Task 3: Key Concepts about Estimating Mean Nutrient Intakes from Supplements

This task relates to estimating mean nutrient intakes from supplements only.  That is, these estimates do not include nutrient intake from dietary (food and beverage) sources.

Nutrient intakes from supplements are estimated using data from the supplement questionnaire.  Very little is known regarding the extent of bias or random errors associated with dietary supplement data.  For that reason, and for practical purposes, the supplement data are generally treated as though none of either type of error occurred.  However, the possibility of both should be noted as a caveat in an analysis of this type.

 

Unlike the data derived from the recalls, there are no data files that provide total daily amounts of each nutrient across all supplements.  Therefore, this must be calculated for each person first.  There are a few key points to note when calculating supplement intake.  First, each supplement could be reported with a different frequency, based on use over the past 30 days, so care must be taken in deriving the intakes from all supplements.  Second, the measurement unit for a given supplement may not be the same across all brands, so conversions may need to be made to combine nutrient values.  Third, nutrients may be listed as compounds and need to be converted to elemental form and amounts (e.g. calcium carbonate would need to be converted to the corresponding amount of elemental calcium in order to determine total calcium); however, this is less of a concern with the supplement data releases since 2001 because recent releases have coded supplements using elemental (common) names rather than compounds.

The variables needed to calculate mean nutrient intake come from the Supplement Data Files 2, 3, and 4.  File 2 provides usage of the supplement, File 3 provides information on the supplement itself, and File 4 provides all the ingredients in the supplement.  (For more information on merging supplement files, see “Module 8: Merge and Append Datasets, Task 1: Merge NHANES data.”) Missing data can be a limitation with several of the dietary supplement variables.  File 2 contains the most missing data.  The number of cases of missing data and the possible remedies vary by the particular variable, as follows: 

• DSD103, which is the number of days the supplement was taken in the past 30 days, has 227 cases of missing data for 2003-2004.  This is because some of the supplements were inadvertently reported in the prescription medication section of the NHANES interview, and that section does not ask participants how often they took the product in the past 30 days.  Because this variable is needed to determine usual intake, analysts can either impute a number of days or drop these records from the dataset.  Imputation requires an assumption that the supplement was taken regularly and is usually based on some other information the respondent provided, such as the number of days that the respondent reported taking certain other types of dietary supplements.

• The variables DSD122Q / DSD122U, which capture, in quantity and units, responses to the question, “On the days you took the supplement, how much did you take?” also have more than 200 missing cases.  This is for the same reason as mentioned in the previous bullet.  That is, this question is not asked in the medication portion of the interview, where some respondents mistakenly report their supplement use.  Analysts may want to impute data, which requires an assumption that the respondent took the serving size listed on the label captured in variables DSDSERVQ / DSDSERVU.

• The supplement name has missing data because no match to the supplement was recorded, no similar supplement exists on the market, or the recorded name was not comprehensible as denoted in the DSDMTCH variable.  However, these records are kept in the data because it is assumed that individuals did take a supplement, even though the name is unknown.  They should be retained for prevalence estimates.  It may be best to exclude these data from analyses in which mean intakes are being estimated.  This action also would reduce missing data for some other variables. 

 

When estimating mean intake from supplements, analysts should decide whether or not they wish to include calcium from antacids.  Antacids are not included in DSD010 (any dietary supplements taken) or in DSDCOUNT (the number of supplements taken).  However, antacids are included in file 2, so the default is that they will be counted.  The variable DSDANTA indicates whether or not the supplement is an antacid and where the antacid data were collected (i.e. during the supplement section of the questionnaire or during the prescription medication section of the questionnaire).  It is necessary to be cautious when using antacid data in analyses because subsequent pilot studies and questions fielded in more recent NHANES suggest antacids are used sporadically and more as a medication than as a supplement.  Including these products in usual intake estimates for calcium may skew results and overestimate usual calcium intake for some individuals.  However, a few antacids were reported during the dietary supplements section of the questionnaire, and they may be assumed to be taken as a supplement.   

Another consideration with estimating mean nutrient intake from supplements is whether you are interested in the mean amount among all persons in the population, or only users of the supplement.  That is, it is necessary to decide whether non-consumers should be included in the estimation.  If you are interested in the per capita amount consumed, you should include the non-consumers with their intake value at zero.  If you are interested in the average amount consumed by users of the supplement, you should exclude the non-consumers.

Means should be examined along with their standard errors, to get an indication of the variation about the mean.  Special statistical procedures are required to get appropriate standard errors when using data from a complex sample such as the NHANES.  In addition, appropriate sample weights should be applied, so the results will represent the population as a whole.  See “Module 13, Estimate Variance, Analyze Subgroups, and Calculate Degrees of Freedom” for more information.