Dietary Risk Index (DRI)

Access the DRI Analytical System HERE 


Introduction

Understanding the scope–and location–of a problem is a key step in solving it. There are over 600 pesticides applied routinely to hundreds of different food crops, under widely different circumstances across millions of farm fields. But which applications pose dietary risks worth worrying about and which do not?

Are pesticide risks going up or down? What about imported food – riskier or safer? Which foods pose the greatest risks, and which the least? To what extent does organic farming reduce pesticide dietary risks?

The U.S. government has invested heavily over the last 20 years in generating the information needed to answers these sorts of questions. The USDA has compiled a massive, high quality pesticide residue data set going back to 1991. It’s Pesticide Data Program (PDP) conducts annual pesticide residue tests for the most important foods accounting for pesticide dietary risks to infants and children and these results are compiled to populate this report.

Since passage of the historic Food Quality Protection Act in 1996, the EPA has completed in-depth dietary risk assessments of nearly all widely used pesticides. We draw on USDA’s residue data and EPA’s risk assessments in calculating the Dietary Risk Index (DRI) values.

Methodology and Data Sources

The DRI value for a given pesticide-food combination is a ratio — the average residue level found in the food (mean of positive samples), divided by the maximum amount of the pesticide that can be in the food without triggering EPA’s “level of concern.”

This level is called the “chronic Reference Concentration” (cRfC), and is based on the grams in a typical serving of the food, the weight of the individual consuming the food, and the chronic Reference Dose (cRfD) or chronic Population Adjusted Dose (cPAD) for the pesticide.

The pesticide dietary risk associated with a serving of a given food is equal to sum of the Dietary Risk Index (DRI) values for each pesticide found in the food.

DRIs can be calculated based on mean residue levels across all positive samples, producing “DRI-Mean” (DRI-M) values. It is appropriate to use DRI-Ms when comparing, for example, the risk levels for pesticide x versus pesticide y in a given food. This formulation of the DRI does not take into account how frequently a positive residue is found in a particular food.

Suppose there are two positive samples out of 600 samples tested for pesticide x in food y, with a mean (average of two) residue level of 0.15 ppm (parts per million). In another year, or in food grown in a different area, suppose there are 450 positive samples out of 600, with a mean residue level of 0.15 ppm. The DRI-M values would be the same for these two sets of samples.

To take the frequency of residues into account, we also calculate a “Food Supply DRI,” or FS-DRI. It is simply the DRI-Mean multiplied by the percent of samples testing positive for each given pesticide.

For further details on how the DRI is calculated and where residue data come from, see the of The Organic Center’s DRI methodology full report and two-page summary.

Levels of Aggregation

The DRI can be applied at several levels of aggregation. The most detailed level quantifies dietary risks associated with the residues of a single pesticide, or its metabolites and/or isomers, in a specific food based on all samples tested in a given year.

Aggregate DRI-Means and FS-DRIs for a specific food can be calculated by adding together the DRI-M and FS-DRI for each pesticide, metabolite, and isomer found in that found during an annual cycle of PDP testing. Aggregate DRI values are reliable estimates of average, total pesticide risk in any given food based on PDP testing results.

The same can be done for specific pesticides. Aggregate, annual pesticide DRI-M and FS-DRI values can be computed by simply adding together the DRI-M and FS-DRI values across all foods in which a given pesticide was found in samples tested by the PDP in a given year. In the case of pesticide DRI-M and FS-DRI values, these are not reliable estimates of overall annual dietary exposure to specific pesticides, because the PDP only tests around 25 foods each year.

Aggregate DRI-Ms and FS-DRIs across all foods tested in a year can be added together. The sum represents the total dietary risk pool reflected by the foods tested by PDP in a given year. These total DRI values across all foods tested are valuable in assessing the relative shares of risk accounted for by:

  • Specific pesticides
  • Pesticide families of chemistry
  • Types of food (fresh fruits, fresh vegetables, processed foods, juices, milk, etc)
  • Domestically grown food
  • All imported food, as well as imports by country, and
  • Conventionally grown food versus organic food.

The Rule of 10

The PDP typically tests 600 to 750 samples of a given food in an annual program cycle. On average, about 40% of fresh fruits and vegetables are imported. PDP identifies the country of origin, so that country-specific DRIs can be computed, as well as DRIs covering “combined imports.” Whenever there are less than 10 samples of a food from a given country, the “rule of 10” prohibits the calculation of a DRI because of an inadequate number of samples to produce a reliable DRI value.

The program also classifies samples according to market claims, with the vast majority falling into two groups – “no claim” (and presumed to be conventionally grown) and organic. But for some foods, the number of organic samples is limited and the rule of 10 is invoked. In the case of imported organic food, there are rarely 10 or more samples in a given year, and hence the rule of 10 makes it difficult to assess differences in DRI values across organic food from multiple countries. This is why most comparisons in DRI values between domestically grown and imported foods are based on “combined imports.”

Technical Issues and Challenges

A number of factors must be taken into account in order to produce the most accurate and reliable set of DRIs, and to properly apply the DRI in tracking changes in pesticide risk levels and distribution over time. Some impact the exposure side of the DRI equation and limits inherent in the PDP dataset, while others reflect issues involving pesticide toxicity and how cRfCs are calculated.

Dealing with Isomers and Metabolites

The PDP tests foods for parent pesticide active ingredients, as well as major metabolites and/or isomers. In many cases, the PDP has changed how it reports the residues of a particular pesticide over time. In general, PDP has added more metabolites and isomers in recent years, in an effort to more completely account for possible residues stemming from a given application of an active ingredient

In a small number of cases, the PDP has reported in some years, but not all, results for “Total” or multiple residues (e.g., “Total Endosulfans” and “Endosulfans,” or “Total Permethrins” or “Permethrins”), in addition to the parent compound and one to several metabolites/isomers combined. In such cases, special care must be exercised to avoid double counting residues and risk in estimating DRIs

In cases where the PDP reports results for individual isomers and metabolites, as well as “total” residues of the pesticide, DRI values for such a pesticide are based just on the “total” residue results. This is necessary to avoid possible double counting of residues present in a given food.

When PDP reports values for the parent compound and one or more isomer or metabolite, a DRI value is calculated for each based on the residue data specific to the parent compound, isomers, and metabolites, and then added together to equal the total pesticide DRI. In most cases, EPA does not request nor evaluate toxicology data on metabolites and isomers, and for this reason calculates risks associated with exposures to metabolites and isomers using the cRfD or cPAD associated with the parent chemical. This same approach is used in estimating DRI values.

Residues Not Linked to Field Applications

Some of the pesticide residues detected in a given food in the course of PDP testing may not stem from applications of pesticides in the field during the growing season. For many fruits and vegetables, one-third to one-half of the residues found are from fungicides applied post-harvest, typically in storage facilities. Clearly, farmers have no control over these applications

Chlorinated hydrocarbon insecticides (e.g., DDT, aldrin, toxaphene, chlordane) that were banned some 30 years ago still account for the majority of residues found in many animal products and some root and leafy green crops. Again, there is little a farmer can do to avoid these residues when crops are grown that are prone to taking up soil-bound organochlorine residues

A significant share of the residues found in organic food samples are from drift, contaminated fog or water, or residues bound in the soil. Growers have little or no control over these classes of residues.

DRI Values in Organic vs. Conventional Food

Avoiding pesticide risk remains the number one reason people switch to organic foods and beverages.  Hence, it is understandable that many people want to know if organic food actually delivers higher margins of safety.

Food companies are constantly seeking more complete, accurate information identifying potential pesticide risk problems in food product and ingredient supply chains. Critics of organic food and farming often assert, generally with no supporting data, that organic food is as risky as conventional food, because organic farmers occasionally use pesticides approved for use on organic farms.

The USDA’s Pesticide Data Program (PDP) strives to sample organic food proportional to its market share. So if organic fresh apples account for 15% of total fresh apple sales, about 15% of the apple samples tested by the PDP in a given year should be organic. The number of organic samples tested by the USDA has grown steadily over the years, but not as fast as organic market share, and hence, organic food has been under-sampled by the PDP.

The DRI system has been used extensively to quantify relative pesticide risks in organic vs. conventional foods. An based on the DRI is hosted on The Organic Center’s website, and can be used to compare the frequency of residues, the average number of residues, and DRI values for conventional vs. organic samples of foods tested by PDP over the years.

“Sustainability” Paper

An article in the journal Sustainability analyzes the pesticide residues found in organic foods in PDP testing. Written by Chuck Benbrook and Brian Baker, the paper provides an overview of all residues found in organic food in the last decade. All residues found in organic food samples were divided into four categories:

  1. Legacy contaminants like the long-banned organochlorine insecticides
  2. Fungicides used post-harvest in packing sheds.
  3. Pesticides approved for use on certified organic farms.
  4. “Other” pesticide residues.

In domestically-grown organic food from 2002-2011, post-harvest fungicides and legacy chemicals accounted for just under 50% of the total 905 residue detected. Organic farmers have little control over either of these sources of residues.

Another 197 residues were found of pesticides allowed for use on organic farms – mostly the bioinsecticide spinosad. “Other” residues accounted for 29.2% (265) of the total number of residues detected (905).

The paper assesses the impact of current National Organic Program (NOP) policies and requirements that target pesticide residues in organic food. The focus is the degree to which the NOP, organic rules, and certifiers are helping to avoid pesticide residues that pose worrisome levels of dietary risk.

Current NOP policy allows food to be sold as organic as long as it contains residues of a prohibited pesticide that are below 5% of the applicable EPA tolerance. So, we assessed the degree to which this policy targets relatively high-risk residues, and concluded it does not.

The Sustainability paper introduces a new concept that will help the NOP and certifiers better target efforts to continuously drive down pesticide risks in organic food. “Inadvertent” residues in organic food are defined as those over which the organic farmer has little or no control — residues in soil or irrigation water, residues that drift onto an organic field from a nearby conventional farm, or residues that find their way onto organic food during packing, storage, or shipment (like post-harvest fungicides).

The vast majority of “inadvertent” residues are present at levels far below those found in conventional foods – generally 10-times to 1,000-times lower. The vast majority of these residues poses very modest risk, and hence should not be the primary target of NOP and certifier enforcement actions focused on reducing pesticide risks.

The paper presents data supporting three encouraging conclusions –

  1. Consumers hoping to reduce pesticide dietary exposures and risk through seeking out organic fruits and vegetables are accomplishing this objective.
  2. The majority of residues found in organic food are low, or very-low risk inadvertent residues over which the farmer has little control.
  3. Data and tools now exist for the organic community to systematically target and avoid possibly risky pesticide residues in organic food, although some changes in NOP policy and requirements will be needed to fully exploit new knowledge in the effort to drive down pesticide risks.

The DRI Analytical System

Pesticide residue levels, frequency, and risks vary greatly across foods and by food form, in different production regions in the U.S. and abroad, over time, and by production system (e.g., organic vs. conventional). The USDA’s Pesticide Data Program (PDP) provides detailed information on each sample collected, allowing analysts to isolate samples that meet certain criteria.

Pesticide residue and risk results also can be quantified at different levels of aggregation, ranging from individual samples, to all residues found in a given food, and all residues in all foods.

The DRI analytical system is composed of over 160 tables that calculate a set of basic descriptive statistics on the frequency, mean level, and DRI values. Each table contains full results for each PDP program year, beginning with the most recent year of testing down to the earliest year in the system.

All PDP samples in a given year are broken into the following four categories: domestically grown samples, all imported samples, imports by country, and all samples together. All tables within the DRI system are calculated four ways, corresponding to these four sets of samples.

All PDP samples are broken into three categories of market claim: organic, no market claim (presumptively conventional), and all samples combined.

Most tables in the system are calculated at three levels of aggregation: individual food-pesticide combinations; aggregate risks by food and by pesticide; and, total risks across all foods and pesticides.

So, there are 36 versions of each table in the DRI system — all combinations of 4 categories, 3 market claims, and 3 levels of aggregation.

The basic tables include:

  • Number of samples by food and origin
  • All residues found in a specific food and DRI values, by food;
  • All foods in which a pesticide was found and DRI values, by pesticide;
  • Aggregate DRI by pesticide including OCs (organochlorine insecticides);
  • Aggregate DRI for OP (organophosphate) insecticides by food; and
  • Aggregate DRI for OPs by pesticide across all foods.

There are 36 versions of each of these 6 tables, resulting in a total of 216 tables. Each table covers 14 years of PDP testing. When printed, the tables range from over 400 pages to around a dozen, and average around 40 pages.

Given the magnitude of the information in the these tables, the DRI analytical system includes a series of look-up tables that draw data from the above set of tables. Access the DRI Analytical System tools here.

The Rule of 10

The PDP typically tests 600 to 750 samples of a given food in an annual program cycle. On average, about 40% of fresh fruits and vegetables are imported. PDP identifies the country of origin, so that country-specific DRIs can be computed, as well as DRIs covering “combined imports.” Whenever there are less than 10 samples of a food from a given country, the “rule of 10” prohibits the calculation of a DRI because of an inadequate number of samples to produce a reliable DRI value.

The program also classifies samples according to market claims, with the vast majority falling into two groups – “no claim” (and presumed to be conventionally grown) and organic. But for some foods, the number of organic samples is limited and the rule of 10 is invoked. In the case of imported organic food, there are rarely 10 or more samples in a given year, and hence the rule of 10 makes it difficult to assess differences in DRI values across organic food from multiple countries. This is why most comparisons in DRI values between domestically grown and imported foods are based on “combined imports.”

History of the DRI

The seminal 1993 National Academy of Sciences report “Pesticides in the Diets of Infants and Children” (NRC/NAS, 1993) set forth the reasons why pesticides can lead to significantly heightened risks when infants, children, and pregnant women are exposed. It called for improved data on pesticide residues in key children’s foods, a need fulfilled in large part by the PDP. The primary recommendations of the 1993 NAS report were incorporated in the historic 1996 Food Quality Protection Act (FQPA), legislation that directed the EPA to dramatically alter the way it sets tolerances governing pesticide residues in food. The FQPA provided the EPA key new tools to assure a “reasonable certainty of no harm” stemming from pesticide exposures via all routes for all vulnerable population groups, and in particular pregnant women, infants, and children.

In cases where the EPA determines that an added safety factor (usually 3-X or 10-X) is required in accord with the provisions of the FQPA, a pesticide’s RfD is reduced 3-X or 10-X and becomes a Population Adjusted Dose (PAD). Dietary risk assessments are then typically based on a pesticide’s cRfD or cPAD, because in most cases chronic-risk benchmarks are lower than acute RfDs or PADs.

The first version of the DRI focused on cancer as the toxicological endpoint, and was developed and applied under the direction of the National Academy of Sciences/National Research Council (NAS/NRC) committee that wrote the 1987 report Regulating Pesticides in Food: The Delaney Paradox (NRC/NAS, 1987).

The DRI concept was further developed and applied to chronic risks in the 1996 book Pest Management at the Crossroads published by Consumers Union (Benbrook et al., 1996).

Dietary risks associated with potato pesticide use has been tracked as part of a collaborative project involving the World Wildlife Fund, the Wisconsin Potato and Vegetable Growers Association, and the University of Wisconsin-Madison (Benbrook et al, 2002).

The EPA’s Office of Inspector General sponsored an independent analysis of the impacts of the FQPA on dietary risks through 2002, based on a version of the DRI designed to track, as closely as feasible, EPA dietary risk assessment policies (OIG, 2006a and 2006b).

The DRI has been applied in multiple other analyses for non-profit organizations, companies, and government agencies.

References

Benbrook, C. M. et al., 2002 “Developing a pesticide risk assessment tool to monitor progress in reducing reliance on high-risk pesticides.” American Journal of Potato Research, 79: 183-99.

Centers for Disease Control and Prevention. 2013. “2 to 20 years: Boys Stature-for-age and Weight-for-age percentiles,” www.cdc.gov/growthcharts, accessed 4/23/12.

Office of Inspector General (OIG). 2006a. “Measuring the impact of the Food Quality Protection Act: Challenges and opportunities.” Report No. 2006-P-00028, August 1, U.S. EPA, Washington, D.C.

Office of Inspector General (OIG). 2006b. Supplemental Report: Details on dietary risk data in support of Report No. 2006-P-00028, Measuring the impact of the Food Quality Protection Act: Challenges and opportunities. U.S. EPA, Washington, D.C.

National Research Council/National Academy of Sciences. 1987. Regulating Pesticides in Food: The Delaney Paradox. Washington D.C., National Academy Press.

National Research Council/National Academy of Sciences. 1993.  “Pesticides in the Diets of Infants and Children.” Washington D.C., National Academy Press.