

                                  Appendix A
                                       
  Consideration of Legally Working Children in Pesticide Exposure Assessment
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                         Office of Pesticide Programs
              Office of Chemical Safety and Pollution Prevention
                     U.S. Environmental Protection Agency
                                Washington, DC
                                       
                                       
                                       
                                  12/22/2016




                               Table of Contents



Executive Summary	3
I.	Introduction	6
II.	Problem Formulation	8
III.	Exposure and Risk Assessment Primer	15
A.	Calculation of Transfer Coefficients	15
B.	How Transfer Coefficients (TCs) Are Incorporated Into Risk Assessments	16
C.	Dermal Exposure Rate Calculation Using Harvester Exposure Monitoring Field Studies	18
IV.	Analysis	22
A.	Mechanistic Approach	22
B.	Monitoring Data	24
i.	Overview of Studies	24
ii.	Analysis of Individual Studies	27
1.	Study Example  -  Pesticide Exposure to Blueberry Harvesters in Southwestern Michigan, 1984 (MRID 47928906)	27
iii.	Summary of Combined Results for Youth in Agriculture Studies	36
V.	Discussion	42
VI.	Conclusions	44



Executive Summary   

Pesticide exposures in agriculture may occur while applying or handling pesticides and from post-application activities, such as harvesting crops previously treated with pesticides, or scouting in fields for pests after the field has been treated with pesticides.  In this appendix, EPA has analyzed how pesticide exposures of children performing agricultural post-application hand labor tasks are considered in its risk assessment process to ensure that EPA's regulatory decisions are health protective.  This analysis focuses on post-application dermal exposures to adults and children working in fields previously treated with pesticides which is known to be the predominant source of exposure for these types of working conditions.  EPA believes its quantitative post-application assessment for adults is protective of working children because considerable monitoring information supports the finding that pesticide exposures to children working in agriculture are highly likely to be much lower than exposures to adults working in agriculture based on a rigorous analysis of a wealth of exposure monitoring data with adults and children working in agriculture.  Additionally, the post-application assessment accounts for concerns regarding average body weight differences between children and the average body weight used in EPA's quantitative assessment for adults.  

This appendix also describes what The Environmental Protection Agency (hereafter referred to as "EPA" or "Agency") knows about the number of children currently engaged in farm labor activities, how such activities are considered in risk assessment, legal and other regulatory requirements associated with children working in agriculture, the methods used to assess the protectiveness of current risk assessment methodologies, the data used to complete this analysis, and the findings.  

This appendix assessed multiple approaches to evaluate its current risk assessment methodologies, including a mechanistic approach, an approach based on an extensive set of  exposure monitoring data collected under a variety of conditions involving both children as young as age 6 through adults up to 85 years old working concurrently in the same fields, and the processes involved in Agency hazard assessment and dose response characterization.  Additionally, the mechanistic approach also considers children of all ages.  The minimum age for farmworkers conducting non-hazardous farm labor is 12 years of age under the Fair Labor Standards Act (FLSA) but it is also clear that children less than 12 years old can be involved in labor activities in farms employing less than 7 people per calendar year (i.e., "family farms"), so they have also been considered.   EPA limits children younger than 18 from being pesticide handlers or early entry workers under the final revisions to the Worker Protection Standard (WPS; 40 CFR Part 170).  The final revisions include an exception from this age requirement only for persons covered by the owner and their immediate family exemption.  However, children less than 18 years of age (12 and above) may still be working in agriculture fields if they are either working on the same farm as a parent or person standing in the place of a parent, or working with parental permission.  Therefore, children can and are engaged in farm labor activities, and this analysis assesses the protectiveness of current post-application pesticide risk assessment methodologies.  It should also be noted that to the extent possible, EPA guidance on consideration of age groups was followed.  

The EPA has a longstanding commitment to Environmental Justice, which is the fair treatment and meaningful involvement of all people regardless of race, color, national origin, or income with respect to the development, implementation, and enforcement of environmental laws, regulations, and policies.  Concern for farmworkers and their families has been a major focus of the Agency's environmental justice efforts in its pesticide program.  EPA is also committed to protecting children's health.  Executive Order (E.O.) 13045 was issued in 1997 and applies to economically significant rules under E.O. 12866 that concern an environmental health or safety risk that EPA has reason to believe may disproportionately affect children.  EPA has also instituted a policy on evaluating risk to children as part of its risk assessments and decision making processes.  These concerns and commitments to children's environmental health have prompted EPA to further evaluate how pesticide exposures of children performing agricultural post-application hand labor tasks are considered in its risk assessment process to ensure that EPA's regulatory decisions are health protective.  However, this analysis does not address other topics related to environmental justice as it relates to farmworkers, such as the potential for take home exposures to pesticide residues or children being cared for by parents while working in fields, which are accounted for in other initiatives (e.g., updates to the Worker Protection Standard).  

As indicated above, EPA wished to evaluate how pesticide exposures of children performing agricultural post-application hand labor tasks are considered in its risk assessment process to ensure that EPA's regulatory decisions are health protective.  This analysis was conducted using a mechanistic analysis and extensive monitoring data.  

The mechanistic analysis was based on known factors about the physical attributes of children.  This analysis should be considered rigorous since the key source of mechanistic information was the EPA Exposure Factors Handbook (EFH).  The EFH is a compendium of nationally representative information about physical and behavioral characteristics of people at different lifestages  -  information that EPA often needs to use in risk assessments.  Specifically, the EFH provides information on a child's skin surface area and their body weight at different ages.  EPA examined how the ratio of surface area to body weight changes in children as they mature, since these attributes, among others, affect exposures to pesticides.  The findings of the mechanistic analysis demonstrate that, assuming other factors (such as level of pesticide residue, amount of time worked, and productivity) are equal, exposure assessments for adults are protective of children who are 12 and older because exposure is directly related to bodyweight and size, which is related to skin surface area.  The overall rate of chemical flux rate into the body through the skin is related to the surface area of the exposed individual.  The larger the person, the more skin surface area they have, resulting in more opportunity for greater total exposure.  When total dermal exposure is normalized for body weight to calculate dose, the resulting dose values are similar for adults and children down to about 12 years old.  This conclusion is based on the results of an analysis completed by Phillips et al.   Phillips et al established an empirical relationship between dermal surface area (SA) and body weight (BW) and illustrates that the relationship remains relatively unchanged with essentially a slope of (0) from about age 12 through adulthood; however, for children younger than 12, the relationship between surface area and body weight changes.  Despite that change, the findings of the mechanistic analysis demonstrate that based on body weight and size, the current risk assessment approaches are protective of the one exposure monitoring condition where total exposures were greater for children than for adults.  Children 12 years old and younger are inherently accounted for in the risk assessment methodology from a mechanistic perspective because they are known to be less efficient workers compared to adults and older children, which reduces their exposures relative to adults (as observed and discussed by the field investigators of the available exposure monitoring studies), even though for children younger than 12, the ratio between skin surface area and body weight is higher.  [Note: A total of 40 children under 12 years of age are included in the exposure monitoring studies; therefore, the conclusions by the field investigators are uniquely applicable to children younger than 12 years old].   

This conclusion that children are accounted for in the risk assessment methodology is supported by the observations of multiple study investigators from the monitoring studies looking at actual exposure information.  The investigators noted that a statically significant difference in productivity was detected between children and adults and that age and productivity (rather than surface area) exhibited significant correlations (full details in Appendix A.4).  

Some examples are provided below:  

 MRID 47923404: "A significant difference in productivity was detected between children and adults.  Age and productivity exhibited significant correlations.  Dermal exposure for children (14 years and younger) was lower than adults".  
 MRID 47923403: "In all three studies, methiocarb on blueberries, benlate on blackberries, and benlate on raspberries, a consistent relationship between age and exposure was observed.  When exposure is measured as total weight of pesticide deposited per hour worked, there is an increase of exposure with age...One study where work rate was available, provides evidence that differences in total exposure with age are attributable to differences in work rate."
 MRID 47923402: "Age and productivity of strawberry pickers appear to correlate positively with dermal exposure; age and productivity are also cross correlated.  Thus one may conclude that increasing age results in higher productivity (experience and motivation); higher productivity results in higher dermal exposure, and consequently, increasing age of pickers results in higher dermal exposure".  
 MRID 47923401: "The study was designed to provide correlations among dermal exposure, age of workers, body size or weight of workers and productivity.  Statistical analyses of the data showed a significant correlation between exposure rate and age and productivity.  Also age appeared to correlate positively with productivity, e.g., older workers seemed to receive higher dermal exposures than younger ones in the same occupational setting."

The second approach was a comprehensive statistical analysis conducted using observational worker exposure monitoring data collected over several years in the 1980s.  These data were generated via a joint funding effort between the EPA and the U.S. Department of Labor.  The research was conducted by 7 different universities in 17 distinct studies in various agricultural production areas of the country.  Specifically, worker monitoring was completed in 8 states in 11 crops for 16 different pesticides (including multiple classes of pesticides).  The individuals who participated ranged in age from 6 to 85 years old and a large percentage (54%) of the overall participants were children 6 to 18 years old.  All totaled, 87 unique exposure conditions were considered and 1472 work days were monitored.  [Note: only 66 of these conditions had adequate data to statistically analyze].  The monitoring data was statistically analyzed using comprehensive statistical models appropriate for the data collected in each study since the designs of the studies and the nature of the results of each varied.  This research was reviewed for compliance with federal ethical standards involving human subjects in research and found to be acceptable.[,]  
The findings of this analysis indicate, when comparing dermal exposures (mg/kg/day) between the two age groups, no statistical differences were observed between adults and working children in the vast majority of exposure conditions.  This means that the current risk assessment methods are protective of children's exposure, regardless of the nature of the study design or resulting data (e.g., the number of dermal sampling areas varied between studies).  In most of the remaining cases, adult exposure rates were higher than those of children, which also indicates that the current risk assessment methods account for children's exposures.  In four exposure conditions, exposures for children exceeded those for adults.  In one of the four exposure conditions, both hand total exposures were greater in children than adults; and in the other three exposure conditions, children exposures were greater than adult exposures for hands only.  Only total exposures (vs. hand or other exposures) are considered in occupational post-application human health risk assessment.  When total exposures are considered, children were found to have greater exposures than adults for only one exposure condition.  Therefore, the number of total exposure conditions where children's exposures are greater than those of adults, represent a very limited portion (2.4%) of the available data.  These conditions do not impact the overall conclusion that the current risk assessment method are highly like to account for children working in agriculture.  Because children were monitored along with adults, and found to have less exposure than adults in the majority (97.6%) of the exposure conditions, current risk assessment methods evaluating adult exposures is protective of exposures received by legally working children.  Even though children could receive exposures from activities other than post-application work (playing, etc), these exposures would be captured by the dosimetry methods because children as well as adults were monitored in the fields over the entire study duration, and usually the work day.  Any dermal contact received from treated foliage during that time would be captured by the dosimeters (i.e., gloves, hand patches, body patches), regardless of activity.  
In summary, the available lines of evidence support the Agency's overall conclusion that the current pesticide exposure and risk assessment methodologies that calculate exposures for adults working in treated agricultural fields are also protective of children performing the same kinds of work.  The researchers that originally conducted the exposure monitoring studies reached a similar conclusion.  It should also be noted these studies were conducted by multiple researchers and institutions, which reduces the potential for bias.  Also, the breadth of the exposure conditions that were encountered (87) in the exposure monitoring studies and the total number of exposure days monitored that were considered (1472) demonstrates the rigor of the analysis and its findings.  

           Introduction
         
Pesticide exposures in agriculture may occur while applying or handling pesticides and from post-application activities, such as harvesting crops previously treated with pesticides, or scouting in fields for pests after the field has been treated with pesticides.  This paper focuses on exposures related to post-application activities since the available monitoring data reflect hand harvesting of agricultural crops by both child and adult farmworkers.  This document evaluates how EPA considers children performing post-application work (hand labor tasks) in agriculture (for the purposes of this paper, referred to as child farmworkers) in the risk assessment process.  This paper describes what EPA knows about the number of children currently actively engaged in farm labor activities, how such activities are considered in risk assessment, legal and other regulatory requirements associated with children working in agriculture, the methods used to assess the validity of current risk assessment methodologies, the data used to complete this analysis, and the findings.  The analysis was completed using a mechanistic approach and with exposure data collected from both child and adult farmworkers.  All the exposure monitoring data included in this analysis have been evaluated and found to be compliant with Federal requirements for involving humans in research.  

This document includes the following:

 Section II, Problem Formulation: This section describes the historical development of methods for assessing risks to agricultural workers from occupational exposure to pesticides, information about children working in agriculture, the current regulations that are applicable to children working in agriculture, and the need to verify that EPA's pesticide risk assessment methodologies are protective of children working in agriculture.  

 Section III, Risk Assessment Primer: This section provides a brief discussion of how pesticide human health risk assessments for farmworkers are performed by EPA, with an emphasis on exposure assessment.

 Section IV, Analysis: This section describes the analyses completed to evaluate how children working in agriculture are considered in the risk assessment process.  The analyses rely on mechanistic data and monitoring data gathered from exposure studies which included children working in agriculture.   

 Section V, Discussion: This section describes the issues that need to be considered in interpretation of the analyses that were completed in order to evaluate the applicability of the current risk assessment methodology to child farmworkers.

 Section VI, Conclusions: This section summarizes the results of the mechanistic and monitoring analyses and discusses how the results may address regulatory and environmental justice concerns for child farmworkers. 

           Problem Formulation

The task of protecting agricultural workers and pesticide handlers from occupational exposure to pesticides is challenging, given the variability of pesticide use patterns and the diversity of tasks that are performed.  To protect both adults and children working in agriculture, EPA relies on the risk assessment and risk management process, and broader public health methods (e.g., providing tools to manage heat stress and musculoskeletal injuries, establishing mechanisms for improving hygiene, and providing guidance for reducing the occurrence of take home exposures).  This section discusses the consideration of environmental justice as it relates to agricultural workers including the development of exposure assessment methods, demographic information of farmworkers in the United States, and the current regulations that are applicable to children working in agriculture. These elements establish the importance of verifying that EPA's pesticide risk assessment methodologies are protective of children working in agriculture.
Farmworker Environmental Justice Movement
Many organizations have voiced concerns for farmworker health and safety from pesticide exposure over time.  To address these concerns, methods have evolved to address exposures in a more refined manner.  Some of the key historical events associated are summarized below.  
As early as the 1960s, researchers such as Durham and Wolfe recognized the need to protect those exposed to pesticides in the workplace and began to develop methods to quantify these exposures.  It was apparent, particularly after the introduction of organophosphate pesticides, that farmworkers were experiencing toxic effects related to exposures to some of those compounds; however, it was difficult to investigate the problem due to the nature of incidents.   Soon after, the farm labor movement was recognized and organized in the 1960s which lead to broader consideration of issues associated with farmworkers.  The movement's efforts, among other accomplishments, improved living conditions, raised public health awareness, described a need for fair wages, and raised occupational safety concerns.  
Occupational Pesticide Exposure Assessment History
In 1974, the President's Council On Environmental Quality (CEQ) established a Task Group which focused on occupational exposure to pesticides.  The CEQ identified a need to evaluate exposures from intensive field operations that involved significant contact with treated plants, such as harvesting or thinning fruit.  This evaluation of exposures led to the development and implementation of restricted entry intervals (REIs), which is the time needed for residues to dissipate before hand-labor activities can be completed without a risk concern.  Key recommendations from the CEQ included requiring worker monitoring data; evaluating geographical and other factors which influence exposures; establishing surveillance systems; and developing lower-risk pesticides as alternatives.  These recommendations stimulated activity in the area of occupational exposure assessment and led to the development of the first occupational exposure assessment guidelines.  
The development of the dislodgeable foliar residue (DFR) sampling method, which is a reproducible method for quantifying pesticide residues on the foliar surfaces was developed in the early 1980s.    The DFR is an indicator of the exposure source but it can be thought of as the amount of residue on the surface of the treated plant available for transfer to a worker's skin.  DFR values can be impacted by a number of factors including the crop, climate, or application method.  
Out of these efforts, a method that quantifies the relationship between field residues and exposures was developed.  This relationship is known as the transfer coefficient (TC).  TCs serve as the basis for how hand labor activities are evaluated in risk assessment.  The TC is an exposure metric that relates exposure for a particular hand labor task (measured as residues on a person or a person's clothing) to the amount of residues on the surface of the crop (measured as DFR levels), and the time worked to experience such exposure.  EPA can calculate exposures for risk assessment purposes using chemical-specific DFR values, an estimate of exposure duration (typically 8 hour workdays), and the TC associated with a particular activity that might be associated with how the chemical of concern is used in the crop.   
The Agency initially used a limited number of TCs for evaluating all farmworker exposures (e.g., fruit tree and strawberry harvesting TCs established by Popendorf and other researchers).   The Agency recognized that the approach of extrapolating TC data from only limited activities was insufficient to represent the large number of possible farmworker activities (e.g., scouting, pruning, harvesting other types of crops, etc).  Therefore, the Agency requested that industry provide a much more extensive database for use in risk assessment.  To this end, EPA developed exposure monitoring guidelines in 1984 (revised in 1998).  EPA sought independent peer review of these methods by the FIFRA Scientific Advisory Panel (SAP) in 1980, 1986, 1998, and 2007.  EPA also required that the regulated industry generate data that could be used to more rigorously quantify these exposures across all areas of agriculture based on the established guidelines.  EPA issued a Data Call-In (DCI) in 1995 which resulted in the formation of the Agricultural Reentry Task Force (ARTF).  The ARTF is a consortium of pesticide manufacturers whose purpose was to address the requirements of the DCI.  The DCI specifically targeted pesticide registrants whose products have the potential for post-application worker exposure.  The data generated by the ARTF provided the capability of evaluating occupational post-application exposures in a systematic manner.  
As early as 1980, EPA was funding research focusing on the potential exposures of child farmworkers that coincided with and was informed by the development of risk assessment methods described above.  EPA conducted this research jointly with the United States Department of Labor (DOL) through an interagency agreement and cooperative research agreements with several universities throughout the country.  The major focus of this research was to perform observational monitoring of farmworkers (children and adults) involved in post-application hand labor activities in many crops and regions.  Even though these data were derived from research conducted over 20 years ago, little has changed in terms of activities performed and many of the pesticides monitored are still used today.  As such, the studies still represent a scientifically credible source of information on exposure of both adult and child farmworkers to pesticides.  
Key Demographic Considerations of Farmworkers in the United States
EPA relies on several sources of information to better understand the composition of the farm labor community.  Based on data from the 2012 Census of Agriculture (National Agricultural Statistics Service (NASS), 2014), EPA estimates that over 300,000 farms employing 1.8 million workers and handlers used pesticides in 2012 (NASS, 2014).  
Another major source of this type of information is the DOL, which conducts an employment-based random-sample national survey of agricultural workers, the National Agricultural Workers Survey (NAWS).  The NAWS collects demographic, employment, and health data in face-to-face interviews.  The survey began in Federal Fiscal Year (FY) 1989 and includes information on over 56,000 agricultural workers ages 14 and up.  Additionally, EPA relies on state reporting of pesticide use where available (e.g., California's Pesticide Use Registry), incident reporting which is considered in risk assessment (e.g., see EPA's 2010 draft Framework for Incorporating Human Epidemiologic & Incident Data in Health Risk Assessment), information on crop production (e.g., crop profiles from the United States Department of Agriculture (USDA)), and input from stakeholders including growers, advocacy organizations, and pesticide manufacturers.
NAWS provides detailed information about farmworkers, including information about the number of children working in agriculture.  NAWS provides information on tasks in crop production (pre-harvest, harvest, and post-harvest), supervising workers, packing crops.  As shown in Figures 1 and 2, NAWS indicates that around 8 percent of respondents (from 2001 to 2009) bring children age 6-12 years old to the fields per year and around 10 percent bring children age 13 to 18 years old to the fields per year.  This data represents children who were brought to the field, however it is unknown how many of them were actually working in the field at the time. 
                                       

                                       

   







NAWS also indicates that horticulture production (garden plants, nursery plants, flowers, and fruits), field crop, and vegetable production appear to be major areas of agriculture where children work (Figure 3).  
Figure 3: Age vs. Crop for Fiscal Year 2008-2009 (as a % of respondents for each crop)

















Figure 4 provides additional information from NAWS related to the types of activities in which working children participate.  NAWS should be consulted for the specific definitions of each of the activities shown in Figure 4.  But, it is likely that children would most often engage in activities described as pre-harvest and post-harvest activities.  These would include tasks such as weeding, pruning, or packaging harvested commodities.  
   
Figure 4: Task by Age for Fiscal Years 2008-2009 (as a % of the total respondents for each task)

















There are legislative mandates which dictate how children can participate in the agricultural workforce as described in the FLSA of 1938.  Some highlights include:

 Children 16 years of age and above may engage in occupations deemed hazardous by the Secretary of Labor, which includes persons handling toxicity category I and II pesticides in agriculture; and
 Children 12 years of age and above may work in agriculture outside of school hours in non-hazardous jobs on farms if they are either working on the same farm as a parent or a person standing in the place of a parent, or working with parental permission.  

EPA limits children younger than 18 from being pesticide handlers or early entry workers under the final revisions to the Worker Protection Standard (40 CFR Part 170); the revisions include an exception from this age requirement only for persons covered by the owner and their immediate family exemption.

Regulatory Authority 

Pesticide regulation is dictated by several statutes, however the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) is most applicable to agricultural worker exposure.  All pesticides distributed or sold in the United States must be registered (licensed) by EPA.  Before EPA may register a pesticide under FIFRA, the applicant must show that using the pesticide according to labeled specifications "will not generally cause unreasonable adverse effects on the environment".  "Unreasonable adverse effects on the environment, as defined by FIFRA and applicable to post-application exposures, means: (1) any unreasonable risk to man or the environment, taking into account the economic, social, and environmental costs and benefits of the use of any pesticide...." 

Environmental Justice Considerations

Since E.O. 12898 was signed in 1994 by President Clinton and E.O. 13045 was issued in 1997, EPA has been striving to attain its goals related to environmental justice and children's health.[,]  EPA also instituted a policy on evaluating risk to children as part of its risk assessments and decision making processes.  In March 2000, the Government Accountability Office (GAO) also considered the occupational health of farmworkers and many of the issues raised in its report have been addressed or are being addressed by improvements in overall worker protection requirements and improved risk assessment methods.  The intent of this paper is to evaluate whether or not the current pesticide exposure assessment methods adequately account for the exposures of child farmworkers.  

In order to comply with its mandate and protect workers who may be exposed to pesticides, EPA relies on a multifaceted approach to protect agricultural workers who may be exposed to pesticides, and continues to strive towards the goals outlined in its environmental justice action plan[,].  To this end, EPA is involved in or has completed many activities aimed at improving living and working conditions as well as the overall health of farmworkers over time.  Some examples include:

 Implementation of the final revisions to the Worker Protection Standard (40 CFR 170), which provides broad protections for farmworkers, including establishing a minimum age of 18 years old for early-entry farmworkers and for handlers, with an exception only for persons covered by the owner and their immediate family exemption.

 Funding aspects of the NAWS such as recent questions focused on time and activity data and information on hygiene practices which allows for more informed risk assessment and risk management decisions.

 Considering and funding a variety of epidemiological research efforts to better understand potential health effects in impacted farmworkers and the surrounding communities in which they live (e.g., EPA/National Institute of Environmental Health Sciences Children's Environmental Health and Disease Prevention Research Centers).

 Training for staff on environmental justice and children's health issues in pesticide risk assessment.

 Development of focused training programs and outreach materials for impacted farmworkers which are accessible given known language barriers.

 Development of training materials and information for clinicians who practice in agricultural areas which aid them in recognizing pesticide related illness and incidents.

 Collaboration with poison information centers and providing more accessible means for reporting possible incidents.

  Use of Federal Advisory Committees such as the Pesticide Program Dialogue Committee (PPDC) to foster communication and collaboration amongst impacted stakeholders (e.g., manufacturers, farmworkers, advocates, grower/producers, public health personnel).

 Recent refinements to risk assessment methods for farmworker activities based on information collected via a broad data call-in that was also evaluated by the FIFRA Scientific Advisory Panel in 2008.

 Evaluation of the validity of current risk assessment methods with regard to how they consider risks for child farmworkers in the regulatory process, which is the focus of this analysis.

           Exposure and Risk Assessment Primer

There are a number of available resources which describe how occupational post-application (farmworker) pesticide exposure and risk assessments are conducted within EPA.  This section will provide a brief overview of how these assessments are completed and provide links where additional information is available.  
      
 Calculation of Transfer Coefficients

In the 2008 SAP review, EPA identified approximately 4,500 unique crop/activity combinations in production agriculture to evaluate in its post-application dermal exposure assessments.  Of these 4,500 crop/activity combinations, EPA completed exposure assessments on about 70% (or ~3,100) crop/activity combinations.  For these combinations, EPA used an approach to calculate post-application exposure for farmworkers based on the residue transferred from the target site (plant) to the farmworker for the particular hand labor task, which is reflected in the TC.  EPA does not, however, evaluate the other 30 percent of the crop/activity combinations (i.e., ~1,400 crop/activities), because exposures from these activities are considered negligible due to the highly mechanized nature of the activities, such as operating a combine during wheat harvest.  

TCs are used generically, regardless of the chemical being evaluated, based on a premise that they essentially represent an exposure rate based on dermal contact for a particular hand labor activity and crop which would occur regardless of the which pesticide happened to be present.  The generalizability of TCs is historically rooted in occupational pesticide exposure assessment and has been reviewed by the FIFRA SAP and supported in a detailed analysis completed by the ARTF (Bruce and Korpalski, 2008; Bruce et al., 2003) which was the basis of the SAP review.  The generic use of TCs can be further illustrated by the following example.  Consider two tomato fields for which the amount of chemical on treated leaf surfaces that can be rubbed off (i.e., can be dislodged) on the skin is the same.  One field has been treated with chemical A while the other field has been treated in a similar manner with chemical B.  If an individual harvests the same amount of tomatoes for the same amount of time in a day in each field, the individual would be expected to be subjected to similar exposure sources on the basis of this "genericness" premise.  The TC would also be expected to be similar for each field and chemical because the ratio of exposure per hour to dislodgeable foliar residue (DFR) would be the same.  If, on the other hand, this same individual were to perform another activity in those fields such as scouting for pests or vegetable tying, the resulting TCs would be different from that estimated for tomato harvesting because the different activities yield different exposures over the course of the day since they have a different amount of contact with the treated foliage.  
      
TCs are calculated based on actual exposure monitoring of individuals involved in a specific hand labor task.  The results obtained from worker monitoring studies via dermal dosimetry are used to determine total dermal exposure, which is then normalized per hour worked.  A ratio is then derived between these values for each worker and the DFR which is a representative concentration of the available surface residue on the crop surface available for transfer to an individual that may come in contact with the crop.  This ratio is the TC and is expressed in units of (cm[2]/hr).  Expressed mathematically, the calculation of a transfer coefficient is as follows:

TC (cm[2]/hr) = [Monitored Exposure (ug)  time worked (hours)]  DFR (g/cm[2])
      
The current WPS provides training requirements for farmworkers which directs them to wear clothing that protects the body from pesticide residues which is typically long sleeved pants, long sleeved shirts, shoes, and socks.  Typically, the monitored exposure used to calculate TCs is based on total body exposure under normal work clothing which is considered to be long pants and long sleeved shirts.  In some studies, exposure monitoring has been conducted using different combinations of dosimeters which allows for flexibility in the presentation of results (e.g., different combinations of clothing). An example of a TC calculation is provided in Table 1.  In this case, subjects were monitored on their face/neck and hands, and all body regions under long sleeved shirts and long pants during tobacco harvesting.   Both the arithmetic mean and geometric mean DFRs are presented in Table 1.  The arithmetic mean is used as the basis for Agency risk assessments.

Table 1:  Tobacco Harvesting  -  Transfer Coefficient Summary[a]
                                Days After 2nd
                                     App.
                                   Worker ID
                                  Time Worked
                                    (hours)
                                 Exposure (ug)
                                DFR (ug/cm[2])
                        Transfer Coefficient (cm[2]/hr)
                                       
                                       
                                       
                             Total Inner Dosimeter
                                     Hands
                                     Face/
                                     Neck
                                  Arith. Mean
                                     Geo.
                                     Mean
                           Based On Arith. Mean DFR
                                 Based On Geo.
                                   Mean DFR
                                       1
                                       A
                                     6.08
                                     30150
                                     3906
                                     24.4
                                     5.03
                                     4.88
                                     1110
                                     1150
                                       1
                                       B
                                     6.08
                                     28784
                                     3334
                                     27.1
                                     5.03
                                     4.88
                                     1050
                                     1080
                                       1
                                       C
                                     6.08
                                     35670
                                     4395
                                     69.3
                                     5.03
                                     4.88
                                     1310
                                     1350
                                       1
                                       D
                                     6.08
                                     26080
                                     2943
                                     18.5
                                     5.03
                                     4.88
                                      950
                                      979
                                       1
                                       E
                                     6.08
                                     32820
                                     4013
                                     19.3
                                     5.03
                                     4.88
                                     1200
                                     1240
                                       1
                                       F
                                     5.05
                                     38140
                                     2672
                                     30.4
                                     5.03
                                     4.88
                                     1610
                                     1660
                                       1
                                       G
                                     6.08
                                     37980
                                     2887
                                     72.3
                                     5.03
                                     4.88
                                     1340
                                     1380
                                       1
                                       H
                                     6.08
                                     27639
                                     3086
                                     21.4
                                     5.03
                                     4.88
                                     1010
                                     1040
[a] Based on the sum of exposure across all body regions (Total Inner Dosimeter + Hands + Face/Neck). 
Transfer Coefficient (cm[2]/hr) = [Exposure (g)  time worked (hours)]  DFR (g/cm[2]).

 How Transfer Coefficients (TCs) Are Incorporated Into Risk Assessments 

The goal in Agency pesticide post-application human health risk assessments is to ensure that a sufficient period of time has elapsed between application and worker reentry into treated areas such that exposures, through contact with foliage associated with a hand labor activity, are below the Agency's threshold of concern.  This construct is formally included on pesticide labeling as REI which codifies such durations.  The REI is developed based on exposure assessment results, the results of the acute toxicity studies, and consideration of the risks relative to the benefits associated with its use as required under FIFRA.  The Agency defines the timeframe it takes for residues to decline to acceptable levels for those who enter previously treated fields to work by coupling information about the environmental levels and dissipation rate of a chemical (i.e., through DFRs) with the TCs for the suite of activities and crops that are associated with the use of a particular pesticide.  

As discussed above, the Agency has identified ~3,100 non-mechanized crop/activity combinations for specific hand-labor tasks which occur in production agriculture.  To facilitate consistency, the Agency developed a series of general scenario-based descriptions for tasks in order to categorize them for risk assessment purposes.  Common examples include: harvesting, scouting, crop maintenance tasks (e.g., irrigating, hoeing and weeding), and turf maintenance (golf course mowing and sod harvesting).  Further categorization of potential exposures was based on agronomics and the types of crops as discussed at the 2008 FIFRA SAP review of the methodology used by EPA; examples of these categories used to group TCs are shown below:

 Field/row crops (e.g., beans, corn, peanuts, peas, sorghum, sunflowers),
 Tobacco,
 Cut flowers (e.g., floriculture crops),
 Orchard crops (e.g., apples, pears, citrus),
 Vine/trellis (e.g., blackberries, blueberries, grapes, kiwi, raspberries), and
 Nursery crops (e.g., container, ball, and burlap ornamentals).

Within each agronomic group, a variety of cultural practices are required to maintain the included crops.  These practices are varied and typically involve light to heavy contact with immature plants as well as with mature plants.  Due to the variety in the hand labor tasks required to produce these crops, different TCs are used to predict the range of exposures within each agronomic group.  This hypothetical matrix of agronomic groupings and associated TCs within each group represents what can be thought of as a library of TCs. 

Once a TC is chosen for risk assessment based on the proposed use pattern or use on a specific crop and expected activity, the next step in the risk assessment process is to calculate dermal exposure (DE).  The TC can be used to assess any similar task performed on a similar crop for which DFR data are available using the following equation:

     DE (mg/kg/day) = [(DFR (ug/cm[2]) x TC (cm[2]/hr) x Hr/Day]/BW (kg)

Where:

      DE (t, task) = Daily exposure or amount deposited on the surface of the skin at time (t) attributable for activity in a previously treated area, also referred to as potential dose (mg ai/day); 
      
      DFR = Dislodgeable foliar residue or the amount of pesticide residue available for transfer to skin from foliar contact based on the Iwata et al sampling method (ug/cm[2]);
      
      TC = Transfer coefficient or an exposure rate which is based the conceptual contact of a worker with treated foliage (cm[2]/hr);
      
      Hr/Day = Hours per day workers are assumed to be engaged in tasks for risk assessment purposes; and
      
      BW = Body weights used to represent a specific group being considered in risk assessment (kg).

Once normalized daily exposure values are calculated, risks (i.e., typically expressed as margins of exposure or MOEs) are estimated by comparing exposure levels to the point of departure (e.g., a no observed adverse effect level (NOAEL) determined in an appropriate animal toxicity study) using the following equation:  

         MOE(dermal, t,task) = POD (mg/kg/day)/DE(t, task) (mg/kg/day)

Where:

      MOE (dermal, t, task) = Margin of exposure; used by the Agency to represent risk or how close a chemical exposure is to being a concern (unitless).  These values are calculated based on an exposure estimate for each day after application and for each task of concern;
      
      DE (t, task) = Daily exposure or amount deposited on the surface of the skin at time (t) attributable for activity in a previously treated area, also referred to as potential dose (mg ai/day); and
      
      POD = Point of departure determined by the toxicity of a chemical, typically the dose level in a toxicity study, where NOAEL in that study or the lowest dose level where an adverse effect occurred (LOAEL) in the study (mg pesticide active ingredient/kg body weight/day).

 Dermal Exposure Rate Calculation Using Harvester Exposure Monitoring Field Studies  

It is important to understand how post-application (farmworker) exposures are calculated to understand the analysis which was completed to evaluate whether or not the current pesticide exposure assessment methods adequately account for the exposures of children working in agriculture.  The agency analyzed the available information used for this analysis in a similar way to that just described; this is because of the intent of the studies and the methods used by the investigators in the research.  Exposure monitoring data from the Pesticide Hazard Assessment Project: Harvester Exposure Monitoring Field Studies (1980-1986) were used to calculate dermal doses (mg/kg/hr) of agricultural harvesters who were either adults or working children (those <18 years old).  The individual studies are described further in Section IV.B and in the appendices of this document.  In each study, exposure monitoring data were collected as amount of active ingredient per glove and/or dermal dosimeter.  The dosimeters were located across the body of the monitored farmworkers in varied fashion, depending upon the design of a particular study, but usually on the arms, chest, and/or back. The worker monitoring techniques used in this study were considered the standards of the time and followed the appropriate, then current, guidance.  Dermal exposures (the focus of this analysis) were monitored using the patch method which is illustrated in Figure 5.  Patches were made of cotton gauze or alpha cellulose.  Exposures to the hands were also measured but with a variety of techniques including cotton gloves collected for analysis or a removal method such as a hand wash using alcohol or aqueous surfactant of some sort (typically dilute soap solution).  Agency guidelines are also available which document the detailed methods used for worker monitoring should more information be desired.  The project also considered appropriate analytical quality control measures and as such, the values which were reported by the investigators were used as the basis for this analysis (e.g., appropriate analytical recovery and stability measures).  No further data corrections or adjustments were made to the reported results.
       Figure 5: Illustration of Patch Placement When Full Body Sampling


The studies included in this research represent 87 distinct exposure conditions (e.g., studies used various combinations of patch locations); all were completed using similar monitoring techniques (i.e., patches with no clothing protection and a hand sampling method  -  both represent exposure to bare skin).  However, different investigators used disparate patch sampling strategies as theoretically illustrated in Figure 5 (i.e., patches were not always placed onto the same part of the body in each study).  As such, the illustration represents 3 different theoretical sampling strategies which could have been employed because of the preferences of the investigators.  It should also be noted that the monitoring approach was opportunistic in that investigators first had to gain access to each prospective study location and then monitor essentially who was working that particular day.  Because this is the case, the subjects who were monitored at each site varied (number of children of each age and the ratio of children to adults varied).  These factors (i.e., sampling disparity and the opportunistic approach to sampling) guided the statistical models and methods used to analyze the data.  Data were not combined between different studies for a variety of reasons; therefore, unless described below, all results were analyzed within each unique exposure condition.  The reasons for not combining data for analysis across different exposure conditions are described below for each study so it will be clear how the data were collected and any caveats that should be considered in the interpretation of the results for each specific study.
                                       
Figure 6: Theoretical Illustration of Disparate Patch Placement (Study Design) in Worker Exposure Monitoring

The patches and gloves are referred to as "monitors" in the summaries below and yield information on the quantity of pesticide residue would have been deposited on the participant's clothing (or on the skin surface, in the absence of clothing) for that region of the body where the monitors are worn.  Exposures to the participant's hands were measured in each study; therefore, there was always a hand exposure rate calculated for all participants unless no residues were measured.  The dermal exposure rate for hands was specifically calculated using the formula below as available and appropriate.  Hand data were not corrected for surface area of the body part if gloves or rinses were used for sampling because they represent the entire surface area of the hands.  However, if there were patches on the hand instead of gloves, the data would be corrected for the surface area of the hand.  

             Hand Dose (mg) = [Left Hand (mg) + Right Hand (mg)] 

Exposure to other areas (arms, chest, back, etc.) were opportunistically calculated when available, and included in the total dermal exposure for each participant.  If multiple regions of the body were monitored, a total dermal exposure rate was calculated using the hand data along with the exposure rates from the other monitored body parts.  Although this combined exposure is referred to as "total" exposure, it only represents the portions of the body where samples were located; however, results for children and adults would always have the same number of patches in the same locations in each study and when  considered in the analysis. The data collected for each body location, except for hands, were corrected for the surface area of the patch (SA Patch) and the surface area of each body location using surface area values from the Exposure Factors Handbook (1997).  The relevant surface area values for adults and children are provided in Appendix A.1, and were incorporated into the dose calculations depending on the gender and age of each participant.  The example calculation illustrates how results from a participant who wore patches on both their right and left arms would be calculated.   

                                Arm Dose (mg) =
      
[Left Arm patch (mg) + Right arm patch (mg)] x [SA arm (m[2]) / SA patches (m[2])]

This dose would then have been normalized for the amount of time the participant worked (i.e., exposure was monitored) and each participant's body weight to yield the dose estimate (mg/kg/hr).  

     Arm Dermal Dose (mg/kg/hr) = Arm Dose (mg) / BW (kg) / Hours Worked

The total dermal dose rate for each participant was calculated incorporating all body regions monitored into a combined total dose rate (mg/kg/hr). 

                        Total Dermal Dose (mg/kg/hr) =
      
Glove dose (mg/kg/hr) + Arm Dose (mg/kg/hr, where available) + Chest Dose (mg/kg/hr, where available).... etc.

In summary, the Agency calculated dermal exposure rates for both adult and children agricultural harvesters using data available in the Pesticide Hazard Assessment Project: Harvester Exposure Monitoring Field Studies (1980-1986).  The nature of the calculations varied depending upon how the study was conducted.  These dermal exposure rates for children and adult workers were then statistically compared as described below in Section IV.B.ii.  


           Analysis

This section describes how the exposures of child farmworkers were quantitatively evaluated.  Two approaches were considered.  The first (Section IV.A) uses a mechanistic approach and  known factors about the physical attributes of children, the relationship between their skin surface area and their body weight as they age, to evaluate current risk assessment methods and how they may apply to children of various ages.  The second (Section IV.B) uses monitoring data which were collected through the joint EPA/Department of Labor research project noted above which observationally measured exposures of children performing post-application hand labor tasks in a variety of crops.

 Mechanistic Approach

The predominant source of exposure for post-application farmworkers is from dermal contact with treated plants.  This has been demonstrated in many monitoring studies conducted for farmworkers across production agriculture. Given the predominant route of exposure and the information that will be discussed below, EPA believes that assuming adult inputs of body size and weight as well as daily work time is protective of children who are legally working.  This can be illustrated mechanistically using the surface area of the body related to age, especially for children who are 12 and older, based on empirical information about how body weights relate to age.  Children who are less than 12 years old are anticipated to account for a very small percentage of legally working children because of criteria outlined in the Fair Labor Standards Act and the information described above in Section II from the NAWS (less than 10% per year).  

The overall rate of chemical flux rate into the body through the skin is related to the surface area of the exposed individual.  The larger the person, the more skin surface area they have, resulting in more opportunity for greater total exposure.  When total dermal exposure is normalized for body weight to calculate dose, the resulting dose values are similar for adults and children down to about 12 years old.  This conclusion is based on the results of an analysis completed by Phillips et al.   Phillips et al established an empirical relationship between dermal surface area (SA) and body weight (BW) and illustrates that the relationship remains relatively unchanged with essentially a slope of (0) from about age 12 through adulthood (Figure 7).  In fact, the correlation coefficient was 0.986 calculated in that research for surface area versus body weight based on a set of 401 observations.  Phillips et al further demonstrates that this ratio changes rapidly from birth to children who are about 12 years old which essentially agrees with the widely recognized phenomenon associated with growing younger children.  To reiterate, since high correlation has been demonstrated in the SA/BW ratio and the slope of the line does not change in ages > 12 years it is reasonable to assume that normalized dermal doses for anyone >12 years old would be similar with all other exposure factors being equal.


12 years of age
12 years of age
            Figure 7: Surface Area to Body Weight Ratio versus Age
A second analysis was also completed using information presented in the EPA's 2011 Exposure Factors Handbook where a comparison of SA/BW ratio for adults and children was calculated based on more recent data (from the National Health and Nutrition Examination Survey; NHANES data up to 2006) than was used previously.  The results of the analysis support the conclusions by Phillips et al.  Therefore, these two analyses indicate that children's surface area to body weight ratios are very similar to those for adults and, in fact, vary by no more than 12 percent from those of adults as illustrated in the last column of Table 2.

Table 2.  Comparison of Adult and Child Surface Areas, and SA/BW ratios
                                   Age Group
                              Surface Area (m[2])
                               Ratio to Adult SA
                               Body Weight (kg)
                               SA/BW (cm[2]/kg)
                             Ratio to Adult SA/BW
                                     Males
                                     Adult
                                    2.07[a]
                                      1.0
                                     86[c]
                                    0.0241
                                      1.0
                                   12 years
                                    1.23[b]
                                     0.59
                                    50.4[d]
                                    0.0244
                                     1.01
                                   14 years
                                    1.60[b]
                                     0.77
                                    63.9[d]
                                    0.0250
                                     1.04
                                   16 years
                                    1.66[b]
                                     0.80
                                    74.4[d]
                                    0.0223
                                     0.93
                                   18 years
                                    1.91[b]
                                     0.92
                                    75.6[d]
                                    0.0253
                                     1.05
                                    Females
                                     Adult
                                    1.82[a]
                                      1.0
                                     69[c]
                                    0.0264
                                      1.0
                                   12 years
                                    1.28[b]
                                     0.70
                                    52.0[d]
                                    0.0246
                                     0.93
                                   14 years
                                    1.55[b]
                                     0.85
                                    59.9[d]
                                    0.0259
                                     0.98
                                   16 years
                                    1.47[b]
                                     0.81
                                    63.0[d]
                                    0.0233
                                     0.88
                                   18 years
                                    1.69[b]
                                     0.93
                                    65.2[d]
                                    0.0259
                                     0.98
    [a]Tables 7-12 & 7-13, Exposure Factors Handbook (EPA, 2011).
    [b]Table 7-10 and 7-11, Exposure Factors Handbook (EPA, 2011). Various percentile values for males and females were selected to represent different ages in each bin (i.e., 10[th] percentile for 12 and 16 year olds and 50[th] percentile for 14 and 18 year olds from their respective bins).
    [c]Default value used by EPA, Office of Pesticide Programs.
    [d]Table 8-14, Exposure Factors Handbook (EPA, 2011). Mean of mean values for males and females using NHANES 1999-2002 Data.


 Monitoring Data

EPA, in conjunction with the U.S. Department of Labor, funded an extensive monitoring program focused on children working in agriculture referred to as the Pesticide Hazard Assessment Project: Harvester Exposure Monitoring Field Studies (1980-1986).  The report from this project consists of a collection of studies that were conducted by university researchers across the country in areas of major agricultural production.  These researchers concluded, overall, that there were no significant differences in pesticide exposures between child and adult farmworkers.  The results of these studies have been re-evaluated in this analysis in order to draw independent and holistic Agency conclusions using the data, and to review the conclusions of the researchers in conjunction with the Agency conclusions.  Section IV.B.i below provides an overview of the research program.  Section IV.B.ii presents a summary of each of the studies considered in this analysis and the statistical findings based on the results.  Section IV.B.iii presents the conclusions derived based on this analysis.
             
                  Overview of Studies

The monitoring data which serves as the basis for this analysis is extensive and covers a variety of potential field exposure situations during hand harvesting.  Although collected over 30 years ago, the data remain relevant for current exposure assessment methodologies and to modern production agriculture since the crops included in the studies are still primarily hand harvested in current production practice.[,]  

Investigators from seven academic institutions conducted this research, including: 
 Mississippi State University, 
 Medical University of South Carolina, 
 Colorado State University,
 University of Iowa,
 University of California, Berkeley,
 Texas Tech University, and 
 University of Florida.
These studies were conducted throughout the United States; including in: 
 California, 
 Mississippi, 
 Michigan, 
 North Carolina, 
 South Carolina, 
 Oregon, 
 Texas, and 
 Wisconsin.  
Hand harvesting was the primary activity monitored throughout this research.  The following crops were hand harvested: 
 Apples, 
 Blackberries, 
 Blueberries, 
 Corn, 
 Cucumbers, 
 Peanuts, 
 Peas, 
 Raspberries, 
 Strawberries, 
 Tobacco, and 
 Tomatoes. 
A total of sixteen pesticide active ingredients were quantified during the conduct of this study.  The following pesticides (names are of the active ingredient) were monitored in the exposure media: 
 Acephate, 
 Azinphos methyl, 
 Benomyl, 
 Captafol, 
 Captan, 
 Carbaryl, 
 Chlorothalonil, 
 Endosulfan, 
 Malathion, 
 Methamidophos, 
 Methiocarb, 
 Methomyl, 
 Methyl parathion, 
 Toxaphene, 
 Trifluralin, and 
 Vinclozolin. 
Since the study's completion, some of the monitored pesticides have been removed from the market.  At the time of the study, however, they were all commercially available and broadly used in agriculture.  
The design of individual monitoring studies varied because of the crop, what pesticides had been applied to the crop, and the opportunistic approach used by the investigators to collect samples observationally.  In some cases, a single pesticide was applied and subjects were monitored on one or more days after the application.  In other cases, multiple pesticides had been applied to fields and samples were collected which provided information from the same monitoring day that represented more than one active ingredient.  Samples were also collected in the days just after application up to several weeks and even months after pesticide application.  Given all of the factors described above, samples were collected in 87 distinct exposure situations where each exposure condition was defined by a number of factors including crop, pesticide, days after application.  For each distinct exposure situation to be utilized in this analysis, both adults and children had to be working concurrently in order to compare exposure rates between the two age groups.

An exposure day (also commonly referred to as a monitoring unit or MU) represents the samples collected from one individual from a single work period, which ranged from a few hours to several hours in each study.  In some cases, an individual's samples were analyzed for more than one pesticide; in those cases results from more than 1 pesticide from the same workday were considered as separate results which makes accounting for all of the permutations of the exposure situations encountered easier to consider.  All totaled, the number of exposure days (hereafter referred to as MUs) included in this analysis was 1472.  This amount of information is similar to that which is available for the surrogate databases used by EPA for routine risk assessment (e.g., the ARTF, the Agricultural Handlers Exposure Task Force; AHETF, and the Pesticide Handler Exposure Database; PHED).  As noted briefly above, the basis of this analysis is a comparison of the exposure rates of children working along-side adults under the same exposure condition.  Figure 8 below illustrates the age distribution in each of the studies.  Ages ranged from 6 years to 85 years old.


     Figure 8: Age Distribution of Youth In Agriculture Study Participants
 

                  Analysis of Individual Studies

Each study within this analysis has been assigned a code in order to facilitate tracking (1 to 17) and the data from each study site has been assigned a "Final Data Tab ID" which can be used to differentiate them (Appendix A.2).  The actual summarized data used to complete the analysis are included in Appendix A.2 with the data for each site in separate excel workbook tabs (Appendix A.2) which also links to the primary study ID.  Each tab may also contain information for more than 1 exposure condition (monitoring on different days after application).  Each tab also includes a color coding system that explains how the specific calculations were completed.  [Note: Separate PDF files for each study have been created.  They can be provided upon request.  Detailed statistical analysis files (SAS code or JMP files) can also be provided upon request.  Appendix A.3 contains a glossary of important statistical terms and concepts useful for interpreting the results of this analysis.]

Appendix A.4 summarizes the design of each individual study (e.g., pesticide application information, location, and activities monitored), outlines the samples that were collected, provides a profile of the participants, provides a summary of the statistical findings, and outlines any major issues for consideration.  As available, the findings of the investigators are also included for comparative purposes.  Calculations using raw data reported by investigators have been completed as described above (see Section III).  In these studies, hands were sampled and a limited number of patches were collected on a few body regions like the forearms and chest; these patches were typically not under clothing.  [Note: placement of patches over or under clothing does not impact the analysis because relative exposures for children and adults were considered for each individual exposure condition].  In some studies dislodgeable foliar samples and/or biological samples were also collected.  These will be noted but no analysis has been completed using this information.  Possible future work could be completed using these data in some circumstances.

Given the considerations above, the approach used in this analysis was to define exposure rates for each MU for each pesticide monitored (sometimes results were reported for more than 1 pesticide from the same sample); always for the hands and then also for total dermal exposures, which would include the hands and a sum of all other sources of exposure calculated from the other patches worn.  These exposure rates were then statistically evaluated, as appropriate, for each distinct exposure condition to determine if the monitored working children were receiving more, less, or equivalent exposures as adults while doing the same task.  Although the sample sizes were small in some studies, the Agency is confident that the data analyzed provided enough statistical power to draw conclusions from the data.  Because so many exposure conditions exist within the data, the analysis focused first on exposure rates within individual study conditions.  However, when the same individuals were monitored in each sampling event and if the exposure conditions were similar, the data were combined within studies as appropriate.  Information from Study 5 is included below as an example of the analysis of individual studies.  The summary, statistical analyses, and findings from all studies are presented in Appendix A.4.

 Study Example  -  Pesticide Exposure to Blueberry Harvesters in Southwestern Michigan, 1984 (MRID 47928906)

An example analysis is outlined in this subsection to demonstrate how exposure rates were generated using specific study monitoring data.  This example illustrates the complexities when analyzing the data and generating exposure rates and statistically analyzing appropriate subpopulations and age groups; only part of the study is described in detail here for illustrative purposes.  Analysis of the entirety of the data can be found in Appendix A.4.  In this example, (Study 5) the study was conducted by the University of Iowa (August 1984) on two farms near South Haven Michigan (i.e., referred to as VK, and GC, respectively).  Workers were monitored while participating in harvesting activities in fields that were treated previously with several pesticides; including malathion (the primary focus of this study), azinphos methyl, terbacil, simazine, triforine, captafol, and methiocarb (Table 3).  Exposures were only quantified for four of the applied pesticides: malathion, azinphos methyl, methiocarb, and captafol.  A total of 42 individuals were monitored during this study but all were involved in multiple monitoring days (Table 3).  Pesticide applications across all of the fields used in this research occurred between April 4, 1982 and July 20, 1982.  Monitored harvesting activities occurred between July 20, 1982 and July 29, 1982 which equates to the same day as a pesticide application up to 83 days after application for the pesticides where exposure was quantified.

Cotton gloves and gauze patches worn on the back, chest, forearm, and shoulder and were used as pesticide exposure monitors.  Table 3 below summarizes the sampling completed at each study site.  Table 3 also contains the names of the data tabs where the specific results for each field reside (Appendix A.2).  

Table 3:  Study Site Summary for Field Study 5, Malathion Blueberry Harvesting in SW Michigan.
                                Study Site/Day
                  Pesticide & Appl. Rate        (lb ai/A)
                           Harvest Days After Appl.
                                 Appendix A.2
                                   Data Tab
                            Worker Characteristics
                                       
                                       
                                       
                                       
                                     Min.
                                      Age
                                     Max.
                                      Age
                              N <18 years old
                               N>= 18 years old
                            Total Monitoring Units
                                     VK/I
                             Azinphos Methyl (0.5)
                                      44
                               (5).VKFieldStudyI
                                       9
                                      62
                                      10
                                      10
                                      240
(Same workers monitored on 3 different days in field treated with 2 pesticides and also monitoring for 2 other pesticides)
                                     VK/II
                                       
                                      45
                              (5).VKFieldStudyII
                                       
                                       
                                       
                                       
                                       
                                    VK/III
                                       
                                      46
                              (5).VKFieldStudyIII
                                       
                                       
                                       
                                       
                                       
                                     VK/I
                                   Malathion
                                  (2, 2, 0.6)
                                   32, 19, 0
                               (5).VKFieldStudyI
                                       
                                       
                                       
                                       
                                       
                                     VK/II
                                       
                                   33, 20, 1
                              (5).VKFieldStudyII
                                       
                                       
                                       
                                       
                                       
                                    VK/III
                                       
                                   34, 21, 2
                              (5).VKFieldStudyIII
                                       
                                       
                                       
                                       
                                       
                                    GC-1/IV
                                   Malathion
                                      (1)
                                      26
                             (5).GC-1FieldStudyIV
                                      10
                                      46
                                       9
                                       9
                                      162
(Same workers monitored on 3 different days in the same field treated with 3 pesticides, all 9 were same individuals at GC-2)
                                    GC-1/V
                                       
                                      27
                              (5).GC-1FieldStudyV
                                       
                                       
                                       
                                       
                                       
                                    GC-1/VI
                                       
                                      28
                             (5).GC-1FieldStudyVI
                                       
                                       
                                       
                                       
                                       
                                    GC-1/IV
                                  Methiocarb
                                     (1.5)
                                       6
                             (5).GC-1FieldStudyIV
                                       
                                       
                                       
                                       
                                       
                                    GC-1/V
                                       
                                       7
                              (5).GC-1FieldStudyV
                                       
                                       
                                       
                                       
                                       
                                    GC-1/VI
                                       
                                       8
                             (5).GC-1FieldStudyVI
                                       
                                       
                                       
                                       
                                       
                                    GC-1/IV
                                Captafol (2, 2)
                                    59, 39
                             (5).GC-1FieldStudyIV
                                       
                                       
                                       
                                       
                                       
                                    GC-1/V
                                       
                                    60, 40
                              (5).GC-1FieldStudyV
                                       
                                       
                                       
                                       
                                       
                                    GC-1/VI
                                       
                                    61, 41
                             (5).GC-1FieldStudyVI
                                       
                                       
                                       
                                       
                                       
                                   GC-2/VII
                                  Methiocarb
                                     (1.5)
                                      14
                             (5).GC-2FieldStudyVII
                                      10
                                      46
                                      11
                                      11
                                      132
(Same workers monitored on 3 different days in the same field treated with 2 pesticides, 9 of 11 were same individuals at GC-1)
                                   GC-2/VIII
                                       
                                      15
                            (5).GC-2FieldStudyVIII





                                    GC-2/IX
                                       
                                      16
                             (5).GC-2FieldStudyIX





                                   GC-2/VII
                                Captafol (2, 2)
                                    67, 47
                             (5).GC-2FieldStudyVII





                                   GC-2/VIII
                                       
                                    68, 48
                            (5).GC-2FieldStudyVIII





                                    GC-2/IX
                                       
                                    69, 49
                             (5).GC-2FieldStudyIX





Notes:
VK/1 through VK/III:  Malathion was applied 3 times in this field.  Methiocarb and captafol residues were also quantified in the worker media even though no applications of these pesticides were noted at this site.  Triforine applications were also described in the study report but worker media were not analyzed for that pesticide.

GC-1/IV through GC-1/VI:  Captafol was applied 2 times in this field.  Terbacil, simazine, triforine, and azinphos methyl applications were also described in the study report but worker media were not analyzed for these pesticides.

GC-2/VII through GC-2/IX:  Methiocarb was applied in this field and captafol was applied twice.  Terbacil, simazine, triforine, azinphos methyl, and malathion applications were also described in the study report but worker media were not analyzed for these pesticides.


A total of 27 exposure conditions are represented in these data since exposures were monitored at 3 different sites, for multiple pesticides, and on differing days after application.  Urine, dislodgeable foliar residue, air, and soil samples along with productivity measures were also collected in this study.  For each exposure condition where quantifiable results were reported, and statistical analysis has been completed.  In some cases, the data were combined across exposure conditions as applicable.  These are noted for each case.  [Note: As a reminder, please refer to Appendix A.3 for a glossary of important statistical terms and concepts useful for interpreting the results of this analysis].

VK I-III Results For Malathion:  The available data for the VK/I-III sites for malathion have been summarized using box and whisker plots using hand exposure rates (mg/kg/day) (Figure 8) and total exposure rates (mg/kg/day) (Figure 9).  Total exposure rates represent all patch exposure locations monitored on the body, including hands.  The same workers were monitored over three different days; therefore, a mixed model was used to account for the repeated measurements of the same worker for multiple days.  The exposure rates were log transformed prior to analysis.  

Age-group (children vs. adult) and exposure day were used as the covariates in the model.  The differences of least square means between adult and children age group were estimated from the model and the results are presented in Table 4 for hand exposure rates and in Table 5 for total exposure rates.  

The conclusions associated with malathion exposure at the VK/I-III site are there is no statistically significant difference between adult and children exposure rates for hand/glove exposure rates.  However the adult age group had statistically significantly higher total malathion exposures rate than monitored children (one sided p value < 0.05).

Figure 9: Box & Whisker Plot, Malathion Hand Exposure Rates (mg/kg/hr)  for Blueberry Harvesters, VK I-III
                                       
 
Table 4: Differences of Least Squares Means and Least Squares Means,  Malathion Hand Exposure Rates for Blueberry Harvesters, VK I-III
                                    Effect
                                   Age Group
                                   Age Group
                                   Estimate
                                Standard Error
                                     DF[1]
                                  t Value[2]
                                Pr > |t|[3]
                                   Alpha[4]
                                   Lower[5]
                                   Upper[6]
                      Differences of Least Squares Means
                                   Age Group
                                    Adults
                                   Children
                                   -0.04881
                                    0.2071
                                      18
                                     -0.24
                                    0.8163
                                     0.05
                                    -0.4839
                                    0.3863
                              Least Squares Means
                                   Age Group
                                    Adults
                                    -3.8643
                                    0.2425
                                      18
                                    -15.93
                                   <.0001
                                     0.05
                                    -4.3738
                                    -3.3548
                                   Age Group
                                   Children
                                    -3.8155
                                    0.2425
                                      18
                                    -15.73
                                   <.0001
                                     0.05
                                    -4.3250
                                    -3.3060
 DF = Degrees of Freedom 
 T Value = Test Statistic
 Pr > t =  p value
 Alpha =  Significance level 
 Lower = Lower 95% Confidence Limit 
 Upper =  Upper 95% Confidence Limit 

Figure 10: Box & Whisker Plot, Malathion Total Exposure Rates (mg/kg/hr) for Blueberry Harvesters, VK I-III
                                       
 
Table 5: Differences of Least Squares Means and Least Square Means, Malathion Total Exposure Rates for Blueberry Harvesters, VK I-III
                                    Effect
                                   Age Group
                                   Age Group
                                   Estimate
                                Standard Error
                                     DF[1]
                                  t Value[2]
                                Pr > |t|[3]
                                   Alpha[4]
                                   Lower[5]
                                   Upper[6]
                      Differences of Least Squares Means
                                   Age Group
                                    Adults
                                   Children
                                    0.3854
                                    0.1301
                                      18
                                     2.96
                                    0.0083
                                     0.05
                                    0.1121
                                    0.6586
                              Least Squares Means
                                   Age Group
                                    Adults
                                    -2.4851
                                    0.09951
                                      18
                                    -24.97
                                   <.0001
                                     0.05
                                    -2.6942
                                    -2.2760
                                   Age Group
                                   Children
                                    -2.8705
                                    0.09951
                                      18
                                    -28.85
                                   <.0001
                                     0.05
                                    -3.0795
                                    -2.6614
 DF = Degrees of Freedom 
 T Value = Test Statistic
 Pr > t =  p value
 Alpha =  Significance level 
 Lower = Lower 95% Confidence Limit 
 Upper =  Upper 95% Confidence Limit 

VK I Results For Captafol:  Methiocarb and captafol residues were also quantified in the worker media even though no applications of these pesticides were noted at this site.  Triforine applications were also described in the study report but worker media were not analyzed for that pesticide.
The available data for the VK/I site for captafol have been summarized using a box and whisker plot based on total exposure which represents  all patch exposure locations (Figure 10).  Additionally, Wilcoxon Rank Sum test has also been completed (Table 6).   

The conclusion associated with captafol exposure at the VK/I site is that the adult age group had statistically significantly higher exposure than children (one sided p value is 0.95/2 = 0.475).  All statistical tests are conducted at significance level of 0.05; a one sided p value < 0.05 is considered statistically significant, while a one sided p value > 0.05 is not considered statistically significant.  Many samples were reported as having no residues; in fact, no separate analysis was completed for only hands because glove exposures were only quantified in a few samples.  Patch monitoring results were also sporadic which should be considered in the interpretation of the captafol results (see Appendix A.2 and applicable data tabs for further information).




Table 6: Wilcoxon Rank Sum Analysis, Captafol Total Exposure Rates for Blueberry Harvesters, VK I
                                       
 
Figure 11: Box & Whisker Plot, Captafol Total Exposure Rates (mg/kg/hr) for Blueberry Harvesters, VK I
                                       
 
  
VK II- III Results For Captafol: 
The available data for the VK II-III sites for Captafol have not been statistically analyzed because in most cases the monitors contained no residues (see Appendix A.2 and applicable data tabs for further information). 

VK I-III Results For Azinphos Methyl and Methiocarb:  The available data for the VK/I-III site for azinphos methyl and methiocarb have not been statistically analyzed because in most cases the monitors contained no residues (see Appendix A.2 and applicable data tabs for further information).

GC-1/IV-VI Results For Captafol:  The available data for the GC-1/IV-VI sites for captafol have been summarized using box and whisker plots based on hand exposure (Figure 11) and total exposure which represents hand and all patch exposure locations in this case (Figure 12).  

The same workers were monitored on different days in the GC-1/IV-VI fields.  In the analysis of the captafol data, the residual correlation from same workers monitored on multiple days was accounted by using a mixed model.  Both hand and total exposure rates were log transformed and used as dependent variables in the model.  Age group and day were used as fixed effects in the model. The difference of least square means between the adult and child age groups was estimated from the model (Tables 7 and 8).  p values associated with the least square difference of mean for glove and total exposure are greater than 0.05.  The conclusions associated with captafol exposure at the GC-1/IV-VI sites are there is no statistically significant difference between adults and children for hand (glove) exposure rates and also for total exposure rates.

Figure 12: Box & Whisker Plot, Captafol Hand (Glove) Exposure Rates (mg/kg/hr) for Blueberry Harvesters, GC-1/IV-VI
                                       
 
  
Table 7: Differences of Least Squares Means Analysis and Least Square Means, Captafol Hand Exposure Rates for Blueberry Harvesters, GC-1/IV-VI
                                    Effect
                                   Age Group
                                   Age Group
                                   Estimate
                                Standard Error
                                     DF[1]
                                  t Value[2]
                                Pr > |t|[3]
                                   Alpha[4]
                                   Lower[5]
                                   Upper[6]
                      Differences of Least Squares Means
                                   Age Group
                                    Adults
                                   Children
                                    0.07188
                                    0.1568
                                      16
                                     0.46
                                    0.6528
                                     0.05
                                    -0.2605
                                    0.4043
                              Least Squares Means
                                   Age Group
                                    Adults
                                    -3.5407
                                    0.1399
                                      16
                                    -25.31
                                   <.0001
                                     0.05
                                    -3.8373
                                    -3.2441
                                   Age Group
                                   Children
                                    -3.6126
                                    0.1409
                                      16
                                    -25.64
                                   <.0001
                                     0.05
                                    -3.9112
                                    -3.3140
 DF = Degrees of Freedom 
 T Value = Test Statistic
 Pr > t =  p value
 Alpha =  Significance level 
 Lower = Lower 95% Confidence Limit 
 Upper =  Upper 95% Confidence Limit 

	Figure 13: Box & Whisker Plot, Captafol Total Exposure Rates (mg/kg/hr) for Blueberry Harvesters, GC-1/IV-VI
                                       
                                       
 
Table 8: Differences of Least Squares Means Analysis and Least Squares Means, Captafol Total Exposure Rates for Blueberry Harvesters, GC-1/IV-VI
                                    Effect
                                   Age Group
                                   Age Group
                                   Estimate
                                Standard Error
                                     DF[1]
                                  t Value[2]
                                Pr > |t|[3]
                                   Alpha[4]
                                   Lower[5]
                                   Upper[6]
                      Differences of Least Squares Means
                                   Age Group
                                    Adults
                                   Children
                                    0.1323
                                    0.3157
                                      16
                                     0.42
                                    0.6807
                                     0.05
                                    -0.5369
                                    0.8016
                              Least Squares Means
                                   Age Group
                                    Adults
                                    -3.4519
                                    0.2232
                                      16
                                    -15.46
                                   <.0001
                                     0.05
                                    -3.9251
                                    -2.9787
                                   Age Group
                                   Children
                                    -3.5842
                                    0.2232
                                      16
                                    -16.06
                                   <.0001
                                     0.05
                                    -4.0574
                                    -3.1110
 DF = Degrees of Freedom 
 T Value = Test Statistic
 Pr > t =  p value
 Alpha =  Significance level 
 Lower = Lower 95% Confidence Limit 
 Upper =  Upper 95% Confidence Limit 
Investigator Conclusions:  It should also be noted that the investigators concluded that children did not receive higher pesticide exposures than adults, overall.  They noted:

 "Harvesters of blueberries  -  children and adults alike  -  were exposed to residues of previously applied pesticides on berries and foliage in the course of berry picking."

 Overall levels of pesticide exposure of children were not remarkably different from exposures of adults.  There was no indication from this study that exposure of children regularly exceeded that of adults.  Child exposure could be somewhat less, on the average, due to smaller body surface area and possibly less vigorous harvesting activity."


 Summary of Combined Results for Youth in Agriculture Studies

The results of each study have been summarized based on the analyses presented in Appendix A.4.  Generally, no statistical differences were observed between dermal dose estimates (mg/kg/day basis) for working children of all ages to those of adults regardless of whether hands only exposures (i.e., typically the primary exposure contributor) or to the total body are considered.  In some cases, adult exposure rates were higher than those of children.  In just four exposure conditions, the exposures for children exceeded those for adults.  Table 9 below summarizes these results and also provides an overview of the exposure conditions encountered in this analysis.  All 87 exposure conditions are not represented here; only 64 had enough information available to statistically analyze.

Table 9: Results of Statistical Analysis of Applicable Observational Worker Exposure Data Involving Children and Adults
                              Exposure Condition
                                   Study ID
                              Pesticide Monitored
                                   Location
                                     Crop
                         Days After Last Pesticide Use
                               Monitoring Units 
                            Per Exposure Condition
                                  (Children)
                                    Results
                                       1
                                       1
                                   Carbaryl
                                    Oregon
                                  Strawberry
                                      15
                                      18
                                     (13)
 No statistically significant difference for hands or total exposure rates[1]
                                       2
                                       1
                                   Carbaryl
                                    Oregon
                                  Strawberry
                                      16
                                      18
                                     (13)
 No statistically significant difference for hands or total exposure rates[1]
                                       3
                                       1
                                   Carbaryl
                                    Oregon
                                  Strawberry
                                      17
                                      18
                                     (13)
 No statistically significant difference for hands or total exposure rates[1]
                                       4
                                       2
                                    Benomyl
                                    Oregon
                                  Blackberry
                                      17
                                      12
                                      (5)
            No statistically significant difference for hands[1],2 
                                       5
                                       2
                                    Benomyl
                                    Oregon
                                  Blackberry
                                      18
                                      12
                                      (5)
           No statistically significant difference for hands[1][,2]
                                       6
                                       2
                                    Benomyl
                                    Oregon
                                  Blackberry
                                      19
                                      12
                                      (5)
           No statistically significant difference for hands[1][,2]
                                       7
                                       2
                                    Benomyl
                                    Oregon
                                  Blackberry
                                      20
                                      12
                                      (5)
           No statistically significant difference for hands[1][,2]
                                       8
                                       2
                                    Benomyl
                                    Oregon
                                  Blackberry
                                      21
                                      12
                                      (5)
           No statistically significant difference for hands[1][,2]
                                       9
                                       3
                                    Captan
                                  California
                                  Strawberry
                                      10
                                      19
                                     (11)
  Exposure rates for Adults > Children for both hand and total exposure[3]
                                      10
                                       3
                                    Captan
                                    Oregon
                                  Strawberry
                                      26
                                      23
                                     (16)
 No statistically significant difference for hands or total exposure rates[1]
                                      11
                                       3
                                    Captan
                                  California
                                  Strawberry
                                       3
                                      20
                                     (12)
 No statistically significant difference for hands or total exposure rates[1]
                                      12
                                       3
                                    Captan
                                  California
                                  Strawberry
                                      48
                                      10
                                      (3)
 No statistically significant difference for hands or total exposure rates[1]
                                      13
                                       4
                                    Captan
                                   Wisconsin
                                    Apples
                                      27
                                       7
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      14
                                       4
                                Azinphos Methyl
                                   Wisconsin
                                    Apples
                                      27
                                       7
                                      (4)
           No statistically significant difference for hands[1][,4]
                                      15
                                       4
                                    Phosmet
                                   Wisconsin
                                    Apples
                                      38
                                       7
                                      (4)
No statistically significant difference for hands or total exposure rates[1][,4]
                                      16
                                       5
                                   Malathion
                                   Michigan
                                   Blueberry
                                       0
                                      20
                                     (10)
          No statistically significant difference for hand exposure,
                  Adults > Children for total exposure[5]
                                      17
                                       5
                                   Malathion
                                   Michigan
                                   Blueberry
                                       1
                                      20
                                     (10)
          No statistically significant difference for hand exposure,
                  Adults > Children for total exposure[5]
                                      18
                                       5
                                   Malathion
                                   Michigan
                                   Blueberry
                                       2
                                      20
                                     (10)
          No statistically significant difference for hand exposure,
                  Adults > Children for total exposure[5]
                                      19
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      N/A
                                      20
                                     (10)
                  Adults > Children for total exposure[6]
                                      20
                                       5
                                  Methiocarb
                                   Michigan
                                   Blueberry
                                       6
                                      18
                                      (9)
                   Adults > Children for hand exposure[7]
                                      21
                                       5
                                  Methiocarb
                                   Michigan
                                   Blueberry
                                       7
                                      18
                                      (9)
                   Adults > Children for hand exposure[7]
                                      22
                                       5
                                  Methiocarb
                                   Michigan
                                   Blueberry
                                       8
                                      18
                                      (9)
                   Adults > Children for hand exposure[7]
                                      23
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      39
                                      18
                                      (9)
 No statistically significant difference for hands or total exposure rates[1]
                                      24
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      40
                                      18
                                      (9)
 No statistically significant difference for hands or total exposure rates[1]
                                      25
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      41
                                      18
                                      (9)
 No statistically significant difference for hands or total exposure rates[1]
                                      26
                                       5
                                  Methiocarb
                                   Michigan
                                   Blueberry
                                      14
                                      22
                                     (11)
 No statistically significant difference for hands or total exposure rates[1]
                                      27
                                       5
                                  Methiocarb
                                   Michigan
                                   Blueberry
                                      16
                                      22
                                     (11)
 No statistically significant difference for hands or total exposure rates[1]
                                      28
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      47
                                      22
                                     (11)
 No statistically significant difference for hands or total exposure rates[1]
                                      29
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      48
                                      22
                                     (11)
 No statistically significant difference for hands or total exposure rates[1]
                                      30
                                       5
                                   Captafol
                                   Michigan
                                   Blueberry
                                      49
                                      22
                                     (11)
 No statistically significant difference for hands or total exposure rates[1]
                                      31
                                       6
                                   Malathion
                                North Carolina
                                   Blueberry
                                      15
                                      20
                                     (10)
 No statistically significant difference for hands or total exposure rates[1]
                                      32
                                       6
                                    Benomyl
                                North Carolina
                                   Blueberry
                                      21
                                      20
                                     (10)
 No statistically significant difference for hands or total exposure rates[1]
                                      33
                                       7
                                   Malathion
                                North Carolina
                                   Blueberry
                                       4
                                      30
                                     (16)
 No statistically significant difference for hands or total exposure rates[1]
                                      34
                                       7
                                   Malathion
                                North Carolina
                                   Blueberry
                                       5
                                      30
                                     (16)
 No statistically significant difference for hands or total exposure rates[1]
                                      35
                                       8
                                    Treflan
                                  Mississippi
                                     Peas
                                      161
                                       7
                                      (5)
    No statistically significant difference for hand exposure rates[1][,9]
                                      36
                                       8
                                   Carbaryl
                                  Mississippi
                                     Peas
                                      35
                                       7
                                      (5)
 No statistically significant difference for hands or total exposure rates[1]
                                      37
                                       8
                                   Toxaphene
                                  Mississippi
                                     Peas
                                       7
                                       7
                                      (5)
     No statistically significant difference for hand exposure rates[1,9]
                                      38
                                       8
                                   Carbaryl
                                  Mississippi
                                     Peas
                                      35
                                       7
                                      (5)
 No statistically significant difference for hands or total exposure rates[1]
                                      39
                                       8
                                   Toxaphene
                                  Mississippi
                                     Peas
                                       7
                                       7
                                      (5)
     No statistically significant difference for hand exposure rates[1,9]
                                      40
                                      11
                                   Carbaryl
                                  Mississippi
                                   Cucumber
                                      16
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      41
                                      11
                                   Carbaryl
                                  Mississippi
                                   Cucumber
                                       1
                                      10
                                      (4)
                        Children > Adults for hands,
          No statistically significant difference for total exposure
                                      42
                                      11
                                   Carbaryl
                                  Mississippi
                                   Cucumber
                                       1
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      43
                                      11
                                   Carbaryl
                                  Mississippi
                                   Cucumber
                                       1
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      44
                                      11
                                   Toxaphene
                                  Mississippi
                                     Peas
                                       5
                                      10
                                      (5)
 No statistically significant difference for hands or total exposure rates[1]
                                      45
                                      11
                                   Toxaphene
                                  Mississippi
                                     Peas
                                       4
                                      10
                                      (5)
 No statistically significant difference for hands or total exposure rates[1]
                                      46
                                      11
                                   Toxaphene
                                  Mississippi
                                     Peas
                                       1
                                      10
                                      (3)
 No statistically significant difference for hands or total exposure rates[1]
                                      47
                                      11
                               Methyl Parathion
                                  Mississippi
                                    Peanut
                                      131
                                      10
                                      (5)
           Children > Adults for both hand and total exposure[6]
                                      48
                                      12
                                  Endosulfan
                                  Mississippi
                                     Peas
                                      11
                                      10
                                      (5)
                        Adults > Children for hands
          No statistically significant difference for total exposure
                                      49
                                      12
                                  Endosulfan
                                  Mississippi
                                     Peas
                                       3
                                      10
                                      (5)
                        Adults > Children for hands
          No statistically significant difference for total exposure
                                      50
                                      12
                                  Endosulfan
                                  Mississippi
                                     Peas
                                       4
                                      10
                                      (5)
                        Adults > Children for hands
          No statistically significant difference for total exposure
                                      51
                                      12
                                  Endosulfan
                                  Mississippi
                                     Corn
                                       2
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      52
                                      12
                                  Endosulfan
                                  Mississippi
                                     Corn
                                       3
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      53
                                      12
                                  Endosulfan
                                  Mississippi
                                     Corn
                                       2
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      54
                                      12
                                  Endosulfan
                                  Mississippi
                                     Corn
                                       2
                                      10
                                      (4)
 No statistically significant difference for hands or total exposure rates[1]
                                      55
                                      12
                                  Endosulfan
                                  Mississippi
                                     Peas
                                       3
                                      12
                                      (6)
 No statistically significant difference for hands or total exposure rates[1]
                                      56
                                      13
                                  Vinclozolin
                                    Oregon
                                  Strawberry
                                      38
                                      18
                                     (13)
             Adults > Children for both hand and total exposure
                                      57
                                      13
                                  Vinclozolin
                                    Oregon
                                  Strawberry
                                      39
                                      18
                                     (13)
             Adults > Children for both hand and total exposure
                                      58
                                      13
                                  Vinclozolin
                                    Oregon
                                  Strawberry
                                      40
                                      18
                                     (13)
             Adults > Children for both hand and total exposure
                                      59
                                      15
                                   Acephate
                                North Carolina
                                    Tobacco
                                       1
                                      17
                                      (8)
             Adults > Children for both hand and total exposure
                                      60
                                      15
                                   Acephate
                                North Carolina
                                    Tobacco
                                       2
                                      17
                                      (8)
             Adults > Children for both hand and total exposure
                                      61
                                      15
                                 Methamidophos
                                North Carolina
                                    Tobacco
                                       1
                                      17
                                      (8)
             Adults > Children for both hand and total exposure
                                      62
                                      16
                                Chlorothalonil
                                     Texas
                                    Tomato
                                       9
                                      39
                                     (13)
           No statistically significant difference for hands[1][,10]
                                      63
                                      17
                                  Endosulfan
                                South Carolina
                                    Tomato
                                       1
                                      10
                                      (5)
                        Children > Adults for hands,
                    Adults > Children for total exposure
                                      64
                                      17
                                  Endosulfan
                                South Carolina
                                    Tomato
                                       1
                                      10
                                      (5)
                        Children > Adults for hands,
                    Adults > Children for total exposure
                   Number of Monitoring Units Analyzed: 948
             Number of Monitoring Units Analyzed for Children: 501
 No statistically significant difference between exposures for children and adults for applicable exposure scenarios analyzed in each study, this indicates the current risk assessment methods are protective of children working in agriculture.  
 Only hand exposure was analyzed for Study 2; only about 55% (12 out of 22) of participants had non-missing exposure values for each exposure conditions for exposures other than hands.  
 Additional data for total exposure was not available; therefore only hand exposure was analyzed.  
 Captan was the only pesticide measured on hands in Study 4. Azinphos methy and phosmet were not found on patches worn on study participant's bodies.  Therefore, only hand residues were identified and analyzed for azinphos methyl and phosmet, not total exposures.  
 No statistically significant difference for hands between adults and children, indicating current risk assessment methods are protective of children working in agriculture receiving dermal exposures via hands.  
 No separate analysis was completed for hands because glove exposures were only quantified in a few samples.  
 Total exposure was not statistically analyzed because monitors were found to contain non-detectable residues for back, chest, shoulders, inside back, and inside forearms.  
 At the time of the study, peanuts were primarily hand harvested.  However, peanuts are now primarily mechanically harvested; greatly reducing/eliminating the concern for exposure to pesticides during hand harvesting. 
 Total exposure, represented by hand, forearm, chest, back, and shoulder exposures was not analyzed since no sample collected contained any detectable residues of treflan. 
 Total exposure was not statistically analyzed because there were too many missing data from forearm and leg measurements.  

           Discussion

In order to properly consider the results of this analysis there are several issues which must be considered.  These include:

 The monitoring data were generated using methods that at the time were considered scientifically state of the art.  Even though current techniques differ somewhat from those used in the research considered in this analysis the results are still appropriate for purposes of comparing the post-application exposures of children and adults.  The hand sampling methods are still commonly employed and the hands represent the majority of the exposures.  Also, a number of exposure values still currently used for risk assessment purposes are based on the patch sampling method and extrapolation across monitored body locations (i.e., most PHED data).
   
 The agricultural activities monitored in the studies (e.g., hand harvest of berries) have remained essentially unchanged in current agriculture from the period when the studies were completed.  This supports the use of these data for evaluation of children working in current agricultural production systems. There is one notable exception - peanut harvesting was conducted by hand when the studies were conducted and in current agriculture it has become highly mechanized.  This exception also represents soil contact versus foliar contact; foliar contact is the more applicable to post-application exposure.  

 The mechanistic monitoring analysis results and those from the monitoring data support each other in that similar conclusions are reached independently that there is generally no meaningful difference in the exposure of adults and children from hand-based agricultural activities.  Therefore, it is reasonable for the Agency to rely on it current exposure assessment methods, which estimate exposure to adults, to account for the exposure to children working in agriculture in field labor tasks.

 Approaches used for hazard identification and dose response analysis by the Agency are also protective of children working in agriculture.  The most sensitive endpoints are always selected for risk assessments and any special sensitivities would also be routinely identified.  Appropriate points of departure and uncertainty factors based on any special sensitivities would be selected for exposure and risk assessment.  Therefore, not only exposure assessment findings but also risk assessment findings completed for reentry workers (e.g., field labor) would be protective of children working in those types of jobs.  Given that the required data are used to evaluate sensitivity in developing children, it follows that not only the exposure assessment process but risk assessment findings account for children working in agriculture.

 There are pesticides that were included in the monitoring studies that are no longer in the market.  It is also likely that perhaps certain use conditions for the remaining pesticides may have also been altered.  These issues do not alter the utility of the data for the purposes of this analysis, since the focus was to look at the potential differences between the pesticide exposure rates between the monitored children and adults as a result of the legal uses of the pesticides involved at that time.

 In order to account for different age group configurations, the Agency re-evaluated a subset of three of the studies using ages <13 as children and >13 as adults to determine whether the overall conclusion of this analysis would change.  Using these age groups, the sensitivity analysis showed that there is no significant difference in exposures between children and adults.  Therefore, using a different age cut off for adults and children will not alter the overall conclusion that there is no difference in exposure between children and adult agricultural workers.  

 It is clear from the study reports that the conduct of the monitoring studies used as the basis for this analysis involved a comprehensive quality assurance and quality control program.  Because of this and the focus of the analysis on relative differences between adults and children, no corrections were made to the exposures based on factors like field recovery.  The corrections for the same exposure condition would be relative for all participants, and so it would not impact the results.

 There is a significant amount of monitoring data that has been included in this analysis that rivals in scope all of the current exposure databases used for regulatory purposes.  This breadth of information adds to the overall rigor of the analysis.

 All the exposure monitoring data included in this analysis have been evaluated and found to be compliant with Federal requirements for involving humans in research. 


           Conclusions

This document confirms the validity and applicability of current EPA pesticide exposure and risk assessment methodology to assess exposures that child farmworkers may have.   Based on multiple sources of demographic information, there are a significant number of children aged 12 through 18 that are currently actively engaged in a variety of farm labor activities, including pre-harvest hand labor and harvesting tasks, and there may be children even younger working on family farms. An analysis was completed using mechanistic information about adults and children, and concluded that using adult body size and weight in risk assessments is protective of child farmworkers.  Additionally, an evaluation of extensive exposure monitoring data gathered from child farmworkers in agriculture showed no significant difference in exposures between children and adults conducting post-application activities.  Based on current Agency policy for risk assessment, any potential toxicity sensitivities of children would be accounted for when selecting the toxicity points of departure for risk assessment.  Therefore, using this weight of the evidence approach, the Agency concludes that children and adults working in agriculture do not have significantly different levels of pesticide exposures, and current risk and exposure assessment methods for pesticides are protective of child farmworkers.    

This analysis should be considered highly rigorous given the breadth of information upon which it is based.  The mechanistic analysis relies on information from the EPA Exposure Factors Handbook, a nationally representative compendium of mechanistic information used in exposure assessments.  Similarly, the exposure monitoring was conducted by multiple researchers and institutions, with the vast majority of the results supporting the Agency's conclusion, only a small percentage demonstrated that children's exposures exceeded those of adults (2.4%).  The monitoring data was statistically analyzed using comprehensive statistical models appropriate for the data collected in each study since the designs of the studies and the nature of the results of each varied.  Also, the breadth of the exposure conditions which were encountered (87) in the exposure monitoring studies and the total number of exposure days monitored which were considered (1472) underlies the rigor of the findings.  


