5
Recommendations

The need for the U.S. Department of Agriculture (USDA), the U.S Department of Health and Human Services (DHHS), and other agencies to address continuing and emerging policy issues related to food consumption has created new challenges for the available data systems. The ever more pressing need for timely data to help guide policy making places a heavy burden on survey data collections sponsored by the federal government, which are not generally geared toward producing data in a short time frame. In the face of increasing obesity in the United States, there are new calls to understand the economic and social factors behind food consumption and nutrition, creating a need to link data on food consumption to data from various sources on food prices, time use, financial resources, food assistance program participation, availability of food outlets and foods, and other potentially relevant factors. Over the past few decades, changes in the kinds of foods that are consumed and how they are prepared have also posed new challenges for food safety. These changes have increased the need for data on the kinds of foods eaten by specific population groups, such as young children, the elderly, and expectant mothers; the extent to which raw and undercooked foods are consumed; and the extent to which potentially hazardous food additives have been used. Also of concern is the extent of consumer knowledge.

In this chapter, we offer recommendations for improvements in the existing data systems. In formulating recommendations, the panel was asked to focus on improvements that could be made on the margins, within the



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 76
Improving Data to Analyze Food and Nutrition Policies 5 Recommendations The need for the U.S. Department of Agriculture (USDA), the U.S Department of Health and Human Services (DHHS), and other agencies to address continuing and emerging policy issues related to food consumption has created new challenges for the available data systems. The ever more pressing need for timely data to help guide policy making places a heavy burden on survey data collections sponsored by the federal government, which are not generally geared toward producing data in a short time frame. In the face of increasing obesity in the United States, there are new calls to understand the economic and social factors behind food consumption and nutrition, creating a need to link data on food consumption to data from various sources on food prices, time use, financial resources, food assistance program participation, availability of food outlets and foods, and other potentially relevant factors. Over the past few decades, changes in the kinds of foods that are consumed and how they are prepared have also posed new challenges for food safety. These changes have increased the need for data on the kinds of foods eaten by specific population groups, such as young children, the elderly, and expectant mothers; the extent to which raw and undercooked foods are consumed; and the extent to which potentially hazardous food additives have been used. Also of concern is the extent of consumer knowledge. In this chapter, we offer recommendations for improvements in the existing data systems. In formulating recommendations, the panel was asked to focus on improvements that could be made on the margins, within the

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies existing data infrastructure, rather than considering major new data collection efforts. A broader consideration of new data collections would require a much more in-depth study. INTERAGENCY WORKING GROUP ON FOOD AND NUTRITION DATA A number of different agencies rely on high-quality data on food consumption, diet, and health. Two departments alone—USDA and DHHS—each have multiple agencies that use these data for different purposes. For some purposes, data on medical and nutritional outcomes and covariates are needed. For others, data about diet and health knowledge and food preparation practices are needed. And for still others, data on prices, expenditures, and financial and other resources are needed. Moreover, for each type of data, there is increasing need for more detail and the ability to link different data sources. With the merging of the National Health and Nutrition Examination Survey (NHANES) and the Continuing Survey of Food Intakes by Individuals (CSFII) in 2002, NHANES now becomes the only large nationally representative dataset of the federal government that collects detailed information on food consumption. This merger has resulted in new efficiencies in data collection, but it has also placed increased demands on NHANES—to fill both the role that CSFII formerly filled and the primary role of NHANES to assess health and nutritional status. The process of merging these two data collection programs is still a work in progress. The National Center for Health Statistics (NCHS), the agency in DHHS that conducts the survey, has a good track record of working with other agencies to address data collection needs. But the NHANES will not be able to address all the data needs for all the agencies. Thus, additions and modifications to other related surveys and data collection efforts will need to be considered. In this regard, we are impressed by the initiative shown by the Economic Research Service (ERS) in USDA, not only in funding the development of the Flexible Consumer Behavior Survey Module to include in NHANES and a food and eating module to include in the American Time Use Survey, but also in undertaking research with scanner data and other initiatives to make the best use of and to enhance the available data infrastructure for food and nutrition-related policy planning and analysis.

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies To build most effectively on the ERS initiative and to take advantage of the current government-wide impetus to deepen understanding of food consumption behavior, its correlates, and its effects on health, safety, and other aspects of society, we recommend an interagency working group on the nation’s food and nutrition data infrastructure. In this, we echo and enlarge on a similar recommendation in Dwyer et al. (2003), which reports on a workshop discussion of the integrated NHANES-CSFII. The proposed interagency group would review the development and collection of new information and make recommendations for design decisions for NHANES and other data sources related to food consumption. It would strive to fill gaps in an effective manner and to reduce unneeded overlaps in data collection. For example, the Food and Drug Administration (FDA) periodically conducts telephone surveys that overlap in content with the kinds of questions on diet and health knowledge and food preparation practices that ERS and NCHS plan to include in the new Flexible Consumer Behavior Survey Module. Perhaps the FCBSM could serve the needs of FDA, or perhaps key question content could be made the same between the FCBSM and relevant FDA surveys so that cross-survey comparisons and validation would be possible. The proposed interagency group should consider how to develop a complete review of the analytical work that has already been done on assessing the effects of food and nutrition programs in order to identify the information needed to address unanswered questions. It should also consider the kinds of testing and validation that should be built into data collection programs to ensure high-quality information while minimizing respondent burden. It should seek as well to facilitate special arrangements for access to linked datasets that cannot be provided in public-use form. We believe the group could usefully be led by the Office of Management and Budget Statistical and Science Policy Office, which has a coordinating role in the federal statistical system. Alternatively, it could be co-led by an agency of the USDA working with an agency of DHHS. The group would include representatives from the various agencies in both DHHS and USDA that have policy and data collection responsibilities related to food and nutrition and also from other federal agencies with related policy responsibilities, such as the Environmental Protection Agency. We understand that interagency working groups are often difficult to make effective because the member agencies have different missions, operations, and cultures. Yet such groups can give visibility to an area, such as food and nutrition policy research, in which coordination and integra-

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies tion of data collection and analysis is needed. Such groups can also provide a venue for systematically considering different perspectives and approaches to collecting quality information in the most efficient and least burdensome manner (see National Research Council, 2005c:12, 44-48). Recommendation 1: An interagency working group, led by the Office of Management and Budget, or co-led by an agency of the Department of Agriculture and the Department of Health and Human Services, should be established and take responsibility for the systematic development and use of diet and food consumption data to address policy and research questions of the federal government. RESEARCH AND DEVELOPMENT To make the proposed interagency working group on diet and food consumption data more effective, we recommend that the group clearly assign lead agency responsibilities for ongoing, sustained research and development programs on data in key areas to inform the group and build a strong base of scientific evidence for its work. Agency research programs should address cost-effective ways to develop high-quality data to remedy data gaps and weaknesses. For example, ERS could usefully have lead agency responsibility for research to develop high-quality, relevant data to understand the economics of food consumption, factors that affect shopping practices, diet and health knowledge, and related consumer behaviors, and how food-related behaviors affect food consumption and socioeconomic well-being. Such a program should include: assessments of the validity and reliability of alternative datasets; research on linkages of relevant survey data with relevant administrative records, neighborhood characteristics, and retail and household scanner data; and the development of protocols for design and testing of new survey content. Similarly, an agency in DHHS could usefully have lead responsibility for research and development on improved data for monitoring and understanding food fortification or food safety issues. Recommendation 2: The proposed interagency working group should assign clear responsibilities to lead agencies for sustained programs of research and development on data in key areas to provide a sound base of scientific evidence for the group’s work to improve the available information on diet and food consumption.

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies ENHANCING FOOD AND NUTRITION-RELATED DATA IN NHANES One of the benefits of previously having two surveys that collected extensive data on food consumption was that each survey could have a different focus. The merged NHANES has not thus far been able to collect all the information that was also on the CSFII, such as information on food expenditures and the information covered by the Diet and Health Knowledge Survey (which was part of the CSFII). Other information, such as data on the relationship between marketing practices, prices, and expenditures for and consumption of food, is also needed. An important task for the proposed interagency group would be to develop priorities and recommend cost-effective methods for adding food and nutrition-related questions to NHANES. The group could consider alternative methods or designs to obtain additional information—for example, rotating modules or administering one module to half of the survey sample while the other half receives a different module. The group’s work would be informed by the ERS research and development program we recommend, which could include the use of small-scale experiments and other methods for testing and validating new survey content in NHANES. The Flexible Consumer Behavior Survey Module (FCBSM) currently under development by ERS and NCHS will significantly enhance the ability of NHANES to support a wide range of food and nutrition-related research. The module is planned to include questions on food shopping, food expenditures, self-assessment of diet quality, frequency of eating food away from home, attitudes toward and knowledge about diet and food safety, use of food labels, and safety-related food preparation practices. We applaud this effort and urge that it go forward without awaiting the appointment of the interagency working group we recommend. Once appointed, such a group should give the highest priority to reviewing the research and development of the FCBSM and how it can serve the variety of needs for data for food and nutrition-related research and policy analysis. We note that an important need for analyses of food consumption and shopping and meal preparation behavior is data on actual food prices. This information would likely be difficult to obtain in a supplemental module to NHANES, although it may be possible to ask respondents to scan in some of their purchases; this idea could be tested. It may also be possible to develop linkages with other sources of price information.

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies Recommendation 3: The proposed interagency working group on diet and food consumption data should consider priorities and methods for obtaining additional food and nutrition-related information in the National Health and Nutrition Examination Survey. The development of the NHANES Flexible Consumer Behavior Survey Module, which will include questions on food expenditures, diet and health knowledge, and other food and nutrition-related topics, should proceed, and research should be conducted on ways to obtain food price information for inclusion in NHANES. DATA LINKAGES Our review of existing data sources on food consumption and related information indicates clearly that no single source can satisfy the full range of data needs. Moreover, it does not appear feasible, even if resources were available, to develop a single all-purpose survey on these topics at the level of detail required for many analyses. In addition to the technical and logistical difficulties, the burden on respondents would be too great. A way to provide needed information at low cost and burden is to look for ways to link data from program administrative records, other surveys, and on-line resources to the NHANES, the Consumer Expenditure Survey, or one of the other datasets (briefly reviewed in Chapter 4) that already contain some relevant food and nutrition-related data. Linkages at the individual level with food assistance program records would provide valuable information on program participation and benefits and associated behavioral effects. Linkages with such sources as area price indexes, census information, and various geographic databases could add metropolitan and neighborhood characteristics that would be helpful for contextual analysis of food consumption behavior. Assessment of the costs, benefits, and methodology for data linkages should be an important component of research and development by ERS in cooperation with other relevant agencies. Both individual and neighborhood linkages require the ability to geocode the addresses of sample households to small areas. Such linkages also require the implementation of efficient, low-cost means of access to the resulting datasets in ways that protect confidentiality, such as through a research data center, remote on-line access, or a licensing arrangement that permits researchers to use confidential data at their institution.

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies Recommendation 4: The proposed interagency working group on diet and food consumption data should consider low-cost ways to enhance the analytic uses of NHANES and other surveys by linkages with food assistance program records and with sources of socioeconomic and food shopping characteristics for the areas in which survey respondents live. A priority should be to work out effective ways to provide access to linked datasets through restricted access mechanisms, such as monitored remote on-line access. USE OF SCANNER DATA Scanner datasets from retail stores and from household panel scanner surveys include very detailed information on the purchase of specific foods, brands, quantities, and the prices paid. Because they are produced for firms interested in the latest market trends, they are usually available within a matter of weeks or months. This kind of detail and the timeliness with which the data are produced are unmatched in federally sponsored surveys. Because some policy and decision-making questions require unusually detailed information or must be made with the most timely data, the scanner datasets represent an attractive opportunity for USDA. In particular, retail scanner data have the potential for understanding short- and long-term trends in aggregate market purchasing of foods. Both these data and the household panel scanner datasets can be used to understand how consumer behavior changes (on both a very short-term and long-term basis) with respect to specific product attributes and price, when new products are introduced, or when labeling regulations change. Proprietary data on household food consumption also have the potential to identify trends in eating patterns and dieting practices. ERS has been exploring the use of these proprietary datasets for several years. It has contracted with ACNielsen for data from the HOMESCAN panel and with the NPD Group for the National Eating Trends data. However, the quality of these data is largely untested. Data quality concerns include the representativeness of the samples of the U.S. population overall and of certain groups, such as low-income, single-adult, and minority households. For scanner data, accuracy of coding and the accuracy of the product scanning process are also of concern. Moreover, the retail scanner data do not cover all retailers, food that does not have a universal product code (UPC), or food purchased in restaurants.

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies It would be beneficial for ERS to further explore the use of these data. Because scanner data are proprietary, it is not possible to make them readily available at low cost to all agencies and researchers that could benefit from using them. ERS could, however, explore with market research firms ways to obtain older scanner datasets that are of less value for the firms’ private-sector clients at a favorable price for redistribution to other federal users. For example, studies that compare food purchases from these datasets with data on food purchases from the Consumer Expenditure Survey (CE) and with food consumption data from the NHANES dietary intake survey could be conducted. Studies on the quality of the data obtained through the scanners could also be conducted. Ways to link scanner data at the neighborhood level with the NHANES or the Consumer Expenditure Survey could be explored as well. As a first step in this area, ERS should consider holding a conference of policy analysts and researchers who have used scanner and related proprietary data to see what has changed since its 2003 conference (see Chapter 3). The conference should address funding and research priorities in three areas: research on data quality (what is known and what is needed); accessibility of the data and how cost and access barriers can be reduced; and what research and policy questions can benefit most from scanner data. Recommendation 5: The Economic Research Service of the U. S. Department of Agriculture should continue to explore the use of data on food purchases, prices, and consumption from proprietary retail scanner systems, household scanner surveys, and household consumption surveys. This work should include a program to examine the quality of the data, consideration of ways to reduce the costs of access, and the determination of priority applications for the information. USE OF OTHER DATASETS Our review has focused principally on the major surveys for food consumption analysis, including the National Health and Nutrition Examination Survey, the Continuing Survey of Food Intakes by Individuals (including the Diet and Health Knowledge Survey supplement to the CSFII), the Consumer Expenditure Survey, and scanner datasets and household food consumption data collected by market research firms. There are many other federal datasets that, while they primarily serve other purposes, include some relevant information and could be useful for food and nutri-

OCR for page 76
Improving Data to Analyze Food and Nutrition Policies tion-related policy analysis and research with modest enhancements. We urge the proposed interagency working group, informed by research and development by the Economic Research Service and other relevant agencies, to consider low-cost ways to exploit surveys such as the Current Population Survey, the American Time Use Survey, panel surveys of specific age groups or the low-income population, and surveys that are designed for the addition of modules to track emerging trends. Recommendation 6: The proposed interagency working group on diet and food consumption data should consider ways to enhance the usefulness of other federal datasets for food and nutrition-related policy analysis and research. Such datasets include the Current Population Survey, the American Time Use Survey, panel surveys that follow families, children, and the elderly over time, and surveys that are designed to include modules to track emerging trends. CONCLUSION This report has reviewed the kinds of information and data needed to more fully understand decisions that the population makes on food consumption and to guide policy makers. We believe the implementation of our recommendations and consideration of the suggestions we make throughout the report will improve the underlying knowledge base for food and nutrition-related policy planning in the United States.