Rice Consumption in the United States

Background Little is known about rice consumption, related food intake patterns, and the nutritional contribution that rice provides in the diets of Americans. Objective To provide information about rice consumption in the United States and the diets of rice consumers. Design Data come from the Continuing Survey of Food Intakes by Individuals (1994-1996) and the National Health and Nutrition Examination Survey (2001-2002). Respondents report 24-hour recall dietary intakes. The amount of rice available in foods is estimated using the Food Commodity Intake Database. Consumers are classified based on the amount of rice they consume in foods. Subjects The analysis includes information from adult individuals: 9,318 from the Continuing Survey of Food Intakes by Individuals and 4,744 from National Health and Nutrition Examination Survey. Statistics Weighted percentages and mean values show the food and nutrient intake amounts. Logistic regression analysis is used to examine relationships among economic, social, and demographic factors that affect rice consumption. Results Rice is consumed by a significant portion of the US adult population. Compared with others who did not consume rice, rice consumers consumed a smaller share of energy per day from fat and saturated fat; more iron and potassium; and more dietary fiber, meat, vegetables, and grains. Race/ethnicity and education are determinants of the probability of consuming rice, and more so than lowincome status. Conclusions Rice consumers choose a diet that includes more vegetables, a smaller share of energy from fat and saturated fat, more dietary fiber and more iron than those who do not consume rice; the differences have remained relatively stable over the last decade. Accounting for race/ethnicity and income levels is important for better understanding of factors that affect food choices and for effective design of dietary interventions. J Am Diet Assoc. 2009;109:1719-1727. R ice is a major staple among two-thirds of the world’s population. By world standards, per capita rice consumption in the United States is not large, although it has increased during the past several decades, reaching a level of 21.0 lb per capita annually today (1). Rice is a grain product and available as a refined grain (white rice) or whole grain (brown rice) (2). Brown rice provides many nutrients, dietary fiber, and trace minerals. In contrast, white rice, as a refined grain, has a finer texture and improved shelf life but lacks dietary fiber, iron, and many B vitamins. Most of the white rice consumed in the United States is enriched with thiamin, riboflavin, niacin, vitamin B-6, and iron to make the nutritional level of the milled product similar to that of the whole grain (brown rice). All enriched rice is additionally fortified with folic acid (3). Although per capita rice consumption in 2006 was nearly three times that of 1970 (1), little is known about rice consumption patterns and the nutritional contribution that rice provides in the diet of Americans. Some factors contributing to the increase include the growing Asian-American and Hispanic-American populations, new and expanded offerings of rice-based food products, and marketing efforts by the rice industry (4). Dietary guidance recommends substituting whole-grain products, including brown rice, for refined products such as white rice (2). This has implications in food assistance and meal programs. Brown rice was included as a whole-grain product and enriched white rice was excluded from recommended changes to the Supplemental Nutrition Program for Women, Infants, and Children program (5), although minority populations who prefer white rice are often participants of that program. The Centers for Disease Control and Prevention recently reported significant declines in blood folate levels of women of childbearing age (6). According to recent data from the National Health and Nutrition Examination Survey (NHANES), women aged 15 to 49 years obtain an average of 151 !g/day food folate and an additional 128 !g/day folic acid through enriched or fortified foods, like white rice, for total intake of folic acid and dietary folate at a level below the 400 !g/day dietary folate equivalents recommendation of the Recommended Dietary Allowance for women of childbearing age (7,8). Enriched white rice is a good source of folic acid and, like other enriched At the time of the study, S. P. Batres-Marquez was an assistant scientist, Center for Agricultural and Rural Development, Iowa State University, Ames. H. H. Jensen is a professor of economics and head, Food and Nutrition Policy Research, Center for Agricultural and Rural Development, Iowa State University, Ames. J. L. Upton is a nutrition communications specialist in Tiburon, CA. Address correspondence to: Helen H. Jensen, PhD, Department of Economics/CARD, 578 Heady Hall, Iowa State University, Ames, IA 50011-1070. E-mail: [email protected] Manuscript accepted: May 8, 2009. Copyright © 2009 by the American Dietetic Association. 0002-8223/09/10910-0005$36.00/0 doi: 10.1016/j.jada.2009.07.010 © 2009 by the American Dietetic Association Journal of the AMERICAN DIETETIC ASSOCIATION 1719 grains, can help boost folic acid intake, especially important in this population that needs increased folic acid to help prevent serious birth defects (8). In addition, food patterns may be more important than a particular food consumed, and the inclusion of a food item, such as rice, may help develop a better food intake pattern compared to another grain choice. Important socioeconomic and demographic factors that affect grain and rice consumption include income, education, sex, region, urbanization, family size and composition, and ethnicity (9-12). Data available from nationally representative surveys of food consumed by individuals in the United States allow comparison of consumption today (2001-2002) with consumption in the mid-1990s and analysis of factors associated with rice consumption. The analysis is based on data from the United States: the Continuing Survey of Food Intakes by Individuals (CSFII) (1994-1996) (13) and the NHANES (2001-2002) (14), both large, nationally representative surveys of individuals in the United States and the foods they consume. This article provides information on current rice consumption patterns in the United States and the dietary intake of rice consumers in general, as well as rice consumers analyzed by income and race and ethnicity. The primary objective of the research was to gain a better understanding of rice consumption in the United States and whether consuming rice has a beneficial effect on total diet; that is, plays a role in fostering a diet that follows dietary guidance for health (15). This was accomplished by examining the contribution of economic, social, and demographic factors related to rice consumption, and changes in consumption between the two survey periods (1994-1996 and 2001-2002). Extension of the analysis of the general population to low-income individuals and individuals of different race and ethnic backgrounds provides insight into factors that affect the food choices among these groups. The results provide information useful to the design of food programs and nutrition education. METHODS AND PROCEDURES The data come from the 1994-1996 CSFII (13), conducted by the US Department of Agriculture (USDA), and the 2001-2002 NHANES (14), conducted by USDA and the US Department of Health and Human Services. The data are existing data from publicly available sources. This study was deemed exempt from Institutional Board Approval under federal regulation 45 CFR §46.101(b), exemption 4 (16). Both surveys are nationally representative and include data collected through in-person interviews with respondents who provide quantitative 24-hour recall information on their food intake. The CSFII survey collected food intakes on 2 nonconsecutive days. The NHANES has food intake data from one interview day. We report data from one day of intake: Day 1 from the CSFII and the only reported day from the NHANES data. The analysis uses information from 9,318 adults in the CSFII and 4,744 adults in the NHANES, aged 20 years and older with complete intake data for the reported day (day 1 for CSFII and observed day for NHANES). Adults were classified by age groups defined for comparison to previous studies (17) and to identify any differences for the younger adults (aged 20 to 24 years) compared to those older. The food intake data were matched to the Food Commodity Intake Database (FCID) (18) through the common set of food codes to identify consumption of foods containing the commodity “rice.” The FCID converts food intakes (reported as eaten) into food commodities (eg, as white rice, tomatoes, and beans rather than “chili with rice and beans”) by linking foods identified by food codes and the amount eaten with commodity codes and the amount of commodity per 100 g food. The food commodities (over 500) are those listed by the US Environmental Protection Agency in their Food Commodity Master List of June 15, 2000. The FCID was used to identify whether a food item contained the commodity rice and, if so, the corresponding amount of rice (measured as a dry weight). One hundred grams of regular white rice, cooked, is 35.709 g rice dry weight. One serving of regular (refined) rice, according to the 1992 Food Guide Pyramid, is equal to 1⁄2 c rice or 79 g (food as eaten) (17). Thus, one serving (1⁄2 c) of regular, cooked rice is equivalent to 28.21 g rice dry weight. Following similar conversions, one serving (1⁄2 c) brown rice is equivalent to 26.24 g brown rice dry weight (19). The Pyramid Servings Database for USDA Survey Food Codes, version 2.0 (20) provided data for the analysis of the Pyramid food groups consumed by individuals. This database includes data on servings for use with the national food consumption surveys and in amounts consistent with the 1992 USDA Food Guide Pyramid recommendations (The Pyramid Servings Database for USDA Survey Food Codes, version 2.0 was produced by USDA’s Community Nutrition Research Group and updates the earlier version). These data characterize the consumption of foods for the two surveys used and allow comparison between consumers of rice vs other consumers in terms of food groups and food components consumed, including discretionary fat and added sugar intake. Discretionary fat includes amounts of fat above that consumed if the lowest-fat choices were made in all the food groups (eg, amount of fat in 2% milk above the amount of fat in skim milk) (20). Added sugars are sugars and syrups that are added to foods or beverages during processing or preparation. This does not include naturally occurring sugars such as those that occur in milk and fruits (2). The Technical Support Database was the database used to code food data collected in the CSFII 1994-1996 and to calculate the nutrient value of those foods. The USDA Food and Nutrient Database for Dietary Studies, version 1.0 (21) was used to process NHANES 2001-2002. The percentages of energy from fat, saturated fat, and carbohydrates were calculated for each individual as the daily intake of energy from fat, saturated fat, and carbohydrates, respectively, divided by total energy intake. Identification and Classification of Rice Consumers Rice consumers were identified and classified based on the amount of consumption (intake) of rice consumed in foods. Information from the FCID was used to identify foods that contained the commodity (a commodity-based ingredient) rice. Only foods that contained white rice, brown rice, and rice flour were included. Rice from other sources (rice bran or baby foods) was not counted in the selection. The actual amount of rice consumed by report1720 October 2009 Volume 109 Number 10 ing individuals came from the individual’s reported food intake (amount) matched to the commodity amounts in the foods. The classification of “rice consumer” was assigned to individuals who reported eating 1⁄4 c (half of a serving) or more of cooked rice during 1 day. This amount was selected based on the distribution of rice in food products. Other individuals were classified as nonconsumers, even though they may have consumed rice in a very small amount on the interview day, or consumed rice on other days. The amounts of rice consumed were compared across rice consumers by using a “consumption index” to indicate the relative levels of intake compared with the average amounts consumed by rice consumers (22). The index equals 100.0 at the average consumption level for consumers of 1⁄4 c rice or more. A value of 76.0 means that the respective group consumed 76% of the average reported consumption level, or 24% less than the average amount. STATISTICAL ANALYSIS Both data sets were weighted to be nationally representative. The CSFII data were analyzed using Linux SAS (version 9.1, 2002-2003, SAS Inc, Cary, NC) and the NHANES data were analyzed using SAS (version 9.1, 2002-2003, SAS Inc, Cary, NC). Statistical tests using analysis of variance test differences in means (both food group and nutrient intakes) were done using WESVAR 4.2 and accounted for the complex survey design used in each of the surveys. Logistic regression analyses, carried out using SAS protocols (23), were used to examine relationships among economic, social, and demographic factors that affect rice consumption. The logistic model allows estimation of a binary (0,1) dependent variable, and uses a transformation (called logit) based on a prediction equation that restricts predicted values to be between zero and one (24). The logistic regression equation predicts the natural log of the odds for an individual being a rice consumer or nonconsumer. Moreover, the regression coefficients in a logistic regression equation can be used to estimate odds ratios for each of the independent variables. For binary response models, the response Y of an individual can take one of two possible values, denoted for convenience 1 and 0 (Y!1 if rice consumer, Y!0 if otherwise). The linear logistic model estimated was: Logit (p)!log (p ⁄ 1"p)#$%&'x%' where x is a vector of socio-demographic explanatory variables and p!Pr(Y!1|x) is the response probability that was modeled; $ is the intercept parameter, & is the vector of slope parameters, and ' is the error term.