that comprises the areas that border the Mississippi River in Arkansas, Louisiana, and Mississippi. Rates of food insecurity have been reported at twice that of the nation. In 1995, USDA’s Agricultural Research Service launched the Lower Mississippi Delta Nutrition Intervention Research Initiative, devoted to assessment of the nutrition and health needs in the Lower Mississippi Delta region and development of sustainable interventions that will ameliorate those conditions among some of our country’s poorest residents.

In this issue of the Journal, Champagne and colleagues examine the diets of adults from this region of the country (26). They report on a telephone survey of 1,607 adults from this region in which data were collected on food intake, household demographics, and food security status to determine whether diet quality differs between food-insecure and -secure individuals. Diet quality was assessed using the Healthy Eating Index (HEI), Dietary Reference Intakes, and energy density; comparisons between groups were adjusted for diet-related variables, including age, sex, race, education, and household income and size.

The 10 component measures of the HEI used in this research are based on recommendations of the 2000 Dietary Guidelines for Americans and the Food Guide Pyramid (27). A revision of the HEI to assess conformance to the 2005 Dietary Guidelines for Americans and MyPyramid is underway at USDA (28). Champagne and colleagues report that although the HEI scores of those who were food-secure were slightly greater than those who were food-insecure (60.6 vs 57.4, respectively), the difference was not significant when adjusted for the diet-related variables. Twenty-two percent of the sample was categorized as food-insecure. Age and education were found to be related to HEI scores, with younger adults having lower scores and those with a college degree having higher scores, a finding similar to HEI scores determined for the population nationwide (27).

The 10 individual component scores of the HEI are of value in that they provide a more detailed explanation of diet quality. In this research, vegetables are identified as a limiting factor in the diet of the food-insecure. Although component scores were significantly lower for those who were food insecure compared with those who were food secure for dairy, vegetable, cholesterol, and variety; vegetable was the only component related to food security after adjustment by the diet-related variables. Adults categorized as food insecure had a significantly lower HEI vegetable score, equating to more than one serving less than food-secure adults.

Champagne and colleagues (26) compared reported one-day dietary intakes to the Estimated Average Requirements (EARs) and Adequate Intakes (AIs) (for those nutrients for which an AI has been established). The authors rightly discuss the limitation of using only one day of dietary data without adjustment for day-to-day variation in dietary intakes. This is a concern when using the EAR to assess diets of groups. When using nutrient intake distributions that do not represent usual intake or have not been adjusted for day-to-day variation, the nutrient intake distributions are wider than they should be. This may result in an overestimation of the prevalence of inadequacy because the distribution of adjusted intakes may be narrower. For nutrients with an AI, it is not necessary to adjust for day-to-day variation because comparison of the group mean intake to the AI is considered appropriate. A commentary in the Journal provides guidance on using the DRIs to assess intakes of groups (29).

The authors suggest that this impoverished population has limited day-to-day variation in their diets. Understanding and accounting for day-to-day variation in the analysis of dietary intake is an important factor in research determining the prevalence of nutrient inadequacy for a group. Estimates of within-individual variances generated from national dietary data have been used to compare usual nutrient intakes of Americans with the DRIs (30) and are available for researchers to use (31).

Regardless of day-to-day dietary variation, comparison of intakes between food-secure and food-insecure adults using the same benchmark provides valuable information. Food-secure adults were more likely than food-insecure adults to meet their EAR and AI for all nutrients studied. After adjusting for diet-related variables, food-secure adults were more likely to meet the EAR for vitamin A and selenium.

Controlling for variables that may impact dietary intake is important in research, as this study has shown. At first analysis, both diet quality measures were shown to be related to food insecurity, but differences were reduced when covariates were considered. This fact does not minimize the importance of the results reported by Champagne and colleagues in relation to food insecurity. Regardless of food security status, both groups of adults had diets that need improvement. It is clear from this study that the lower Mississippi Delta region is a place in the United States where diets continue to be inadequate. Poverty, lack of resources, and limited education are all factors that contribute to poor diets.

Conclusion 

Although the majority of individuals in the United States are food secure, food insecurity has remained a fact of life for too many people (1). Food insecurity has been shown to have harmful health and behavioral impacts, particularly for pregnant women, children, seniors, and other nutritionally vulnerable groups. Research reported by Cassady and colleagues (21) and Stimpson and colleagues (23) provide further understanding about the factors related to food intake of low-income individuals. Champagne and colleagues’ study (26) further highlights the pressing need for effective nutrition intervention strategies to improve the diets of those in poverty-stricken regions of the country. Most importantly, these authors have done justice to emphasizing the important role that nutrition research must play to unravel the answers about barriers and strategies to improve the diets of the poor.