Click for next page ( 13


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 12
The Proposed System It is proposed that an ongoing system be developed and implemented by the federal agencies with major interests in food consumption and health. This system should encompass these components or subsystems: A continuous collection, processing, and review of food intake data from a stratified probability sample of the U.S. population. This is the core of the food consumption data. The collection of health status indicator data from currently available sources and their collation for population strata analogous to those from which the dietary information is derived. The ongoing examination of available aggregate data from commercial and governmental sources on market food disappearance from regional and economic strata comparable to the above. Collation, analysis, and reporting of information from these three sub- systems should be the responsibility of a lead agency in the federal govern- ment. This agency must have adequate personnel and fiscal capability to undertake special purpose analyses of the data bases generated by the above subsystems. There should be provision also for government and non- government investigators, after appropriate review of requests, to have ac- cess to the data base for meritorious projects deemed to be in the public interest. The lead agency must also have the human and fiscal sources necessary to initiate or promote field studies or other specific investigative projects in response to food-health problems or potential problems revealed by the ongoing monitoring of the data system. 12

OCR for page 12
The Proposed System The system as outlined above, and the data bases so generated, will be capable of: monitoring the nutritional adequacy of the food practices and food supply of the general population; monitoring general health trends and their potential relationship to food usage; identifying sectors of the population in which particular food intake- related risks may be high and about which particular concerns should exist or in which special in-depth studies should be undertaken; providing a data base for testing probable effects of fortification pro- grams, food additive regulations, nutritional standards or guidelines for food products, and similar federal needs; monitoring potential risk associated with selected food components or contaminants; and providing a sound data base for studies for the examination of relation- ships between food intake and chronic disease in segments of the U.S. population. FOOD INTAKE OF INDIVIDUALSSUBSYSTEM I It is recommended that this subsystem be established as an ongoing opera- tion, with rapid processing of data and with capability to collate and analyze data on a "moving average" basis. The establishment of data collection, processing, and analysis as an on- going operation will permit the responsible lead agency to develop and maintain a highly qualified unit operating on a continuous basis. The neces- sary experience in the various aspects of these operations can be highly developed and the unit fully employed on an ongoing basis. This will im- prove operational efficiency and eliminate administrative problems inherent in a system that is only periodically involved in such studies. For purposes of sampling design, it is suggested that the time frame for collection of a statistically adequate sample of the stratified U.S. population be 5 years. It is recommended that the design provide for the collection and processing of these data at the rate of 20 percent of the full sample per year. It is recommended that analysis and interpretation be updated each year on the basis of data accumulated for the preceding 5 years (a moving 5-year average). The annual analysis also will provide trend information for the 12-month period as an "early warning" of developing problems. This sys- tem of sample collection is portrayed graphically in Figure 2.

OCR for page 12
14 ASSESSING CHANGING FOOD CONSUMPTION PATTERNS YEAR Collection ~~ ~ ~ 2 ~ 3 ~ 4 ~ 5 ~ 6 ~ 7 ~ ~ ~ 9 ~ ~o] Analysis 1 1 - 5 2 -6 3 - 7 L 5-9 1 1 r 6- 10 _ . _ 1 7 - 1 1 _ 1 1 FIGURE 2 Time frame for sample collection. Measurement of Food Intake The data collection system should provide for replicated 1-day observations of the same individuals in sufficient number and stratification to provide an estimate of the distribution of usual intakes among the selected strata of the population. The recommended data collection technique is a 24-hour recall by trained interviewers using standardized instruments and food models to help respondents estimate quantity of food consumed and probe for precise information. This recall may be supplemented with a 3-day food record and a food frequency questionnaire to provide additional data concerning fre- quency of food use. The data collection and processing system must provide for internal quality control and should provide external validity checks as appropriate procedures are developed. Four replicated 24-hour recall observations on the same individual within the 1-year sampling period may be necessary to provide a measure of the usual food intake of that individual (and of the variation currently experi- enced) adequate for use in determining the usual food intake pattern of the population group. Research should be initiated to determine the measure- ment frequency required for stipulated levels of reliability of the estimate of usual food intakes of individuals in a population and for estimation of the extremes of food consumption patterns. The precision required will vary with each study. Sampling The stratification of sampling should be designed to permit statistically reliable estimates of risks (low or high intake in comparison to actual needs

OCR for page 12
The Proposed System TABLE 1 Stratification of Sampling 15 Age and Physiologic Population Socioeconomic Race and State Density Region Level Ethnicity 0-5 yr Metropolitan Northeast High White 6-10 yr Urban Northwest Middle Black 11-20 yr Periurban Southeast Low Span~sh- Male Rural Southwest Poverty American Female Other Pregnant Lactating 21-45 yr Male Female Pregnant Lactating 46-65 yr Male Female 66+ yr Male Female or tolerances), as well as group averages. The population may be stratified, for example, according to the characteristics listed in Table 1. It is recognized that it may not be feasible to design all of these and other desired features into a sampling frame and stratification design. It is also recognized that it will not be feasible to provide adequate numbers of indi- viduals for distribution analyses in all cells potentially generated by the above 26 traits (14 x 4 x 4 x 4 x 4 = 3,548) and that in data analyses it will be necessary to collapse cells. Therefore, it is recommended that an experienced design group be charged with the development of an optimal design, taking into account first the questions to be put to the data and second the cost and logistical considerations. An auxiliary consideration in the sampling and analysis design should be the feasibility of collating health status indicator data for analogous subpopulations since it is the intent of the overall system that data from different sources be linked at the level of the subpopulation group and not at the level of individuals. Data Base Requirements It will be necessary to charge specific individuals or organizations with the development of requisite data bases for the proposed analysis system from existing information. These data bases will include distribution of nutrient

OCR for page 12
16 ASSESSING CHANGING FOOD CONSUMPTION PATTERNS requirements, distribution of food component tolerances, and food compos- ition data including both nutrients and nonnutrients. A cooperative effort by government, academia, and industry will be needed. It is to be expected that over time the current data bases will improve in quality. The food consump- tion data base must be maintained in a format that permits recalculation of the intakes and reassessment of risks as other data bases improve or other needs for the data arise. Coding and Data Analysis The descriptors of consumed food that are entered and retained in the data base files must be sufficiently precise to allow answers for questions that may be posed at a later time. The descriptors are necessary even though information about food composition may not yet be available for all foods that can be precisely identified. While it would be desirable to identify manufacturers in food descriptions, such reporting would greatly expand the detail of the food coding system and the size of the data base storage requirement and, as a result, might reduce the reliability of data by increas- ing the chance of error. It is recommended that, as the system is developed, a special advisory group be convened to address the questions of precision of food identification and coding in the collection and storage of current and future food consumption data. These decisions must be made with an awareness of priority questions that are and will be put to the data base. It seems probable that certain questions will not be answerable from the data base. For example, it may be difficult to generate reliable estimates of the distribution of intakes of infrequently used foods or of intakes of par- ticular brands of foods. In these cases, the agency may choose to undertake specific consumption or frequency studies either as additions to dietary interviews proposed or as separate studies. The agency, on the other hand, may turn to other agency or commercial sources of data when these sources meet the requirements. HEALTH STATUS INDICATOR DATA- SUBSYSTEM II The stated objective of this system is to yield information that may be valuable in maintaining or improving the state of health, including nutri- tional status, of the U.S. population. Therefore, it is necessary to relate food consumption patterns to nutritional and health status. However, knowledge of the relationships between food consumption, nutritional status, and gen- eral health status is limited, and many areas are controversial. Careful examination of the factors to be studied and the data to be collected is

OCR for page 12
The Proposed System 17 required to ensure relevance to the issues under examination and to permit appropriate interpretation. Existing health data bases can be used in conjunction with food intake data. While it is not necessary to collect both types of information from the same individuals, the collation, transfer, and summarization of health data according to population characteristics that can be duplicated in food con- sumption studies will be necessary. The coordinated studies will focus initially on identified or postulated food-related health problems and will expand as relationships become clear. The health indicator data base should include information relevant to these acknowledged and possible relationships; however, there are practical limi- tations on the extent to which health conditions can or should be identified and retained in the data base of this subsystem. Nutritional status is one component of health. From the proposed system, the usual food intake of individuals will be known, and this information can be used to estimate the risk of nutritional inadequacy (or excess) but not to identify the nutritional state or health of specific individuals. While predic- tion of the nutritional status of a population group can be assisted by limited use of anthropometric and biochemical measures, it is not recommended that this proposed monitoring system include these costly measurements for the total population sample. Rather, the collection of anthropometric data (such as height for weight, height for age, or height and weight for age) and biochemical measures (such as hemoglobin or hematocrit) should be limited to a population subsample. A detailed dietary and/or health history of indi- viduals contacted for dietary recall information may, on occasion, be desir- able. More detailed or extensive evaluations may be indicated if pilot studies identify prevalent problems in population segments. In order to increase the potential for identifying relationships between food consumption and health status, it is recommended that available statis- tics be examined to identify population segments whose health indicators appear to be abnormal relative to the population average and to averages of other subgroups. Special studies of food consumption should be carried out in these population segments and/or an appropriate sampling of this popula- tion subgroup should be included in a regular monitoring program that combines information on diet and health. Reliable statistics are available from many sources, but the data are not collated in the manner necessary to identify potential groups for study. It is recommended that a task force consider the existing sources of health infor- mation and the optimal approach to the identification and transfer of these data for use in the proposed system. The relationship of the health data to income status of population groups should also be evaluated to determine if,

OCR for page 12
18 ASSESSING CHANGING FOOD CONSUMPTION PATTERNS at least for some indicators, sampling based on economic stratification may yield the desired information. Identification of population segments to be studied should be made with a recognition of, first, the priority questions likely to be put to the composite data base, and second, the logistics and cost. AGGREGATE FOOD DATA SUBSYSTEM III A number of sources of aggregate data on food disappearance and food purchase now exist within the government and the private sector. Some data are available at a national level only. Others provide data grouped according to regional or consumer characteristics. Sources in the private sector com- monly provide information about specific commodities and specific brands and are designed to be rapidly responsive to market trends. Comparison of aggregate food disappearance or food purchase data with the composite of food intake data over the time period should provide a basis for predicting relationships between the data bases. With this information, the aggregate data bases may prove to be very useful as an early warning component of the proposed system by detecting changes in foods reaching the consumer. They may be of additional use for special purpose studies, such as tracking specific commodities used in commercial products. It is recommended that uses of aggregate data bases to supplement information needed for a total overall monitoring system be identified. IMPLEMENTATION OF THE TOTAL SYSTEM Many parts of the proposed system, i.e., data bases and collection systems, already exist within various agencies of government. However, the parts have not been coordinated or applied as effectively as might be possible. It is recommended that interagency discussions be undertaken with the objective of developing a coordinated system of the type described herein. Primary implementation responsibility should be assigned to one lead agency or to an interagency institution. This responsibility should relate not only to the coordination and analysis of data, but also to control of the quality of sampling, data collection, and data processing. The final conclu- sions can be no stronger than the weakest link between the unit that supplies information and the final processing of data to be examined.