Cover Image

PAPERBACK
$48.00



View/Hide Left Panel

Appendix E
Background on the National Health and Nutrition Examination Surveys and Data Analysis Methods

OVERVIEW

In the 1960s, the National Health Examination Survey began to assess the health status of individuals ages 6 months through 74 years, including measures of hypertension, elevated serum cholesterol, and overweight. Nutritional intake was added as a survey component in the 1970s. Therefore, the 1970s mark the time during which information on sodium intake became available from this survey, beginning with the first National Health and Nutrition Examination Survey, known as NHANES I (1971–1974).1 The U.S. Department of Agriculture’s (USDA’s) food composition database has provided the sources of information that allow the estimates of food intake collected in the NHANES to be translated into quantitative nutrient intake (Bodner-Montville et al., 2006; Briefel, 2006). The NHANES and related food intake surveys conducted by USDA were integrated in 2002, and at that time the dietary reports from the integrated survey became known as What We Eat in America (WWEIA).

The NHANES reflects a continuous and standardized data collection based on a representative sample of the U.S. population and major subgroups (including those related to race/ethnicity and income) and provides critical diet and health measures for federal program planning and policy

1

The U.S. Department of Agriculture (USDA) Household Food Consumption Surveys conducted in 1977–1978 and 1987–1988 provide a snapshot of food sources of sodium in the 1970s and the 1980s at the household-level. More recent data have not been collected, although USDA plans to collect similar household food acquisition and food cost data in 2010 or 2011.



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 409
Appendix E Background on the National Health and Nutrition Examination Surveys and Data Analysis Methods OVERVIEW In the 1960s, the National Health Examination Survey began to assess the health status of individuals ages 6 months through 74 years, including measures of hypertension, elevated serum cholesterol, and overweight. Nutritional intake was added as a survey component in the 1970s. There- fore, the 1970s mark the time during which information on sodium intake became available from this survey, beginning with the first National Health and Nutrition Examination Survey, known as NHANES I (1971–1974).1 The U.S. Department of Agriculture’s (USDA’s) food composition database has provided the sources of information that allow the estimates of food intake collected in the NHANES to be translated into quantitative nutrient intake (Bodner-Montville et al., 2006; Briefel, 2006). The NHANES and related food intake surveys conducted by USDA were integrated in 2002, and at that time the dietary reports from the integrated survey became known as What We Eat in America (WWEIA). The NHANES reflects a continuous and standardized data collection based on a representative sample of the U.S. population and major sub- groups (including those related to race/ethnicity and income) and provides critical diet and health measures for federal program planning and policy 1 The U.S. Department of Agriculture (USDA) Household Food Consumption Surveys con- ducted in 1977–1978 and 1987–1988 provide a snapshot of food sources of sodium in the 1970s and the 1980s at the household-level. More recent data have not been collected, al- though USDA plans to collect similar household food acquisition and food cost data in 2010 or 2011. 0

OCR for page 409
0 APPENDIX E making. The survey relies on the gold standard for dietary measures, two or more 24-hour dietary recalls per person (IOM, 2000). NHANES is unique in that it collects and tracks both dietary intake and health measures in a nationally representative sample of Americans. Dietary intake estimates are limited by survey respondents’ abilities to accurately report foods and amounts consumed and by the accuracy, specificity, and currentness of the food composition databases used to code foods reported in the survey. They are also prone to underreporting intake (IOM, 2000). Issues related to es- timation of usual intake related to WWEIA-NHANES have been carefully reviewed by others (Dwyer et al., 2003). During the four decades that dietary intake has been tracked in nation- ally representative cross-sectional surveys of the population, there have been changes in the data collection methods and protocols used to estimate dietary intake. The quality of data has improved, but of course some bias and measurement error still exist given that the estimates must rely on self- reported data. Beginning with NHANES III (1988–1994), improvements were made in dietary data collection to produce population-level estimates of total sodium intake to track progress in meeting Healthy People objec- tives for the dietary guidelines for sodium.2 These improvements included the collection of more than 1 day of intake on at least a subsample of the population and questions about tap water consumption and water soften- ing, dietary supplement use, and salt added at the table, including the type of salt. These additional survey questions were intended to produce more complete estimates of dietary sodium intake. Beginning in 2003–2004, two 24-hour diet recalls were collected and released for each person, allowing for estimates of usual nutrient (sodium) intake in the population using statistical software to account for the large day-to-day variations in individual intake (Dodd, 1996). The improvements in dietary data collection and the availability of statistical techniques to assess dietary intake allow for estimates of the population’s usual sodium intake from food sources. When the available statistical software is not applicable to the measure(s) of interest, data on the basis of a 1-day mean are reported. This applies to analyses focused on food categories, sodium intake from earlier studies, and measures of sodium density. ESTIMATION OF CURRENT SODIUM INTAKE FROM NHANES 2003–2006 Current estimates of intake are derived from information available from two recently completed NHANES: 2003–2004 and 2005–2006. Those data sets were combined for this report to provide larger sample sizes for 2 Available online: http://www.healthypeople.gov/document/pdf/tracking/od19.pdf (accessed November 14, 2009).

OCR for page 409
 APPENDIX E subgroup analysis. Analysis weights were appropriately revised following the recommended procedures for combining NHANES survey data.3 Data are weighted to reflect population estimates. Unweighted sample size is shown in the data tables in Appendix F. Sodium from Foods Estimates of food intake are derived from two 24-hour dietary recalls.4 The day 1 interview was conducted in person in the Mobile Examination Center of NHANES. The day 2 interview was conducted by telephone 3–10 days later. As part of the NHANES 24-hour recall interview on day 1, re- spondents are probed to provide details of the food consumed. These probes elicit information such as brand names, preparation method, the form of the food (such as frozen, canned, or fresh), and the type of food to assist in clarifying levels of sodium in the food consumed, as well as fat, calories, and other components and where the foods were obtained. Questions are not asked about salt used in cooking, recipes, or food preparation as part of the 24-hour recall. Rather, a set of health-related questions is administered separately and includes questions about salt use. Respondents are asked how often salt is used in cooking inside the home; this question refers only to ordinary or seasoned salt and not “lite” salt or salt substitutes. Response options include “never,” “rarely,” “occasionally,” and “very often.” This information is applied to algorithms for recipes and sodium absorbed in cooking (Moshfegh, 2009). A statistical method for estimating usual intake distributions and the proportion below or above defined cutoff values has been developed at Iowa State University and makes use of the second-day dietary recall for this purpose (Carriquiry, 2003). This method was applied to estimates of sodium intake for this study, where possible. Certain measures of interest were not applicable to use of this software and were reported as 1-day means; these include estimates of sodium intake from other sources such as tap water. Sodium from Other Sources Other sources of sodium include salt added at the table, tap water, and dietary supplements. The approach used to estimate this intake was developed for the Healthy People 2010 Progress Review, Focus Area 19, 3 Available online: http://www.cdc.gov/nchs/data/nhanes/nhanes_03_04/nhanes_analytic_ guidelines_dec_2005.pdf (accessed March 25, 2010). 4 Available online: http://www.ars.usda.gov/Services/docs.htm?docid=13793 (accessed Oc- tober 26, 2009).

OCR for page 409
 APPENDIX E presented April 20085,6 and described as part of the Healthy People 2010 tracking system.7 Accurate reporting of salt used at the table relies on subjects’ ability to estimate the quantity and frequency with which salt is added to foods. For NHANES, respondents are asked to indicate how often salt is added at the table.8 Response options include “never,” “rarely,” “occasionally,” and “very often.” Sea salt, flavored salts (such as garlic or onion salt), and seasoning salts were counted as ordinary salts. So-called lite salt was recorded as such and has a reduced sodium content. Salt substitutes do not contain sodium. When an analysis incorporates use of salt at the table in the estimation of sodium intake, the amount of sodium depending on salt type is multiplied by the frequency value (i.e., sodium in type of salt multiplied by frequency amount of sodium from table salt added per day) to obtain a daily amount for each person. Regarding type of salt, a zero sodium value is assigned for reports of “none” and “salt substitute.” When “very often” was reported for use, ordinary salt is assigned as 290 mg sodium for persons ages 2–19 years, and 580 mg for persons over 20 years of age. If salt use was reported as “occasionally,” the value for “very often” was multiplied by one-half; for reports of use as “rarely,” the value for “very often” was multiplied by one-fourth. When an analysis incorporates sodium from drinking water, water derived from a water softening or conditioning system is identified as containing 3 mg of sodium per fluid ounce. Otherwise, water is counted as unsoftened. One mg sodium per fluid ounce is used for “regular” mu- nicipal water based on the USDA food composition database. In WWEIA 2003–2004, only sweetened bottled waters were captured in the 24-hour recall. Information on plain water, tap water (and source), and plain car- bonated water was captured in survey questions following the 24-hour recall. For the analysis in this report, waters were categorized as tap water (for the tap water contribution), and other bottled and sweetened waters were categorized as foods in the beverage category. Finally, data on dietary supplements are collected as part of NHANES, but the incorporation of sodium from dietary supplements requires ad- ditional data permutations to link the dietary supplement data set to the foods intake data set. To make this calculation, the content of each dietary supplement reported by the respondent and the frequency of use in the past 5 Available online: http://www.healthypeople.gov/data/2010prog/focus19/Default.htm (ac- cessed November 14, 2009). 6 Available online: http://www.cdc.gov/nchs/healthy_people/hp2010/focus_areas/fa19_nutrition2. htm (accessed November 14, 2009). 7 Available online: http://www.healthypeople.gov/Document/html/tracking/od19.htm (ac- cessed November 14, 2009). 8 Available online: http://www.healthypeople.gov/document/pdf/tracking/od19.pdf (accessed November 14, 2009).

OCR for page 409
 APPENDIX E month are combined to estimate a daily amount of sodium from supple- ments per person. Antacids are included in the estimates of sodium intake from dietary supplements. Usual Sodium Intake Comparison to Dietary Reference Intakes As part of describing current sodium intake, means and distributions of usual intake from foods and from all dietary sources in NHANES 2003–2006 were compared to the Dietary Reference Intakes (DRIs) (IOM, 2005)—that is, the Adequate Intake (AI) and the Tolerable Upper Intake Level (UL) for sodium. If the usual mean intake exceeds the AI, the group is assumed to have adequate intake levels (Murphy, 2003). The proportion of the population that exceeds the UL is determined to be at risk of adverse effects from an excessive intake (Murphy, 2003). The statistical method for estimating the proportion below or above defined DRI cutoff values devel- oped at Iowa State University was used (Carriquiry and Camano-Garcia, 2006). Special Subgroups The NHANES collects information on race/ethnicity on the basis of self-reported categories as follows: non-Hispanic whites, non-Hispanic Af- rican Americans, and Mexican Americans. Income for the survey is also reported on a category basis and is analyzed consistent with standards for reporting nutrition and statistical data for the evaluation of nutrition assistance programs: low-income is defined as an annual household in- come level of 130 percent of poverty or less, the income eligibility for the Supplemental Nutrition Assistance Program, formerly called the food stamp program; higher-income is defined as an annual household income above 185 percent of poverty, the eligibility cut-off for free- or reduced-price school meals and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC); and intermediate income is between 130 and 185 percent of the poverty line. Mean sodium intake from foods is highest among low- and higher-income adults ages 19–30 years and higher-income adults ages 31–50 years (Appendix F, Table F-6). Hypertension was defined as an elevated blood pressure (systolic pressure ≥ 140 mm Hg and diastolic pressure ≥ 90 mm Hg) and/or the taking of antihypertensive medications at the time of the individual’s medical examination in the NHANES Medical Examination Center (NCHS, 2009).

OCR for page 409
 APPENDIX E ESTIMATION OF TIME TRENDS FOR SODIUM INTAKE FROM NHANES Trends in mean sodium intake have been reported for age/gender sub- groups from age 1 year through 74 years from 1971–1974 to 1999–2000 (Briefel and Johnson, 2004). NHANES collected single 24-hour dietary recalls in 1971–1974 and estimated mean sodium intake from foods by age group and gender for the household-based population ages 1 year through 74 years. The age range was expanded to 2 months and older with no up- per age cutoff in NHANES III (1988–1994), and from birth on starting in NHANES 1999. Nutrient intake is not reported for breastfeeding infants. To update the Briefel and Johnson analysis (2004) and allow for com- parison to current estimates of intake, estimated 1-day mean sodium intake from foods9 in NHANES 2003–2006 was derived using analytic techniques comparable to those used in the earlier analysis. A table was then generated to compare intake estimates from NHANES 2003–2006 to the existing time trend analysis (Briefel and Johnson, 2004). APPROACH TO CHARACTERIZING SOURCES OF SODIUM IN NHANES 2003–2006 The food category analysis used data from NHANES 2003–2006 and relied on the food categorization scheme used in a previous NHANES analysis by Cole and Fox (2008). In brief, all foods reported in the 24-hour dietary recalls are grouped into 11 major categories and into 154 food groups. Nearly 4,000 unique food codes were used to code foods reported in NHANES 2003–2004 (n = 3,894 foods). The estimates of sodium from foods include salt used in cooking and food preparation, but not salt added at the table. Further, foods are not disaggregated at the ingredient level, and salt that was used in recipes is also included in the sodium content of the food “as prepared.” Foods are recorded as reported by consumers, for example, as an apple, a mixed dish, or a sandwich. In some cases, individual components were reported, and it was not always possible to aggregate or disaggregate all reported foods at the same level. One-day dietary recall data were used to estimate food sources of sodium and mean daily sodium and sodium density by home versus away food source using the population proportion method as described by Krebs- Smith and colleagues (1989). For this report, the committee classified food sources as “Home,” “Away,” and “Other” based on the food source cat- egories listed in Box E-1. “Home” sources are those foods obtained from the store and assumed to be consumed at home. “Away” sources include restaurants (which include those with waiter service, fast food and pizza 9 This estimate includes salt used in cooking and in food preparation.

OCR for page 409
 APPENDIX E BOX E-1 Food Source Categories Home includes: Other includes: • Foods prepared at home • Child care center • Foods purchased from the • Family/adult day care center store • Meals on Wheels • Community food programs Away includes: • Vending machine • Restaurant with wait staff • Common coffee pot or snack tray • Restaurant fast food/pizza • From someone else/gift • Bar/tavern/lounge • Mail order purchase • Restaurant, no additional • Residential dining facility information • Grown or caught by you or someone you know • Cafeteria not at school • Fish caught by you or someone you know • Cafeteria at school • Sport, recreation, or entertainment • Street vendor, vending truck • Fundraiser sales places, and bar, tavern, or lounge categories) and cafeterias (school and non-school). “Other” sources for purposes of this analysis represent an av- erage of 22 percent of the daily sodium in 2003–2006 and include sources such as child care centers, vending machines, street vendors, sports events, and community food programs. The data used to characterize sodium intake by contributing source are largely obtained from self-reported intake surveys coded using composition databases. As such, they are subject to the same limitations described for estimating intake by self-report. As discussed earlier, the constantly evolving food supply and increasing globalization make it a challenge to maintain updated food composition databases or databases for supplements that may also undergo formulation changes. Furthermore, new technologies for analyzing samples may change established values for the nutrient content of certain foods. Each food item listed in self-reported intake surveys has a code that corresponds to an entry in the database. However, foods may not have unique food codes; they are often grouped with similar foods within a food group and assigned an aggregate nutrient content based on the market share of the items in the food code. The categorization of foods in the database can affect the ability to track the contribution of individual food items to sodium intake over time. How researchers decide to categorize and report foods can also have a major influence on the rank ordering of which foods are the greatest con- tributors to sodium intake and can make data comparisons across studies difficult (Cole and Fox, 2008; Cotton et al., 2004; Subar et al., 1998).

OCR for page 409
 APPENDIX E REFERENCES Bodner-Montville, J., J. K. C. Ahuja, L. A. Ingwersen, E. S. Haggerty, C. W. Enns, and B. P. Perloff. 2006. USDA Food and Nutrient Database for Dietary Studies: Released on the web. Journal of Food Composition and Analysis 19(Supplement). Briefel, R. R. 2006. Nutrition monitoring in the United States. In Present Knowledge in Nu- trition. 9th ed, edited by B. Bowman and R. Russell. Washington, DC: ILSI Press. Pp. 838-858. Briefel, R. R., and C. L. Johnson. 2004. Secular trends in dietary intake in the United States. Annual Reiew of Nutrition 24:401-431. Carriquiry, A. L. 2003. Estimation of usual intake distributions of nutrients and foods. Journal of Nutrition 133(2). Carriquiry, A. L., and G. Camano-Garcia. 2006. Evaluation of dietary intake data using the tolerable upper intake levels. Journal of Nutrition 136(2). Cole, N., and M. K. Fox. 2008. Diet quality of American young children by WIC participation status: Data from the National Health and Nutrition Examination Surey, –00. Document No. PR08-67. Washington, DC: Abt Associates, Inc. and Mathematica Policy Research. Cotton, P. A., A. F. Subar, J. E. Friday, and A. Cook. 2004. Dietary sources of nutrients among U.S. adults, 1994 to 1996. Journal of the American Dietetic Association 104(6): 921-930. Dodd, K. W. 1996. A technical guide to C-SIDE: Software for intake distribution estima- tion. Dietary Assessment Research Series, Report 9 Technical Report 96-TR 32. Ames: Iowa State University Department of Statistics and Center for Agricultural and Rural Development. Dwyer, J., M. F. Picciano, D. J. Raiten, P. P. Basiotis, M. M. Bender, B. K. Bindewald, A. L. Carriquiry, A. K. Courtney, N. T. Crane, K. W. Dodd, K. Egan, K. C. Ellwood, S. E. Gebhardt, J. F. Guthrie, J. M. Harnly, J. M. Holden, C. Johnson, S. M. Krebs-Smith, P. M. Kuznesof, C. E. Lang, M. McDowell, A. Moshfegh, P. R. Pehrsson, K. Radimer, A. F. Subar, C. A. Swanson, and W. R. Wolf. 2003. Estimation of usual intakes: What we eat in America-NHANES. Journal of Nutrition 133(2). IOM (Institute of Medicine). 2000. Dietary reference intakes: Applications in dietary assess- ment. Washington, DC: National Academy Press. IOM. 2005. Dietary Reference Intakes for water, potassium, sodium, chloride, and sulfate. Washington, DC: The National Academies Press. Pp. 269-423. Krebs-Smith, S. M., P. S. Kott, and P. M. Guenther. 1989. Mean proportion and population proportion: Two answers to the same question? Journal of the American Dietetic As- sociation 89(5):671-676. Moshfegh, A. 2009. Surveillance and monitoring sodium intakes. Presented at the Institute of Medicine Committee on Strategies to Reduce Sodium Intake’s Public Information- Gathering Workshop, March 30, Washington, DC. Murphy, S. P. 2003. Collection and analysis of intake data from the integrated survey. Journal of Nutrition 133(2). NCHS (National Center for Health Statistics). 2009. Health, United States, 00 with special feature on the health of young adults. DHHS Publication No. 2009-1232. Hyattsville, MD: National Center for Health Statistics. Subar, A. F., S. M. Krebs-Smith, A. Cook, and L. L. Kahle. 1998. Dietary sources of nutri- ents among US adults, 1989 to 1991. Journal of the American Dietetic Association 98(5):537-547.