The first analysis would typically be a Tier 1 analysis. For a Tier 1 analysis, consumer intake could be assessed by using screening methods based on conservative assumptions. A conservative screening method might be based on the 95th-percen-tile corn consumer, assuming all of the protein remains in the food at the time it is eaten, including fractions such as corn sugar.

Corn example:

Tier 1:

The intake of our novel protein by the person who eats 154 g/corn/day (USDA, CSFII 1994-98 using DEEM™ software or might further consider a subgroup with high corn consumption, such as Hispanics (164 g/corn/day). The consumption could then be combined with the estimates of protein in corn from USDA and the fraction of that protein that would be the introduced protein to conservatively estimate intake of the introduced protein. In this example, it would be 154 g x 8.12 g/100 g x 500 |g introduced protein/100 g corn protein = 63 |g introduced protein/day. In this example, the results are a screening value that overestimates typical intake and also assumes that all of the protein was still in the food as it was consumed. These could easily be refined by excluding non-protein-containing food products. Another refinement could be made if data were available to show that the protein in question was degraded during processing. For example, the preparation of tortillas using nixtamilization degrades most proteins.

The results would then be compared to some measure of safety — perhaps comparison to an upper reference value for protein intake or to the results of animal feeding studies.

If additional analyses are desired to refine the screening intake, it is possible to refine both the consumption and composition values. In this example, the fractions of corn that do not contain protein, such as sugar and HFCS, would be excluded from the analysis.

In the sections that follow, examples of the available methods have been organized (somewhat arbitrarily) into categories to assist the reader in selecting the most appropriate framework and the desired methods for each step of the framework. The methods are divided into those that provide single (point) estimates and those that characterize the full distribution of consumer intakes.

Characterizing the full distribution of consumer intakes is the most resource-intensive assessment, since data are required that are characteristic of the range of consumer consumption practices as well as the range of introduced protein levels in the foods that are eaten. Therefore, such methods are usually reserved for later steps. When the methods are employed, appropriate statistical models are used to evaluate the data and to describe the range of consumer intakes and the associated probabilities of consumers having each level of intake. These intake assessments are generally referred to as probabilistic or Monte Carlo intake estimates.

For substances requiring further refinement beyond screening methods or point estimates of intake as described above, a probabilistic analysis of the variability in intakes can be conducted. Conceptually, the population's intake must be thought of as a range of values rather than a single value because individual members of the population will consume different amounts, and even the same individual will consume different amounts on different days. Factors that contribute to this variability include age (due to differences in body weight and the type and amount of food consumed), gender, ethnicity, nationality and region, and personal preferences, among others. Variability in dietary intake is often described using a frequency distribution. The differences in point estimates and distributions are further described in the following sections.

9.4.2 Point Estimates of Dietary Intake

A point estimate is simply a single value that describes some parameter of a consumer's intake (e.g., the average U.S. population's intake of protein "x"). For example, an average consumer's intake is calculated as the product of the average consumption of the foods of interest and the average levels of the introduced protein in those foods. The resulting estimate can be further adjusted by additional adjustment factors as appropriate (processing factors, etc.). A point estimate that estimates a high consumer's intake (such as the upper 90th-percentile consumer) can also be calculated, provided the appropriate data are available.

A point estimate is not inherently "conservative" or "realistic." The conservatism incorporated into the analysis is determined by the data and the assumptions that are used in calculating the estimate. Point estimates can range from initial screening methods which use very little data and generally include very conservative assumptions, to refined intake assessments which include extensive underlying data in order to realistically calculate the desired estimates of intake.

Dietary intake assessments can be based on a food consumption distribution determined empirically from a food consumption survey and a single-point estimate to represent the concentration of the introduced protein in the relevant food product. Each point of a distribution curve of food consumption can be multiplied by the concentration level in the relevant food commodity. Conversely, it is possible to have a single-point estimates for consumption and an empirical distribution of introduced protein concentrations in that food. Finally, it is possible to have sufficient data to determine the distribution profile for both the amounts of food consumed and the levels of the introduced protein in those foods.

An example of a conservative point estimate of intake would be one that is derived from food disappearance data (often referred to as food balance data). Food balance data are generally available for most countries. These data include the amounts of foods available for human consumption derived from national statistics on food production, disappearance, or utilization, such as those compiled by the USDA's Economic Research Service32 or the Australian Bureau of Statistics.18 The Food and Agriculture Organization of the United Nations (FAO) FAOSTAT database is a compilation of similar statistics for more than 250 countries. The data are compiled, or estimated when official data from member countries are missing, from national food production and utilization statistics.33

Fantastic Organic Food Facts

Fantastic Organic Food Facts

Get All The Support And Guidance You Need To Be A Success At Utilizing Organic Foods. This Book Is One Of The Most Valuable Resources In The World When It Comes To Getting The Right Information About Eating Healthy With Organic Food.

Get My Free Ebook

Post a comment