1
Introduction

Changes in the U.S. public’s food consumption and diet-related attitudes and information, together with advances in medical knowledge of dietary effects on health, have heightened awareness of the importance of understanding what people eat and why they eat it. At the same time, there are new challenges for diet, health, and food safety. Biotechnological innovations in food production have raised concerns among some consumers about the safety of some foods. Pathogens, such as E. coli O157:H7 and bovine spongiform encephalopathy (BSE) or mad cow disease, and the threat of terrorism have amplified concerns about food safety. And while nutrient deficiencies in the population remain, among the most pressing dietary problems today are overconsumption of trans and saturated fat, sodium, refined carbohydrates, and total calories, and underconsumption of fruits, vegetables, and whole grains (see www.ers.usda.gov/Briefing/DietAndHealth [June 2005]).

These developments raise important and intriguing policy and research questions. What has caused the increase in overweight and obese Americans? Are people eating more, eating the wrong foods, exercising less, or some combination of these? How do changes in food markets—food prices and availability—affect what people consume? How do other factors, such as income, time resources, and consumers’ preferences and knowledge, affect food consumption decisions, and how have they changed over time? How do factors outside of homes, such as the availability of stores and restaurants, food and food preparation technology, food marketing



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies 1 Introduction Changes in the U.S. public’s food consumption and diet-related attitudes and information, together with advances in medical knowledge of dietary effects on health, have heightened awareness of the importance of understanding what people eat and why they eat it. At the same time, there are new challenges for diet, health, and food safety. Biotechnological innovations in food production have raised concerns among some consumers about the safety of some foods. Pathogens, such as E. coli O157:H7 and bovine spongiform encephalopathy (BSE) or mad cow disease, and the threat of terrorism have amplified concerns about food safety. And while nutrient deficiencies in the population remain, among the most pressing dietary problems today are overconsumption of trans and saturated fat, sodium, refined carbohydrates, and total calories, and underconsumption of fruits, vegetables, and whole grains (see www.ers.usda.gov/Briefing/DietAndHealth [June 2005]). These developments raise important and intriguing policy and research questions. What has caused the increase in overweight and obese Americans? Are people eating more, eating the wrong foods, exercising less, or some combination of these? How do changes in food markets—food prices and availability—affect what people consume? How do other factors, such as income, time resources, and consumers’ preferences and knowledge, affect food consumption decisions, and how have they changed over time? How do factors outside of homes, such as the availability of stores and restaurants, food and food preparation technology, food marketing

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies and labeling policies, and incorporation of advances in dietary knowledge into health care delivery, affect what people are consuming and the consequences for their health and safety? Given that the prevalence of obesity is greater among low-income than other households, what effects have food assistance and educational programs had on the nutritional quality of diets of those served by the programs, and are these programs effective in improving diets and health? How do food consumption patterns affect food markets—for example, how do different weight-loss programs affect the purchasing and consumption of different foods? Where do people buy and consume food, and how does food preparation affect food safety? How does consumption of specific foods change after a food safety outbreak? Many different kinds of information are needed to address these questions and to formulate or adjust policy: information on food expenditures, food consumption, food prices, where food is purchased and consumed, food preparation, diet and health knowledge, and possible sources of contamination. For example, to understand if foods with relatively high energy content and low nutrient content are being consumed rather than healthier foods because they are relatively cheaper or more readily available, data on food consumption, prices, and availability are needed. For some purposes, longitudinal data on the behavior of the same households over time are needed. For other purposes, data are needed on a very timely basis in order to make decisions based on current or recent market conditions. While there are rich sources of data on food consumption and related issues, gaps exist, and no single source contains all of the information needed to answer these questions. For example, there are good data from the Consumer Expenditure Survey (CE) on food purchases by households. However, these data do not have information on who in a household consumes how much of the food; they do not contain detailed information on food consumed away from the home; and they do not include information on prices paid for specific quantities of particular foods. The National Health and Nutrition Examination Survey (NHANES) collects critical information on food consumption and health and nutritional status that has many uses for policy making and research. However, it does not now collect data on how much was spent on the food that was consumed. The ability to link such economic information as food purchasing, food consumption, and household socioeconomic characteristics to survey information on what consumers know about diet, health, good food preparation practices, and food safety issues is also lacking.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Data from government-sponsored national surveys, which are the primary sources available for policy and decision making, usually take a few years to collect and process. Yet for some policy-making decisions, data on the most recent market conditions are needed by the U.S. Department of Agriculture (USDA). Data on the supply of food, from growers and food manufacturers, are available on a relatively timely basis. However, with existing data sources, actual consumption of this food can only be inferred on an aggregate level. The most timely data on food purchases and food consumption are collected by market research firms to assess trends in food consumption. These data are collected for proprietary reasons, are of uncertain quality, and require payments, unlike data from government surveys, which are generally free to users. Because of the lack of access to timely data, the secretary of agriculture has less information for making some decisions affecting markets, programs, and the health of U.S. citizens than the executives of many companies in the private sector. Interagency efforts have been undertaken to fill data gaps and develop comprehensive nutrition, health, and food consumption monitoring data systems. Most notably, in 1993, the Interagency Board on Nutrition Monitoring and Related Research (IBNMRR), chaired by senior officials of the U.S. Department of Health and Human Services and USDA, prepared a 10-year plan for a comprehensive nutrition monitoring and related research program. Mandated by the National Nutrition Monitoring and Related Research Act of 1990 (P.L. 101-445), the plan’s goals were to further the collection of continuous, coordinated, timely, and reliable data by federal and state agencies; foster the use of comparable methods for collecting data and reporting results; promote related research; and disseminate and exchange information with data users. The IBNMRR published directories of federal and state nutrition-related datasets and monitoring activities and commissioned reports on the dietary and nutritional status of the U.S. population (see, e.g., Federation of American Societies for Experimental Biology, 1995; National Center for Health Statistics, www.cdc.gov/nchs/about/otheract/nutrishn/nutrishn.htm [June 2005]). The mandate for the board expired in 2003, although work remains to be done to achieve its goals. In addition, work is required to link other kinds of data, such as prices, to information on consumption choices and the consequences for diet and health.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies PANEL CHARGE AND CONTEXT In recognition of existing data gaps, the Economic Research Service (ERS) of the USDA received funding from Congress to improve the data infrastructure on food consumption and nutrition. As part of this effort, ERS asked the Committee on National Statistics of the National Academies to convene a panel of experts to review data needs to support research and decision making for USDA food and nutrition policies and programs. The panel was also charged to assess the adequacy of the current data infrastructure and to recommend enhancements to improve it. For both tasks, the panel was asked to consider improvements to the current data systems, rather than new data systems. The primary basis for the panel’s deliberations, given limited resources, was a workshop on Enhancing the Data Infrastructure in Support of Food and Nutrition Programs, Research, and Decision Making, which the panel convened on May 27-28, 2004. The workshop served as a forum for the USDA and other federal agencies with related policy responsibilities to discuss continuing and emerging data needs for policy and decision making, the data sources available to address these issues, and possible improvements or alternative data sources. During the workshop, representatives from six USDA agencies, the Food and Drug Administration (FDA), the National Institutes of Health (NIH), and the Environmental Protection Agency (EPA) discussed current and emerging data needs for policy and decision making related to food consumption. Representatives from the two key federal statistical agencies that produce food consumption and expenditure datasets, the National Center for Health Statistics (NCHS) and the Bureau of Labor Statistics (BLS), and representatives from private firms that produce data on food consumption and expenditures, the NPD Group and ACNielsen, discussed the strengths and limitations of their data. Outside researchers gave their reactions to the presentations and suggested possible improvements to the data infrastructure. The workshop summary is included as Appendix A. This report is based on the discussions at the workshop and the deliberations of the panel. The report outlines key data that are needed to better address questions related to food consumption, diet, and health; discusses the available data and some limitations of those data; and offers recommendations for improvements in those data. The panel was charged to consider USDA data needs for policy making and the focus of the report is on those needs. It is important, however, to recognize that many policy

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies issues and the data required to address them fall under the purview of other agencies with related policy missions. For example, EPA and FDA both share responsibility with the Food Safety and Inspection Service of USDA to ensure the safety of the nation’s food supply. Likewise, several agencies within DHHS have missions related to the diet and nutritional adequacy of food consumed by Americans. This report does not explicitly cover the food consumption data needs of these other agencies, but it includes their needs when they overlap with those of the USDA. The panel was charged to consider incremental changes in existing data systems that could be implemented (1) in a relatively short time frame, (2) at modest expense, (3) for general analytical use. A more comprehensive study would also consider data needs for assessing specific problems and population groups of interest, regardless of which agency was responsible and putting aside considerations of resources and time for development—as one example, the data requirements to fully understand the causes and consequences of obesity. Another example is the data requirements for assessing the health and nutrition of low-income people who receive (or are eligible for but do not receive) benefits from food assistance programs.1 The panel’s study was conducted with the benefit of previous and continuing work of other studies from the National Academies. In the late 1990s, the Committee on National Statistics convened a workshop on evaluating food assistance programs in an era of welfare reform. The workshop participants considered data needs and research methods for assessing the effects of programs and changes in programs on household economic and food security and individual health and well-being (National Research Council, 1999). The workshop report discussed most of the relevant federal datasets on food and nutrition in some detail but did not make recommendations for changes in them. Workshop participants also discussed the utility of linking survey data with relevant administrative records, as well as the need for more state-level data. In 2003 the Board on Agriculture and Natural Resources held a work- 1   For this purpose, Logan, Fox, and Lin (2002) reviewed almost 100 data sources for their potential to support food assistance program outcomes research, identifying 13 data sources that are clearly useful, another 13 sources that could be useful if they were expanded in one or more ways, and 70 sources that are not useful because they are outdated, restricted to specific populations, lack sufficient content on program participation, nutrition, and health, or comprise administrative records that would be difficult to link to datasets with program participation information.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies shop to consider ways to more efficiently and effectively conduct food and health research and education for promoting better health. In particular, the workshop considered how to begin to integrate relevant input from the agricultural and health sciences, although it did not explicitly consider the data infrastructure needs for integration (National Research Council, 2004). The Institute of Medicine’s Food and Nutrition Board and Board on Health Promotion and Disease Prevention recently completed a study to assess the factors responsible for the epidemic of obesity in children and identify promising approaches for prevention efforts. That study reviewed the scientific literature on the causes of childhood obesity and on obesity prevention programs and recommended a research and action agenda to assist in the prevention of obesity (Institute of Medicine, 2005). The study was not specifically charged to address improvements in the data infrastructure for evaluation of obesity prevention programs, but it did recommend that the federal government “strengthen support for relevant surveillance and monitoring efforts, particularly the National Health and Nutrition Examination Survey” (Institute of Medicine, 2005:6). Finally, the Committee on National Statistics (CNSTAT) has under way a study of the measurement of food insecurity and hunger by the USDA; that measurement is obtained from a supplement to the December Current Population Survey (see Nord and Bickel, 2002; Nord, Andrews, and Carlson, 2004). The CNSTAT study recently completed a first phase report (National Research Council, 2005b) and will issue a final report that examines in depth the current food insecurity measure and possible alternatives to it. So as not to duplicate work, this report on improving data to analyze food and nutrition policies does not consider food insecurity measurement, although there is a brief discussion in Chapter 4 of the food insecurity scale and other food-related information in the Current Population Survey December supplement. The rest of this chapter provides a context for consideration of data needs by briefly reviewing important factors that influence food purchasing and consumption behavior and the consequences for diet and health. It also provides a limited review of food consumption data needs for food safety policies and programs. Box 1-1 lists the acronyms used throughout the report. Table 1-1 summarizes key features of the major public and private surveys of data on food consumption, expenditures, and store sales that are considered in the report. Chapters 2-4 examine the three sources that can provide data relevant to the panel’s charge to examine the data infrastructure for food and

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies BOX 1-1 Acronyms Used in the Report ACS American Community Survey APHIS Animal and Plant Health Inspection Service, USDA ATUS American Time Use Survey BEA Bureau of Economic Analysis BLS Bureau of Labor Statistics BMI Body Mass Index BRFSS Behavioral Risk Factor Surveillance System BSE Bovine Spongiform Encephalopathy (mad cow disease) CDC Centers for Disease Control and Prevention, DHHS CE Consumer Expenditure Survey CPS Current Population Survey CSFII Continuing Survey of Food Intakes by Individuals DHHS U.S. Department of Health and Human Services DHKS Diet, Health, and Knowledge Survey ECLS Early Childhood Longitudinal Study EPA Environmental Protection Agency FCBSM Flexible Consumer Behavior Survey Module FDA Food and Drug Administration, DHHS FSIS Food Safety and Inspection Service, USDA HHANES Hispanic Health and Nutrition Examination Survey HRS Health and Retirement Study IBNMRR Interagency Board on Nutrition Monitoring and Related Research IRI Information Resources, Inc. MEC Mobile Examination Center, NHANES MSA Metropolitan Statistical Area NET National Eating Trends NET Nutrition and Education Training Program NCHS National Center for Health Statistics, DHHS NHANES National Health and Nutrition Examination Survey NIH National Institutes of Health NIS National Immunization Survey OMB Office of Management and Budget PCE Personal Consumption Expenditures PDP Pesticide Data Program PSID Panel Study of Income Dynamics RDC Research Data Center SLAITS State and Local Area Integrated Telephone Survey UPC Universal Product Code USDA U.S. Department of Agriculture WIC Special Supplemental Nutrition Program for Women, Infants, and Children

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies TABLE 1-1 Overview of Major Federal and Private-Sector Surveys on Food Consumption, Food Purchases by Consumers, and Food Sales in Stores Type of Survey Food Consumption (Intake) Surveys Survey National Eating Trends (NET) National Health and Nutrition Examination Survey (NHANES) Sponsor NPD Group DHHS/USDA Description Food intake by individuals from 2-week diary; socioeconomic characteristics of all household members; where food purchased and eaten Food intake by sampled persons from 24-hour recall for 2 nonconsecutive days beginning 2002 (first in-person, second by phone); household characteristics; food assistance program participation; socioeconomic characteristics, where food eaten, and detailed health measures from medical tests and examinations for sampled persons Sample Nationally representative; 2,000 households sent diaries each year Nationally representative; some groups oversampled; 5,000 people examined each year Frequency and Timeliness Ongoing panel; 3-month lag between collection and release Continuing since 1999; released approximately every 2 years with 2-year lag from collection; 2001-02 available (does not include day 2 of dietary intake) Response Rate (approximate) Not available 82 percent of eligible sample persons interviewed; 76 percent examined (1999-2000 round) NOTE: For more information on NHANES, CSFII, and DHKS, see Chapter 2; for NET, see Chapter 3.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Discontinued Consumption Surveys Continuing Survey of Food Intakes by Individuals (CSFII) – Integrated with NHANES in 2002 Diet and Health Knowledge Survey (DHKS) – Source of questions for new Flexible Consumer Behavior Survey Module in NHANES USDA USDA Food intake by sampled persons from 24-hour recall for 2 nonconsecutive days (in-person, 1994-1996 round); household characteristics; food assistance program participation; last month’s food expenditures; socioeconomic characteristics of members aged 15 and older; where food eaten and health measures for sampled persons Supplement to CSFII in 1989-91 and 1994-96; questions about diet and health knowledge and attitudes, use of food labels, factors in shopping, food preparation practices Nationally representative; oversampling of low-income people; 5,000 people per year over 3 years One adult aged 20 and older who completed a dietary intake from each household in CSFII sample Conducted most recently in 1989-91, 1994-96, and 1998 (children aged 0-9 only); most questions now in NHANES Conducted as supplement to CSFII in 1989-91 and 1994-96; not currently available 80 percent of sampled household members completed first day of dietary intake; 76 percent completed second day (1994-1996 round) 74 percent of eligible adults (see “Sample” above)

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Type of Survey Food Purchases by Consumers Survey Consumer Expenditure Survey (CE) Combined Outlet Consumer Panel Source BLS Information Resources, Inc. (IRI) Description Household Survey: household characteristics; socioeconomic characteristics for members aged 15 and older; household food assistance benefits; usual monthly or weekly expenditures on food by type of outlet Diary Survey: Usual weekly food expenditures; price of purchased food by type and outlet Household panel members scan their food purchases from retail outlets; includes prices, quantities, promotion information, and demographics Sample Nationally representative; 7,500 consumer units per year in Household Survey (5 quarters of information); 7,500 households per year in Diary Survey (two 1-week diaries) Nationally representative; panel of 50,000 households Frequency and Timeliness Continuing since 1980; released annually with 1-year lag from collection, 2003 available Monthly data with 12-day lag between collection and release Response Rate (approximate) 78 percent, 2001 Household Survey; 75 percent, 2001 Diary Survey Not available NOTES: For more information, see Table 2-2 in Chapter 2 on the CE; see Chapter 3 on the Combined Outlet Consumer Panel, HOMESCAN Consumer Panel, and CREST; see Chapter 4 on other surveys that include limited data on food purchases.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Consumer Report on Eating Share Trends (CREST) HOMESCAN Consumer Panel NPD Group ACNielsen Prepared food purchases by individuals at commercial restaurants and other outlets, including fast food outlets; includes prices; identifies outlets Household panel members scan their food purchases from retail outlets; includes prices, quantities, and promotion information for items with UPC codes; item identification and weights for items lacking UPC codes; and demographics Online survey, weighted to be nationally representative; 3,000 adults and 500 teenagers daily Nationally representative, panel of 61,500 households (only one-quarter report both UPC and non-UPC purchases) 3-month lag between collection and release Monthly data with 3-week lag between collection and release 40 percent 85 percent

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Type of Survey Food Sales in Stores Survey Custom Store Tracking Scantrack Services Source Information Resources, Inc. (IRI) ACNielsen Description Point-of-sale data for food stores, food/drug combinations, and mass merchandisers Point-of-sale data for food stores, food/drug combinations, and mass merchandisers Sample Nationally representative; 32,000 retail outlets across the U.S. Nationally representative; 4,800 stores representing more than 800 retailers in 52 major markets Frequency and Timeliness Monthly data with 12-day lag between collection and release Monthly data with 10-day lag between collection and release Response Rate (approximate) Not applicable Not applicable NOTES: See Chapter 3 for discussion of Custom Store Tracking and Scantrack Services. nutrition research and policy: federal datasets on food and nutrition (Chapter 2); proprietary sources of food consumption and expenditure data (Chapter 3); and federal datasets that might provide some of the otherwise missing data (Chapter 4). These chapters also consider the limitations of the various data sources in addressing questions about food consumption patterns. Chapter 5 presents the panel’s recommendations. FOOD CONSUMPTION DECISIONS The question of “What shall we do for dinner tonight?” has probably occupied more time and thought and generated more tension than most people are willing to admit. The process of answering this question usually goes something like this: What are you hungry for? What food do we have

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies at home? Do we have the time or energy to cook? Will the kids eat it? Can we afford to go out? Will the grocery store or restaurant be crowded at this time of day and do we have enough time to stop there? What would be healthy to eat or do we care about that right now? It is clear that a number of factors go into making this daily decision. They include not only individual- and household-level factors (such as income, time resources, knowledge, skills, and preferences), but also factors outside of the household (prices and the availability of stores and restaurants). At perhaps a lower level of consciousness, the larger policy and media environment probably also play a role. For example, concern over mad cow disease may (or may not) trump a craving for the big juicy burger and fries just advertised on a television commercial. The purpose of considering the context of food decisions is to highlight the types of data that need to be collected in order to understand food consumption choices and to address the questions about food consumption that are listed above. In making decisions regarding food consumption, households and individuals consider their resource levels. These resources include monetary resources (income and asset levels), which are not always adequate for food consumption. Evidence from the current USDA measurement of food insecurity indicates that people in as many as 11 percent of U.S. households in 2003 worried about their ability to obtain adequate food for the family and were not always able to do so due to economic deficits (Nord, Andrews, and Carlson, 2004:3). Resources also include time—the amount of time available for food preparation and eating and for other activities. In 2003, Americans spent an average of 1.2 hours per day on food consumption and 0.5 hours per day on food preparation and cleanup (Bureau of Labor Statistics, 2004c:Table 1). These averages mask differences in time constraints for food preparation and consumption among different kinds of households, such as a family with two working adults compared to a family with one adult working outside and one inside the home or to a family with only one adult. Household members may also have a set of skills or informational resources available—for example, information on which foods are healthy and food preparation skills. Finally, individual household members have different food preferences (and allergies or aversions to some foods). There are other factors that contribute to food consumption decisions. The amount and types of foods that can be consumed given a household’s resources depend on the prices households face and the availability of different types of food (for example, the presence and types of grocery stores,

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies restaurants, and food retailers and the variety of foods they carry). Packaging may also affect food consumption decisions, as may labeling that identifies the ingredients and nutrient elements that individuals may or may not prefer to eat or may perceive as harmful. Technology for food production and preparation is also a factor for households, both for their own use and for the production of food away from home. Some theories about why low-income populations have a higher prevalence of being overweight or obese focus on the contextual factors. For instance, one argument is that the lack of major grocery stores and health food stores in low-income neighborhoods contributes to higher prices and lower availability of healthier foods, such as fruits and vegetables, while convenience stores and fast-food restaurants are plentiful in these neighborhoods. In other words, tasty, energy-dense, and low-micronutrient foods are readily available at low cost. These foods may be a logical choice if healthful foods are difficult to obtain and if a family must be fed with limited economic resources. However, research shows a more complex picture. For example, in a 1996 survey of low-income households, most people were able to shop at supermarkets within or close to their neighborhood, though one-third traveled more than 4 miles to shop for food, most often citing high prices and lack of stores in their neighborhood as reasons (Ohls et al., 1999:xiii-xiv; see also Cole, 1997). Changes in technology have also been examined as possible contributors to consumption trends (Lakdawalla and Philipson, 2002). Specifically, it is hypothesized that technological advances in food preparation have reduced time costs and increased the quantities and varieties of foods that are produced both in the home and by mass producers (Cutler et al., 2003). Some of these advances, however, have likely contributed to less healthy eating habits. Households also make decisions within a policy environment. Policies directly related to food consumption include food and nutrition assistance and education programs, most of which are under USDA, as well as food standards. Food assistance programs include the Food Stamp Program, the National School Lunch and Breakfast Programs, the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), and several smaller programs, including the Child and Adult Care Food Program, the Commodity Supplemental Food Program, the Emergency Food Assistance Program (TEFAP), the Food Distribution Program on Indian Reservations, Meals on Wheels for the elderly (a DHHS program), the Special Milk Program, the Summer Food Service Program, and the WIC Farmers’

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Market Nutrition Program. Food education programs include Food Stamp Nutrition Education, the Nutrition Education and Training (NET) Program, Team Nutrition for School Meals, and the nutrition component of the WIC Program. Food standards include sugar restrictions on cereals in the WIC Program and nutrition standards for school meals. (See also Institute of Medicine, 2004, on criteria for selecting WIC food packages.) Policies toward the marketing and labeling of food, such as nutrition labeling guidelines and guidelines for health claims of foods, may also contribute to consumption decisions. Public education campaigns for healthful eating, such as the Food Guide Pyramid, and on food safety may also have an impact on the foods that Americans eat and the methods they use to prepare them. In addition, medical education standards and health care delivery policies and practices regarding nutrition and diet may affect decisions that health care consumers make about eating. Finally, food safety policies, including federal and state regulations for food production and preparation, inspections, and surveillance of food-borne illness, affect the quality and variety of foods that reach the consumer. This overview makes it clear that a great deal of information is needed to fully understand food consumption decisions and their consequences for diet and health. In addition to data on what foods people eat and what they prefer to eat, information on household resources—income, assets, time, education, health and diet knowledge, and food preparation skills—is needed. Environmental-level information is also needed—that is, information on prices of food and related goods; availability of different foods; availability of grocery stores, food retailers, and restaurants; and marketing practices (such as amount of advertising exposure, target audiences, coupons or other incentives, packaging and display). Finally, information on policy interventions, such as food stamps and other public assistance programs, government-assisted marketing efforts (such as the Dairy Board, over which USDA has oversight), labeling regulations, and public health initiatives related to diet is also needed. FOOD SAFETY ISSUES A complete assessment of data needs to support food safety policy is beyond the scope of this panel’s report, but a few comments regarding the current state of data to support policy analysis are important for context.

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies The main point of interface between this panel’s charge and food safety policy is the use of food consumption data to support exposure assessment for food safety hazards. Dramatic changes in what food is eaten and how food is prepared during the last few decades have altered the type and incidence of food safety risks for U.S. consumers. For example, over the past 40 years, Americans markedly increased the percentage of their total food dollar that was spent on food eaten away from home (from 29 percent in 1963 to 47 percent in 2003), as well as the percentage of their dollar for food eaten away from home that was spent at fast-food outlets (from 26 percent in 1960 to 38 percent in 2000). Since 1970, annual consumption of red meat has dropped by 18 pounds, while annual consumption of chicken has increased by 37 pounds (see www.ers.usda.gov/Briefing [June 2005]). These kinds of changes mean that exposure assessment in the future—whether to food-borne pathogens or pesticide residues—will need to be based on the most current food consumption patterns and should include information about methods of preparation, including consumption of undercooked, raw, or unwashed foods, and whether prepared at home or obtained from a retail outlet. Ideally, such data should provide sufficient detail to distinguish populations at particular risk for some hazards, such as expectant mothers, young children, the elderly, and the immunocompromised. Food-Borne Pathogens It is now widely recognized that microbial food-borne pathogens are the most important food-borne hazard. Food-borne pathogens continue to evolve and adapt, with such new hazards emerging as Salmonella enteritidis in eggs, E. coli O157:H7 in ground beef, and the potential for BSE prions in beef (National Research Council, 2003). Exposure assessment is still in its infancy for these hazards. In food-borne illness outbreaks, the food source is often unidentified. Furthermore, there is imperfect understanding of the dose-response relationship for many pathogens, and attempts to estimate these relationships are not well developed. A risk assessment by the USDA Food Safety and Inspection Service (FSIS) for E. coli O157:H7 used data from multiple sources for risk characterization and exposure assessment. These sources included: reported illnesses and food sources from the 10 surveillance sites in the FoodNet system (located in California, Colorado, Connecticut, Georgia, Maryland,

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Minnesota, New Mexico, New York, Oregon, and Tennessee), which is run cooperatively by the Centers for Disease Control and Prevention, the FDA, and the FSIS; the incidence of the pathogen in the food supply at various points along the supply chain from data provided by the FSIS and the USDA Animal and Plant Health Inspection Service (APHIS); and intake data regarding ground beef consumption (U.S. Department of Agriculture, 2001). The ability to link data from different sources for such exposure assessments, however, is far from ideal. FoodNet monitors illness outbreaks in its 10 sites or catchment areas by obtaining information on the incidence of different food-borne pathogens at laboratories in these areas, but the laboratory data may not indicate the exact food product source. (FoodNet validates outbreak data by reviewing the methods and tests used by the FoodNet laboratories and surveying physicians in the catchment areas to determine under what circumstances they order samples to be analyzed by these laboratories.) The FSIS monitors food-borne pathogens in meat and poultry slaughter and processing plants, but it does not make these data public, except for periodic summaries. The APHIS monitors pathogen incidence in animals on farms, but not on a regular basis for all pathogens of potential human health interest. More importantly, there is no systematic monitoring of food-borne pathogens in the food supply at the retail and household level. The National Health and Nutrition Examination Survey (NHANES) obtains much relevant data for exposure assessment, including 2-day dietary recall begun in 2002. However, the survey does not currently obtain information on food preparation (for example, washing raw foods) and other topics thought to be most useful for this purpose (see Chapter 2). The FDA periodically carries out a random-digit dialing Food Safety Survey of 2,000-4,000 households to monitor perceptions of individual and societal risk related to food consumption, food-handling practices in home-prepared food, understanding and use of food product safety labels, food allergies, consumption of potentially risky foods, attitudes toward new food technologies, perception, knowledge, and experience of food-borne illness, and sources of food safety knowledge. However, the Food Safety Survey, which was fielded in 1988, 1993, 1998, and 2001, does not include data on actual food consumption, its data on household characteristics are limited, and its response rates are low (61 percent in 2001—see www.cfsan.fda.gov/~lrd/ab-foodb.html [June 2005]). FDA has conducted periodically since 1982 the Health and Diet

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies Survey, which addresses more broadly the health-related knowledge and attitudes of people aged 18 and over in households. Last conducted in 2004, it has been used to study people’s awareness of relationships between diet and risk for chronic disease, consumer use of food labels, weight loss practices, and the effectiveness of the National Cholesterol Education Program. Like the Food Safety Survey, the Health and Diet Survey does not collect data on food consumption or economic characteristics of respondents’ households, and its response rates are very low (41 percent in 2002—see www.cfsan.fda.gov/~1rd/ab-nutri.html [June 2005]). FoodNet includes a population survey component in which residents of its catchment areas are contacted and asked about recent diarrheal disease, treatment sought, and whether foods causing known outbreaks of food-borne illness have been consumed. However, the FoodNet population information does not include questions about food preparation and storage or residents’ knowledge of food safety issues, and information on household characteristics is limited (www.cdc.gov/foodnet/what_is.htm [June 2005]). The gaps in the available data mean that heroic assumptions are required to link together pathogen incidence data, intake data related to specific subpopulations, and food preparation data in order to carry out exposure assessments for one or more food-borne pathogens. Attention to filling these gaps to improve the validity and reliability of such linkages will be important for future food safety risk assessment and policy analysis. Pesticide Residues Residues from pesticides in food can also be a problem, although it is not clear which kinds of residues are currently the most problematic or may become so in the future. Exposure assessments for pesticides currently use a methodology developed by the EPA to carry out its mandate under the 1996 Food Quality Protection Act to periodically review tolerance limits and reregister pesticides for compliance with updated standards. The Act required EPA to give special consideration to children’s exposure and exposure to multiple residues with similar toxicity. EPA’s exposure assessment process first uses a conservative assumption of residues equal to maximum residue limits, such that data on actual residues are not required. When this initial assessment indicates a potential risk from a particular pesticide, EPA refines its assessment using realistic exposure data. Actual pesticide residue data are collected through the USDA’s Pesticide Data Program (PDP) and the FDA’s Total Diet Study. The PDP began in

OCR for page 7
Improving Data to Analyze Food and Nutrition Policies 1991. It operates in 10 states and is designed to capture information on actual residues in the food supply as close as possible to when food is actually eaten. The Total Diet Study began in 1961. It obtains samples of food purchased by FDA personnel in selected cities, which are then analyzed by FDA laboratories; the results are used to estimate exposures by weighting to food consumption patterns from the Continuing Survey of Food Intakes by Individuals (the CSFII food consumption data are now part of NHANES—see Chapter 2). The PDP and the Total Diet Study represent important sources of information about the incidence of residues in the food supply over time. They have been used, for example, to assess actual risks as a way of better understanding outcomes of pesticide regulation (Day et al., 1995; Kuchler et al., 1997). Evaluating Consumer Education One more point on food safety information is worth noting. Data on consumer food safety knowledge and practice would be crucial for developing any consumer education or labeling efforts aimed at safety. Questions in these areas could be part of the addition of a health knowledge component to the NHANES that ERS is developing and that we support (see Chapters 2 and 5). They could also be added to the FoodNet population survey component. Conversely, more detailed household characteristics, as well as some information on food consumption, could be useful to add to the FDA Food Safety and Health and Diet Surveys, perhaps asking questions of subsamples to reduce burden on respondents.