Key Concepts About Deciding Which Environmental Chemical Variables to Include in Your Dataset
Data files are organized into four categories, as listed below. All environmental chemical variables are included in the Laboratory section and organized by sample specimens and chemical groups. For each two-year data release included in an analysis, it is important to review the laboratory section because you will find the general information for overall laboratory measurements and specific information related to each environmental chemical may change.
- Demographics files contain survey design variables such as sample weights, strata and primary sampling units, as well as demographic characteristics, such as sex, age, and race.
- Examination files contain information collected through physical exams (including anthropometric and physical activity data, dental exams, and 24-hour dietary recall interviews.)
- Laboratory files contain results from analyses of blood, urine, hair, tuberculin skin test, air, and household dust and water specimens. These files also include subsample data and subsample weights that should be used in analyses of these analytes.
- Questionnaire files contain data collected through household and MEC interviews.
|
|
It is advisable to read the Documentation to fully understand
this important information and identify the correct variables
for your analysis.
For each two-year cycle of NHANES, review the general information and the Variable List to decide which environmental chemical variables are needed in your analysis. Use the search feature to speed up your search for variables relevant to your analysis. Carefully read the Documentation and related Laboratory Procedure Manuals for each of your search results.
Special Note on Limit of Detection (LOD) Variables
For many enviromental chemical measures, a variable is present in the data file to indicate if the measured value of the environmental chemical is beyond the limit of detection of the laboratory assay. It is helpful to include this variable when creating an analytic dataset. See Module 11 for further information on the LOD variables.
Special Note on Lipid and Creatinine-adjusted Variables
In reviewing the documentation you will notice that some of the blood measurements are lipid-adjusted and some are unadjusted. In data files for serum dioxins, furans, polychlorinated biphenyls, polybrominated diphenyl ethers, and organochlorine pesticides, each analyte is expressed as two forms, lipid-adjusted and unadjusted, because these compounds are lipophilic and concentrate in the body’s lipid stores, including the lipids in serum. Serum levels reported per gram of total lipid reflect the amount of these compounds that are stored in body fat.
Similarly, for chemicals measured in urine, levels can be presented two ways: per volume of urine and per gram of creatinine. Urinary levels are expressed both ways in the literature and used for different purposes. Levels per gram of creatinine (i.e. "creatinine corrected" values) adjust for urine dilution. For example, if one person has consumed more fluids than another person, his or her urine output is likely higher and the urine more dilute (has less urine creatinine) than that of the other person. Interpretation of creatinine corrected results should also recognize that creatinine correction can also partially adjust for differences in lean body mass or renal function among persons. Unlike the lipid-adjusted measurements, which are included on the data files, "creatinine corrected" values need to be created by the data user with the information available on the data file. Further details about lipid and creatinine adjustment is available in Module 13.
References
- For more information about lipid and creatinine adjustment, please see Module 13 "Using Urine Creatinine or Blood Lipid Adjustments of Environmental Chemical Data" of this tutorial.
- For more information about LOD variables, please see Module 11 "Important Analytic Considerations and Limitations Regarding Environmental Chemical Data Analyses" of this tutorial.
