Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
L - 5 CONJOINT DATA USE: A COORDINATING COMMITTEE PERSPECTIVE BACKGROUND In March 19 78, the U. S . Depar tment 0 f Agr icul Lure (USDA) and the U. S . Department of Health and Human Services (DHHS ~ submitted a proposal to the Congressl5 for the repeated collection and systematic reporting of information on the food, nutrition, ant health status of the U.S. popu- lation. At the core of this proposed system, called the National Nutrition Monitoring System (NNMS), are the USDA food consumption surveys (the NFCS, Nationwide Foot Consumption Surveys, with its Household Food Use and Individual Dietary Intake components) and the DHHS health status surveys (in particular, the NHANES, National Health and Nutrition Examination Survey, made up of nutritional and health status components). COMMITTEE PERSPE CT IVE A nationwide system for monitoring food consumption and nutrition, their determinants, and their impacts must include estimates of food acquis ition and consumption, as well as nutrient and health conditions . These estimates are obtained from a variety of national surveys, including the core surveys (NFCS and NHANES) and such others as the Department of Labor Consumer Expenditure Survey, and non-HANES health surveys. Figure 5-1 illustrates these interrelationships. Because national food and nutrition concerns cannot be dealt with solely by the large-scale core surveys, nationwide food and nutrition monitoring must also include information obtained from more specific surveys -- such as the Hispanic HANES, a nationwide health and nutrition survey of Hispanic populations. The scope of this report is limited to recommendations for only a part of this nationwide food and nutrition monitoring: the food consumption and nutrient intake aspects of two national surveys, NFCS and NHANES . Data collection and reporting methods must meet the existing and future needs of users of each of the separate core survey data bases, although it is equally important that the data bases feeding information into the larger monitoring system meet monitoring system needs, as well. The purposes of each of the separate core surveys provide perspective on its contributions to the overall purposes of a monitoring system. The relation of these separate core survey data bases helps to characterize the monitoring system and, more importantly, to define its limitations.
- 54 - PURPOSES OF CORE SURVEYS The NFCS provides data needed to estimate amounts of foods and nutrients in U.S. diets, as well as data useful for es timating dietary adequacy. These estimates are reported both as aggregated data on average comparisons and as nonaggregated data that describe intake distributions. Survey repetition provides data that permit estimates, within limits, of changes in foods and nutrients in U.S. diets, as well as measurements of changes useful for estimating trends in dietary adequacy. The NFCS obtains information on both the kinds and the quantities of foods. From this, estimates can be made of the nutrients in diets ingested by survey subjects -- adults and children who are stratified within the data bases by sex, age, and other descriptors. In addition, these surveys obtain information on various household characteristics, including factors that can contribute to assessment of the socioeconomic and demographic factors associated with food consumption and dietary status. An essential and unique feature of the NFCS that must be preserved and improved with high priority is the collection of data representing multiple days of food intake by individual survey subjects. The Com- mittee recognizes that it is this feature that, coupled wi th appropriate statistical techniques, permits examination of the distributions of intake among individuals. Information from either survey , sub ject to adequate sample size, can be used to estimate average intake of defined groups. Only the NFCS data can yield the distribution analyses needed to estimate nutrient adequacy or food component status (see pp. 96-98~. In comparison, the NHANES provides data needed to estimate the preva- lence of physical states or disease conditions in the U.S. population. Normative or descriptive data, such as data on weight and stature, are collected. The NHANES also collects biochemical data that serve as descriptors of biochemical and, perhaps indirectly, of phys iologic s tatus . These biochemical and anthropometric measures can provide data needed to estimate nutritional status of the survey population. Successive surveys permit the repeated measurements needed to estimate, within limits, changes over time in health and in the biochemical and anthropometric aspects of nutritional status linked to specific nutrient intake. PURPOSES OF A NATIONWIDE MONITORING SYSTEM In the view of the Committee, the purpose of a nationwide food and nutrition monitoring system is to provide repeated collection and effective dissemination of usable data on and descriptors of the food,
55 dietary, nutritional, ant biochemical factors that contribute to the understanding of th~nutritional status of the population. The terms in this Statement of purpose can be described as follows: · Repeated data collection. Similarity of measurements over time promotes the capability of analyzing trends an well as conducting retro- spective studies. · Effective data dissemination. Both the diverse uses of the core survey data and the needs of a monitoring system require the organized and systematic management and release of data. · Usable data. The use of varied data bases requires compati- bility and comparability at sufficient data points to ensure necessary linkages between data bases. o Food and dietary factors. The system requires information not only about food and nutrition themselves, but also about socioeconomic and demo- graphic factors that influence the food, dietary, and nutritional status of the population and its subgroups. · Nutritional status. Nutritional status is a health-related judgment based on measurements that directly or inferentially reflect the dynamic pro- cesses of food ingestion and digestion, the transport, absorption, and metab- olism of food components and their products, and the excretion of non-used food components and metabol ic products . In contrast with the use of per capita disappearance data, the advantage of the conjoint use of NFCS and NHANES data bases for nationwide food and nutrition monitoring is that it allows information from separate data bases to be linked at the level of nonaggregated information. This permits data on the dis tributions of food or nutrient intake among individuals within one popu- lation group (e.g., identified by sex, age, and income) to be linked to other data on similar persons in the population. USES OF A MONITORING SYSTEM A system that provides regular information about the population's food and nutrient consumption with health information provides the fundamental data needed to: · Describe the adequacy and safety of the U.S. diet and determine whether there are any "problems.'" · Describe the general nutritional health status of the population and determine if there are any '"problems."
- S6 - · Identify the population subgroups in which problems exist. · Aid in developing strategies for resolving these problems. · Monitor trends in nutritional status and dietary intake to determine whether the observed prob lems are changing. Conjoint use of data from the separate core surveys provides information of the national scope needed to describe the status of food consumption, nutrient intake, and nutritional health. These core survey date help to identify national food-, diet-, or nutrient-related public concerns and their trends along the continuum from inadequate intake to excess consumption of specific nutrients. A national data collection system (involving data from both the NFCS and NINES and other data) provides the information necessary to assess the food and nutrient-related risks to health that may be amenable to intervention. Moreover, the system can provide information to assist in selecting among possible options for intervention, such as the following: · The design of food programs -- food stamp programs, supplemental food programs, food fortification, or food guidance. · The targeting of programs -- to population groups by sex' age, income, or dietary pattern. · Determination of the education content of programs -- based on the kinds of excesses or inadequacies and their association with food and nutrient intake patterns. · Regulation -- setting of tolerances for substances in food or approvals for addition of substances to food. DATA NEEDS OF A MONITORING SYSTEM An effective system for monitoring food and nutrition concerns will involve other data bases, as well. For example, socioeconomic data of the type and extent gathered by the Department of Labor and reported as the Consumer Expenditure Survey can contribute needed perspective on questions on the socioeconomic determinants of food patterns. Data on the composition of . foods constitutes a separate data base that contributes fundamental infor- mation for all other uses of the food consumption and dietary intake survey data. The DHHS, through its National Center for Health Statistics collects separate data base information on occurrence of diseases (morbidity), their prevalence, and death from disease, as well as other information that contributes to the needs of a national monitoring system. However, none of these contributing data bases can substitute for the information obtained from the core survey (NFCS and NHANES) data bases. As Figure 5-1 illustrates, each
of the other data bases can contribute data needed to meet the goals of a national monitoring~9yatem. Yet, without the information currently obtained from the NFCS and NHANES data bases, a monitoring sys tem cannot exist . Given the Scope of the public concerns to be addressed through nationwide food and nutrition monitoring, neither the NFCS nor the NHANES can contribute all the needed information. For some concerns, neither the NFCS nor the NHANES alone can provide the needed information; clearly, data from both surveys will be needed. Figure 5-2 illustrates this point, us ing dietary observations an an initial point for assessing food and nutrition concerns involving separate survey data bases. In like manner, Figure 5-3 illustrates the concept of conjoint use with biochemical observations as the initial point for assessing the food- and nutrition-related concerns. It should be noted that in either case (assessment beginning with dietary or biochemical observation) data from both surveys will be needed eventually. RATIONALE FOR NOT RECOMMENDING A SINGLE DIETARY SURVEY Many have suggested that it is redundant to collect national dietary intake data through two separate surveys -- NFCS and NHANES. The Committee considered this matter. One possibility would be a single dietary survey. There would be three ways to attain such a single survey: eliminate dietary data collection from one of the existing surveys, merge data collection activities of the two surveys into one, or design a new survey after elim- inating one or another (or both) of the two existing surveys. Each of these approaches raises serious problems . ELIMINATE DIETARY INTAKE DATA COLLECTION FROM ONE OF THE EXISTING SURVEYS Eliminating dietary data from the NFCS is tantamount to eliminating this survey ant losing the uses that require information collected only in the NFCS, such as extensive socioeconomic data, food purchasing data, data on seasonality, and replications . E1 iminating dietary data from the NHANES would reduce interview time by about 15%, but would not diminish NHANES cos ts proportionately, because NHANES fixes costs are high. This elimination would not be desirable, because some current and important emerging uses for the entire NHANES data base -- including use of the biochemical, clinical, and anthropometric data (which were not a subject of this Committee's mandate) -- depend on the dietary information in the NHANES. Furthermore, there are other important uses that require linkages between the detailed socioeconomic data collected through the Household Food Use and Individual Dietary Intake components of the NFCS and the physiologic and health measures of the NHANES. These linkages can be established either through personal descriptors or through the dietary data, themselves. Dietary data are required for both these linkages:
58 · Dietary data are needed from the NHANES to confirm that groups that are defines by their personal descriptors (age, sex, income, etc. ) and that are apparently similar in their personal characteristics are indeed similar in their dietary intake before interpretive conclus ions can be drawn across the two surveys. This use links the survey data through personal descriptors. · Another important link is through the dietary information itsel f, in that groups with similar dietary intake are examined for their socioeconomic characteristics in the NFCS and their health and nutritional characteristics in the NHANES. This requires the same dietary information in the NHANES and the NFCS. MERGE THE DATA COLLECTION ACTIVITIES OF THE TWO SURVEYS INTO ONE The NFCS is a less clustered survey sample, because it is an in-home process; the NHANES, on the other hand, must be clustered around its clinical examination units. Thus, NHANES data collection cannot be added to the NFCS. If NFCS data collection is imposed on the NHANES, if would substitute the NFCS dietary interview, adding 2 hours of dietary data collection to the present NHANES half-hour interview and thus make the NHANES a 5-hour process. This substantially greater respondent burden would be likely to affect response rates adversely. NHANES examination response rates are adversely affected by bad weather. Therefore, the NHANES moves north and south with the seasons. This would preclude the NFCS monitoring of seasonal patterns, because season and geo- graphic area are inextricably confounded in the NHANES. Because of the greater clustering in the NHANES than in the NFCS, the merged survey would have to interview more (probably at least twice as many3 people than does the present NFCS to obtain the present precis ion of NFCS mean estimates. Thus, a single, merged survey would probably cost more than the two separate surveys. The exact increase in costs cannot be estimated at present, because the necessary calculations cannot be made without NFCS cluster identifiers, which are not now recorded in NFCS data. The Committee recommends that these cluster identifiers be retained in the future and that the appropriate cost analyses be done. DESIGN A NEW (SINGLE) SURVEY From the above discussions, it is clear that the surveys, because they have different purposes and uses, cannot be merged into a single entity with- out the risk of compromising some of the fundamental purposes of one or the other. In addition, it is unlikely that any single survey can be designed to
59 satis fy all the needs for survey tata on foot consumpt ion and dietary intake without being much more expensive than the present surveys. These considerations and the information obtained from the users' con- ferences let the Committee to believe that a more appropriate course was to enumerate a series of technical recommendations to improve comparability and compatibility between the separate surveys at currently existing points of potential linkage. Such a course could meet the users' highest-priority needs for more timely information and better information about who is eating what. While at the same time accepting the realities of continuing independent but mutually compatible surveys, the Committee recognizes the merit of con- tinuing consideration of the desired features of a composite single survey. To this end, the Committee developed a systematic approach to examining issues of definition of data needed and of requisite precision and reliability of data (Appendix A). The approach was used in the workshop. As a result, certain issues became apparent with regard to most purposes (e.g., documen- tation, timeliness) and other issues were more use-specific (e.g., required quality of measures and precis ion of estimates). The Committee found its approach (Appendix A) useful and urges that it be made an on-going process in the continuing review of surveys, their purposes, and their design. The Committee does not preclude the possibility that, at some future time and with explicit statement of priority uses, a single survey could be designed and implemented. The Committee notes only that this should not be expected to serve the present scope of uses reported by participants in the workshop and symposium. NEEDS OF DATA USERS As indicated in Table 4-2, some needs for data are common to all data users. Thus, in the view of the Committee, recommendations for modifications to the existing surveys that increase the utility of data developed from these fundamental questions would have high priority. For maximal effectiveness, the data should fulfill the needs of both the core surveys and nationwide food and nutrition monitoring. Each of the questions shown in Table 4-2 identifies data collection needs and implicitly, therefore, aspects of survey use worthy of review. TheThe data uses and data needs common to most data users, to the separate core . surveys and to nationwide food and nutrition monitoring were the basis for the Committee's assessment of the highest priority options for modifying the existing surveys. In the view of the Committee, an improved data collection system is needed to ensure the effective use of the separate survey data bases that are needed for a system of nationwide food and nutrition monitoring. However, it is equally important to bear in mind that each of the separate core surveys has
t 60 - uses that are independent of its contributions of data for nationwide food and nutrition monitoring. Essential independent survey purposes and uses for the data from the separate core surveys must not be compromised by modifications intended solely to increase nationwide food and nutrition monitoring capabilities. Rather, des ign implications that evolve from common data uses ant needs implicit in the priority questions of who is eating what form the basis for Committee recommendations for modifications to the existing food consumption and dietary intake surveys . The specific recommendations of the Committee are dis cussed in detail in Chapters 6 and 7. SU=ARY The National Nutrition Monitoring System (NNMS) proposes to collect and report data that permit estimates of the food, nutrition, and health status of the U.S. population. At the core of this system are separate surveys: the USDA's Nationwide Food Consumption Surveys (with its Household Food Use and Individual Dietary Intake components ~ and the DHHS surveys (in particular, the National Health and Nutrition Examination Survey ~ . Data from separate surveys must be used to provide information of the scope needed to assess the food, dietary, nutritional, and biochemical factors associated with the nutritional health of the U.S. population and its sub- groups. Comparable and compatible (linked) data from the separate core survey data bases provide the information needed to assess the foodand nutrition- related risks to health that may be amenable to intervention and to assist in selecting from among possible intervention options. Modifications of the existing surveys should ensure more effective con- joint use of data from them without compromising their essential independent uses. Information that addresses fundamental questions common to the needs of all data users provides the basis for survey modifications of high priority.
61 ~ ~ 1 · ~ c ~ o O AL ~ a` ~ t.'b A A :~- ~ -= ° ~ ~ ~ e Liz e: ~ s 4! so ~ \ Hi= Ike `i ~ °' ~ ~ ~ ~ ~ ~ \ \ ., ~ - to c · - o · - c o · - c 0 ~ 0 I ~ E Us ~4 _ _ en O . - C o . - · - C C C ~ · - o o O · - 0 C o . - LJ O I: · - - \ 0 C 'a i: · - O · - O ~ C) U: :~: U: Z ~ ' - ' . ~; o ~ ~ Z ~ ~ C. ~ ~ O \N _ ~ ~ <: ~ ~ ~ '. - ~ · - O CO ~ ~ Z ·. O 1 ~ o 1 / ° 1 -; ~ C 1 ·- ~ ~ c' O · - c -¢ u, u, z o ~ ~ 6~d ~ o ~ ~ - ~ ~ c ~ o ~ o ~ o o o , e e U, _ b 40 O "8 8 U, C:, O b Z C, O CO e b O D O ~: ~: C: Z _ a ~ ~e 0 ~ Q z U: ¢ Z O U) ~ 8 z - o o :^ U, C o · - · - ~ C o 0 ~ _ 0 ~ . c e eo ~ 0 . C, C~ _ / . Jo ,. O ~4 . · e ~ . / . .
- 62 - Figure 5-2 Conjoint uses of data where concern to be addressed originates in observations of dietary intake. ________~>NECS -- Identification of concern through use of dietary data , NFCS -- Identification of the population subgroup(~) of concern through use of dietary data / i NHANES -- Corroboration of observations by use of biochemical measures if appropriate measurements can be identified and sample is large enough ~ -- whether to l NFCS -- Selection from tag iOceC~e~C Ian alternatives | Education · Fortification 0 Food distribution programs ~ Regulation NFCS -- Monitoring trends; through succes ~ ive surveys NHANES -- Monitoring trends through successive surveys .)
- 63 - Figure 5-3 Conjoint use of data where concern to be addresses originates in biochemical measures. - NH, AS -- Identification of concerns through deviant biochemical measures < if appropriate measurements can be identified and sample is large enough / NHANES -- . ~ i . Assessment of dietary association on the basis of group means NFCS -- Distribution of dietary association in population subgroup(~) with dietary data 1 1 DECISION -- whether ~ to intervene NFCS -- Selection from among intervention alternatives · Education · Fortif ication · Foot assistance programs · Regulation NFCS -- Monitoring trends ~ I ~ NHANES -- Monitoring bents through successive through successive surveys surveys ~ . .
