Using GIS and historical records to reconstruct residential exposure to large-scale pesticide application

Evaluating the accuracy of satellite-based methods to estimate residential proximity to agricultural crops

BACKGROUND

Epidemiologic investigations increasingly employ remote sensing data to estimate residential proximity to agriculture as a means of approximating individual-level pesticide exposure. Few studies have examined the accuracy of these methods and the implications for exposure misclassification.

OBJECTIVES

Compare metrics of residential proximity to agricultural land between a groundtruth approach and commonly-used satellite-based estimates.

METHODS

We inspected 349 fields and identified crops in current production within a 0.5 km radius of 40 residences in Idaho. We calculated the distance from each home to the nearest agricultural field and the total acreage of agricultural fields within a 0.5 km buffer. We compared these groundtruth estimates to satellite-derived estimates from three widely used datasets: CropScape, the National Land Cover Database (NLCD), and Landsat imagery (using Normalized Difference Vegetation Index thresholds).

RESULTS

We found poor to moderate agreement between the classification of individuals living within 0.5 km of an agricultural field between the groundtruth method and the comparison datasets (53.1-77.6%). All satellite-derived estimates overestimated the acreage of agricultural land within 0.5 km of each home (average = 82.8-148.9%). Using two satellite-derived datasets in conjunction resulted in substantial improvements; specifically, combining CropScape or NLCD with Landsat imagery had the highest percent agreement with the groundtruth data (92.8-93.8% agreement).

SIGNIFICANCE

Residential proximity to agriculture is frequently used as a proxy for pesticide exposure in epidemiologic investigations, and remote sensing-derived datasets are often the only practical means of identifying cultivated land. We found that estimates of agricultural proximity obtained from commonly-used satellite-based datasets are likely to result in exposure misclassification. We propose a novel approach that capitalizes on the complementary strengths of different sources of satellite imagery, and suggest the combined use of one dataset with high temporal resolution (e.g., Landsat imagery) in conjunction with a second dataset that delineates agricultural land use (e.g., CropScape or NLCD).

Role of land use and land cover in residential exposures to agricultural pesticide models

Exposure of the general population to pesticides, especially in agricultural areas, is a major public health concern. This review analyses the role of Land Use and Land Cover (LULC) in Residential Exposure to Agricultural Pesticides (REAP) and how it is measured and modelled. Some epidemiological studies have shown that basic LULC variables, such as distance to a crop and field size, are relevant for explaining REAP. However, the potential of LULC mitigation elements, such as vegetation barriers, grassy strips and buffer zones, to mitigate REAP has been poorly studied. The availability of recent low-cost and high-quality geospatial data enables REAP models to include alternative and more precise LULC variables. This review also highlights the need for (i) generic environmental sampling protocols, (ii) exposure and spraying datasets and (iii) assessment of the mitigation capacity of LULC to improve REAP modelling significantly.

Pesticide toxicity assessment and geographic information system (GIS) application in small-scale rice farming operations, Thailand

The objectives of the study were to evaluate the impact of pesticide exposure on farmer health during non-active rice farming and active rice farming periods and present the change in the individual cholinesterase activities (%reduction) on the geographic information system (GIS) mapping in Nakhon Ratchasima Province, Thailand. Acetyl- and butyryl-cholinesterase (AChE and BuChE) activities were monitored during both study periods using Test-mate ChE (Model 400). The location of paddy fields was specified using Garmin geographic positioning system MAP 62s. Fifty-eight farmers who participated in this study had an average age of 49.2 ± 6.9 years. Higher prevalence of all health symptoms was observed among farmer participants during the active rice farming period comparing to the non-active rice farming period (p < 0.01). Furthermore, farmers had significantly lower activities of AChE and BuChE during the active rice farming period comparing to the non-active rice farming period (p < 0.01). Our findings indicate that the GIS mapping indicate that the cases with a significant enzyme inhibition have dispersed across the agricultural and the nearby residential areas. This, investigation can be used to promote safer use of pesticides among farmers and mitigate pesticide exposure among residents living in close proximity to a rice field.

Dietary Risk with Other Risk Factors of Breast Cancer

Breast cancer is one of the most common and feared cancers. The incidence of breast cancer is persistently on the rise due to urbanization and lifestyle changes. Breast cancer cannot be prevented fully but can be effectively treated and controlled if risk factors are determined accurately. Detection of breast cancer at an early stage along with the advancement in treatment options has provided a greater chance of survival.

Passive environmental residential exposure to agricultural pesticides and hematological malignancies in the general population: a systematic review

Incidence rates of hematological malignancies have been constantly increasing over the past 40 years. In parallel, an expanding use of agricultural pesticides has been observed. Only a limited number of studies investigated the link between hematological malignancies risk and passive environmental residential exposure to agricultural pesticides in the general population. The purpose of our review was to summarize the current state of knowledge on that question. A systematic literature search was conducted using PubMed and Scopus databases. We built a scoring scale to appraise relevance of each selected articles. We included 23 publications: 13 ecological studies, 9 case-control studies and a cohort study. Positive associations were reported between hematological malignancies and individual pesticides, pesticide groups, all pesticides without distinction, or some crop types. Relevance score was highly various across studies regardless of their design. Children studies were the majority and had overall higher relevance scores. The effect of passive environmental residential exposure to agricultural pesticides on hematological malignancies risk is suggested by the literature. The main limitation of the literature available is the high heterogeneity across studies, especially in terms of exposure assessment approach. Further studies with high methodological relevance should be conducted.

Association between Pesticides in House Dust and Residential Proximity to Farmland in a Rural Region of Taiwan

Pesticide drift was reported in many international studies, but rarely studied in Taiwan. We conducted a study in a rural region of Taiwan to examine the associations between pesticides in house dust and nearby agricultural areas using geographic information system (GIS). A questionnaire regarding home characteristics and pesticide use, and indoor and outdoor dust samples were collected from 47 rural homes. Dust samples were analyzed for six pesticides, and agricultural land data for GIS analysis were retrieved from a national website. All but prallethrin were frequently detected from indoor dust samples (>50%), and the maximum concentrations were all below 1000 ng/g. Detection frequencies and concentrations of pesticides in outdoor dust samples were even lower than that in indoor dust samples. Only "work involving pesticides" in the questionnaire was significantly associated with four pesticides in house dust (p < 0.05). Carbofuran and tetramethrin in house dust were significantly correlated with rice cultivation area at certain buffer distances (ρ > 0.33, p < 0.05), and chlorpyrifos was found to be associated with abandoned cultivation area, suggesting the occurrence of pesticide drift. Despite the low levels of pesticides in house dust, residents in the rural region should be cautious of pesticide drift from nearby active or abandoned farmlands.

Using geographic information systems to estimate potential pesticide exposure at the population level in Canada

Residents in close proximity to agricultural land are at risk of higher pesticide exposures. The purpose of this study was to generate national population-level exposure estimates for Canada for three commonly applied pesticides that are suspected carcinogens (2,4-dichlorophenoxyacetic acid (2,4-D), glyphosate and chlorothalonil). Using geographic information systems, pesticide exposure was estimated for every - census subdivision (CSD) in Canada (n = 5054) by combining raster-based surfaces for the distribution of crops with average crop-specific pesticide application rates data. Analyses examined all identified crops in combination with different pesticide application rates to obtain a cumulative potential total exposure. Specifically, the number of acres of particular crops were calculated for each CSD and then multiplied by the average pesticide application rates data, summed across crops, and combined with population data by CSD to provide a potential pesticide exposure estimate for each CSD. Results demonstrate that the population exposure varies greatly depending on agricultural production by CSD region. For example, in Ontario, the 2,4-D application rate was an average of 361 kg/km², while in Saskatchewan, which primarily grows field/cereal crops, 2,4-D application rates were much higher (3810 kg/km²). The highest potential exposures to all three pesticides were in the prairie provinces (Alberta, Saskatchewan, Manitoba) along with Prince Edward Island, Southern Quebec and British Columbia. This work can be used in conjunction with other exposure assessment approaches to better understand overall pesticide exposure among Canada's general population.

"Good Epidemiology Practice" Guidelines for Pesticide Exposure Assessment

Both toxicology and epidemiology are used to inform hazard and risk assessment in regulatory settings, particularly for pesticides. While toxicology studies involve controlled, quantifiable exposures that are often administered according to standardized protocols, estimating exposure in observational epidemiology studies is challenging, and there is no established guidance for doing so. However, there are several frameworks for evaluating the quality of published epidemiology studies. We previously developed a preliminary list of methodology and reporting standards for epidemiology studies, called Good Epidemiology Practice (GEP) guidelines, based on a critical review of standardized toxicology protocols and available frameworks for evaluating epidemiology study quality. We determined that exposure characterization is one of the most critical areas for which standards are needed. Here, we propose GEP guidelines for pesticide exposure assessment based on the source of exposure data (i.e., biomonitoring and environmental samples, questionnaire/interview/expert record review, and dietary exposures based on measurements of residues in food and food consumption). It is expected that these GEP guidelines will facilitate the conduct of higher-quality epidemiology studies that can be used as a basis for more scientifically sound regulatory risk assessment and policy making.

Assessment of residential exposures to agricultural pesticides: A scoping review

The assessment of residential exposure to agricultural pesticides is a major issue for public health, regulatory and management purposes. In recent years, research into this field has developed considerably. The purpose of this scoping review is to provide an overview of scientific literature characterizing residential exposure to agricultural pesticides and to identify potential gaps in this research area. This work was conducted according to the JBI and PRISMA guidelines. Three databases were consulted. At least two experts selected the eligible studies. Our scoping review enabled us to identify 151 articles published between 1988 and 2019 dealing with the assessment of residential exposure to agricultural pesticides. Of these, 98 (64.9%) were epidemiological studies investigating possible links between pesticide exposure and the onset of adverse health effects, principally cancers and reproductive outcomes. They predominantly used Geographic Information Systems and sometimes surveys or interviews to calculate surrogate exposure metrics, the most common being the amounts of pesticides applied or the surface area of crops around the dwelling. Twenty-six (17.2%) were observational measurement studies conducted to quantify levels of pesticide exposure and identify their possible determinants. These studies assessed exposure by measuring pesticides in biological and environmental matrices, mostly in urines and house dust. Finally, we found only eight publications (5.3%) that quantified the risk to human health due to residential exposure for management purposes, in which exposure was mainly determined using probabilistic models. Pesticide exposure appears to be largely correlated with the spatial organization of agriculture activities in a territory. The determinants and routes of exposure remain to be explored to improve the conduct of epidemiological and risk assessment studies and to help prevent future exposures. Improvement could be expected from small-scale studies combining different methods of exposure assessment.

Occupational exposure to pesticides: development of a job-exposure matrix for use in population-based studies (PESTIPOP)

Occupational exposure to pesticides concerns a wide population of workers not only in agriculture. The reliability of self-reported information on pesticide use is questionable because of the diversity of use. The PESTIPOP job-exposure matrix has been designed to assess pesticide occupational exposure in the general population. The matrix is composed of two axes: the first axis corresponding to jobs (combinations of occupations and industries) and the second one to pesticide exposure. The estimated exposure metric is the probability of exposure coupled with a reliability assessment (low, medium or high). These metrics were defined by combining different sources: (1) an a priori expert assessment (Agricultural industry experts); (2) data from a multicenter case-control study on brain tumors in the general population (occupational history, specific questionnaires); and (3) an a posteriori expert assessment based on the data of a case-control study. So far, 2559 jobs have been identified and 209 (8%) were found to be exposed to pesticides. Jobs with agricultural exposure had a higher exposure probability than jobs with non-agricultural exposure (wood preservation, park maintenance, pest control). Indirect exposure was more frequent than direct exposure. The PESTIPOP matrix will be transcoded into international classifications for use in epidemiological studies.

Assessment of residential environmental exposure to pesticides from agricultural fields in the Netherlands

We developed a spatio-temporal model for the Netherlands to estimate environmental exposure to individual agricultural pesticides at the residential address for application in a national case-control study on Parkinson's disease (PD). Data on agricultural land use and pesticide use were combined to estimate environmental exposure to pesticides for the period 1961 onwards. Distance categories of 0-50 m, >50-100 m, >100-500 m and >500-1000 m around residences were considered. For illustration purposes, exposure was estimated for the control population (n=607) in the PD case-control study. In a small validation effort, model estimates were compared with pesticide measurements in air and precipitation collected at 17 stations in 2000-2001. Estimated exposure prevalence was higher for pesticides used on commonly cultivated (rotating) crops than for pesticides used on fruit and bulbs only. Prevalence increased with increasing distance considered. Moderate-to-high correlations were observed between model estimates (>100-500 m and >500-1000 m) and environmental pesticide concentrations measured in 2000-2001. Environmental exposure to individual pesticides can be estimated using relevant spatial and temporal data sets on agricultural land use and pesticide use. Our approach seems to result in accurate estimates of average environmental exposure, although it remains to be investigated to what extent this reflect personal exposure to agricultural pesticides.

Environmental exposure to pesticides and the risk of Parkinson's disease in the Netherlands

BACKGROUND

Exposure to pesticides has been linked to Parkinson's disease (PD), although associations between specific pesticides and PD have not been well studied. Residents of rural areas can be exposed through environmental drift and volatilization of agricultural pesticides.

OBJECTIVES

Our aim was to investigate the association between lifetime environmental exposure to individual pesticides and the risk of PD, in a national case-control study.

METHODS

Environmental exposure to pesticides was estimated using a spatio-temporal model, based on agricultural crops around the residential address. Distance up to 100m from the residence was considered most relevant, considering pesticide drift potential of application methods used in the Netherlands. Exposure estimates were generated for 157 pesticides, used during the study period, of which four (i.e. paraquat, maneb, lindane, benomyl) were considered a priori relevant for PD.

RESULTS

A total of 352 PD cases and 607 hospital-based controls were included. No significant associations with PD were found for the a priori pesticides. In a hypothesis generating analysis, including 153 pesticides, increased risk of PD was found for 21 pesticides, mainly used on cereals and potatoes. Results were suggestive for an association between bulb cultivation and PD.

CONCLUSIONS

For paraquat, risk estimates for the highest cumulative exposure tertile were in line with previously reported elevated risks. Increased risk of PD was observed for exposure to (a cluster of) pesticides used on rotating crops. High correlations limited our ability to identify individual pesticides responsible for this association. This study provides some evidence for an association between environmental exposure to specific pesticides and the risk of PD, and generates new leads for further epidemiological and mechanistic research.

Pesticide exposure assessed through agricultural crop proximity and risk of amyotrophic lateral sclerosis

BACKGROUND

Epidemiologic studies have raised the possibility that some pesticide compounds induce the neurodegenerative disease amyotrophic lateral sclerosis (ALS), though the available evidence is not entirely consistent.

METHODS

We conducted a population-based case-control study in two Italian populations to assess the extent to which residence in the vicinity of agricultural crops associated with the application of neurotoxic pesticides is a risk factor for ALS, using crop acreage in proximity to the residence as an index of exposure.

RESULTS

Based on 703 cases and 2737 controls, we computed an ALS odds ratio of 0.92 (95% confidence interval 0.78-1.09) for those in proximity to agricultural land. Results were not substantially different when using alternative exposure categories or when analyzing specific crop types, with the exception of a higher risk related to exposure to citrus orchards and olive groves in Southern Italy, though based on few exposed subjects (N = 89 and 8, respectively). There was little evidence of any dose-response relation between crop proximity and ALS risk, and using long-term residence instead of current residence did not substantially change our estimates.

CONCLUSIONS

Though our index of exposure is indirect and subject to considerable misclassification, our results offer little support for the hypothesis that neurotoxic pesticide exposure increases ALS risk.

Use of job-exposure matrices to estimate occupational exposure to pesticides: A review

The health effects of pesticides have been extensively studied in epidemiology, mainly in agricultural populations. However, pesticide exposure assessment remains a key methodological issue for epidemiological studies. Besides self-reported information, expert assessment or metrology, job-exposure matrices still appear to be an interesting tool. We reviewed all existing matrices assessing occupational exposure to pesticides in epidemiological studies and described the exposure parameters they included. We identified two types of matrices, (i) generic ones that are generally used in case-control studies and document broad categories of pesticides in a large range of jobs, and (ii) specific matrices, developed for use in agricultural cohorts, that generally provide exposure metrics at the active ingredient level. The various applications of these matrices in epidemiological studies have proven that they are valuable tools to assess pesticide exposure. Specific matrices are particularly promising for use in agricultural cohorts. However, results obtained with matrices have rarely been compared with those obtained with other tools. In addition, the external validity of the given estimates has not been adequately discussed. Yet, matrices would help in reducing misclassification and in quantifying cumulated exposures, to improve knowledge about the chronic health effects of pesticides.

Accuracy of two geocoding methods for geographic information system-based exposure assessment in epidemiological studies

BACKGROUND

Environmental exposure assessment based on Geographic Information Systems (GIS) and study participants' residential proximity to environmental exposure sources relies on the positional accuracy of subjects' residences to avoid misclassification bias. Our study compared the positional accuracy of two automatic geocoding methods to a manual reference method.

METHODS

We geocoded 4,247 address records representing the residential history (1990-2008) of 1,685 women from the French national E3N cohort living in the Rhône-Alpes region. We compared two automatic geocoding methods, a free-online geocoding service (method A) and an in-house geocoder (method B), to a reference layer created by manually relocating addresses from method A (method R). For each automatic geocoding method, positional accuracy levels were compared according to the urban/rural status of addresses and time-periods (1990-2000, 2001-2008), using Chi Square tests. Kappa statistics were performed to assess agreement of positional accuracy of both methods A and B with the reference method, overall, by time-periods and by urban/rural status of addresses.

RESULTS

Respectively 81.4% and 84.4% of addresses were geocoded to the exact address (65.1% and 61.4%) or to the street segment (16.3% and 23.0%) with methods A and B. In the reference layer, geocoding accuracy was higher in urban areas compared to rural areas (74.4% vs. 10.5% addresses geocoded to the address or interpolated address level, p < 0.0001); no difference was observed according to the period of residence. Compared to the reference method, median positional errors were 0.0 m (IQR = 0.0-37.2 m) and 26.5 m (8.0-134.8 m), with positional errors <100 m for 82.5% and 71.3% of addresses, for method A and method B respectively. Positional agreement of method A and method B with method R was 'substantial' for both methods, with kappa coefficients of 0.60 and 0.61 for methods A and B, respectively.

CONCLUSION

Our study demonstrates the feasibility of geocoding residential addresses in epidemiological studies not initially recorded for environmental exposure assessment, for both recent addresses and residence locations more than 20 years ago. Accuracy of the two automatic geocoding methods was comparable. The in-house method (B) allowed a better control of the geocoding process and was less time consuming.

Validity of geographically modeled environmental exposure estimates

Geographic modeling is increasingly being used to estimate long-term environmental exposures in epidemiologic studies of chronic disease outcomes. However, without validation against measured environmental concentrations, personal exposure levels, or biologic doses, these models cannot be assumed a priori to be accurate. This article discusses three examples of epidemiologic associations involving exposures estimated using geographic modeling, and identifies important issues that affect geographically modeled exposure assessment in these areas. In air pollution epidemiology, geographic models of fine particulate matter levels have frequently been validated against measured environmental levels, but comparisons between ambient and personal exposure levels have shown only moderate correlations. Estimating exposure to magnetic fields by using geographically modeled distances is problematic because the error is larger at short distances, where field levels can vary substantially. Geographic models of environmental exposure to pesticides, including paraquat, have seldom been validated against environmental or personal levels, and validation studies have yielded inconsistent and typically modest results. In general, the exposure misclassification resulting from geographic models of environmental exposures can be differential and can result in bias away from the null even if non-differential. Therefore, geographic exposure models must be rigorously constructed and validated if they are to be relied upon to produce credible scientific results to inform epidemiologic research. To our knowledge, such models have not yet successfully predicted an association between an environmental exposure and a chronic disease outcome that has eventually been established as causal, and may not be capable of doing so in the absence of thorough validation.

An evaluation of metrics for assessing maternal exposure to agricultural pesticides

We evaluate the use of three different exposure metrics to estimate maternal agricultural pesticide exposure during pregnancy. Using a geographic information system-based method of pesticide exposure estimation, we combine data on crop density and specific pesticide application amounts/dates to create the three exposure metrics. For illustration purposes, we create each metric for a North Carolina cohort of pregnant women, 2003-2005, and analyze the risk of congenital anomaly development with a focus on metric comparisons. Based on the results, and the need to balance data collection efforts/computational efficiency with accuracy, the metric which estimates total chemical exposure using application dates based on crop-specific earliest planting and latest harvesting information is preferred. Benefits and drawbacks of each metric are discussed and recommendations for extending the analysis to other states are provided.

Expert assessment of historical crop specific pesticide use in the Netherlands

OBJECTIVES

Retrospective assessment of environmental pesticide exposure is challenging. Exposure measurements or information on crop-specific pesticide use are often lacking historically. We applied expert assessment to reconstruct historical pesticide use patterns in the Netherlands, and evaluated reliability and accuracy of this procedure.

METHODS

For six main crops in the Netherlands, two experts per crop individually rated the probability (percentage of farmers applying) and frequency of use of authorised active ingredients between 1961 and 2005 per 5-year period. Inter-rater agreement was investigated by the percentage overall agreement and weighted Cohen's κ's (κ(w)). Experts' ratings were compared with self-reported pesticide use from recent farmer surveys to determine accuracy of the ratings.

RESULTS

Inter-rater agreement on the probability of use varied between crops (κ(w) 0.25 to 0.69), as well as agreement on the frequency of use (κ(w) 0.32 to 0.64). Inter-rater agreement was marginally higher for herbicides and fungicides than insecticides. Comparisons with survey data indicated fair to good accuracy of the experts' ratings for the probability (κ(w) 0.48 to 0.65) and frequency of use (κ(w) 0.38 to 0.68). For all crops except fruit, the specificity of the experts' ratings was higher than the sensitivity.

CONCLUSIONS

Overall inter-rater agreement between experts was fair to good and experts' ratings were reasonably accurate. Results of this study indicate that expert assessment can be used to derive information on historical pesticide use, which is essential for epidemiological studies evaluating the effect of (past) environmental exposure to pesticides on health.

The effect of administrative boundaries and geocoding error on cancer rates in California

Geocoding is often used to produce maps of disease rates from the diagnosis addresses of incident cases to assist with disease surveillance, prevention, and control. In this process, diagnosis addresses are converted into latitude/longitude pairs which are then aggregated to produce rates at varying geographic scales such as Census tracts, neighborhoods, cities, counties, and states. The specific techniques used within geocoding systems have an impact on where the output geocode is located and can therefore have an effect on the derivation of disease rates at different geographic aggregations. This paper investigates how county-level cancer rates are affected by the choice of interpolation method when case data are geocoded to the ZIP code level. Four commonly used areal unit interpolation techniques are applied and the output of each is used to compute crude county-level five-year incidence rates of all cancers in California. We found that the rates observed for 44 out of the 58 counties in California vary based on which interpolation method is used, with rates in some counties increasing by nearly 400% between interpolation methods.

Generating land cover boundaries from remotely sensed data using object-based image analysis: overview and epidemiological application

Satellite imagery and aerial photography represent a vast resource to significantly enhance environmental mapping and modeling applications for use in understanding spatio-temporal relationships between environment and health. Deriving boundaries of land cover objects, such as trees, buildings, and crop fields, from image data has traditionally been performed manually using a very time consuming process of hand digitizing. Boundary detection algorithms are increasingly being applied using object-based image analysis (OBIA) technology to automate the process. The purpose of this paper is to present an overview and demonstrate the application of OBIA for delineating land cover features at multiple scales using a high resolution aerial photograph (1 m) and a medium resolution Landsat image (30 m) time series in the context of a pesticide spray drift exposure application.

Using Landsat satellite data to support pesticide exposure assessment in California

BACKGROUND

The recent U.S. Geological Survey policy offering Landsat satellite data at no cost provides researchers new opportunities to explore relationships between environment and health. The purpose of this study was to examine the potential for using Landsat satellite data to support pesticide exposure assessment in California.

METHODS AND RESULTS

We collected a dense time series of 24 Landsat 5 and 7 images spanning the year 2000 for an agricultural region in Fresno County. We intersected the Landsat time series with the California Department of Water Resources (CDWR) land use map and selected field samples to define the phenological characteristics of 17 major crop types or crop groups. We found the frequent overpass of Landsat enabled detection of crop field conditions (e.g., bare soil, vegetated) over most of the year. However, images were limited during the winter months due to cloud cover. Many samples designated as single-cropped in the CDWR map had phenological patterns that represented multi-cropped or non-cropped fields, indicating they may have been misclassified.

CONCLUSIONS

We found the combination of Landsat 5 and 7 image data would clearly benefit pesticide exposure assessment in this region by 1) providing information on crop field conditions at or near the time when pesticides are applied, and 2) providing information for validating the CDWR map. The Landsat image time-series was useful for identifying idle, single-, and multi-cropped fields. Landsat data will be limited during the winter months due to cloud cover, and for years prior to the Landsat 7 launch (1999) when only one satellite was operational at any given time. We suggest additional research to determine the feasibility of integrating CDWR land use maps and Landsat data to derive crop maps in locations and time periods where maps are not available, which will allow for substantial improvements to chemical exposure estimation.

Use of land surface remotely sensed satellite and airborne data for environmental exposure assessment in cancer research

In recent years, geographic information systems (GIS) have increasingly been used for reconstructing individual-level exposures to environmental contaminants in epidemiological research. Remotely sensed data can be useful in creating space-time models of environmental measures. The primary advantage of using remotely sensed data is that it allows for study at the local scale (e.g., residential level) without requiring expensive, time-consuming monitoring campaigns. The purpose of our study was to identify how land surface remotely sensed data are currently being used to study the relationship between cancer and environmental contaminants, focusing primarily on agricultural chemical exposure assessment applications. We present the results of a comprehensive literature review of epidemiological research where remotely sensed imagery or land cover maps derived from remotely sensed imagery were applied. We also discuss the strengths and limitations of the most commonly used imagery data (aerial photographs and Landsat satellite imagery) and land cover maps.

Development of a GIS-based indicator for environmental pesticide exposure and its application to a Belgian case-control study on bladder cancer

We developed two indicators to assess the exposure of residents to agricultural pesticide use and applied it in a case-control study on bladder cancer in the province of Limburg, Belgium. The first indicator used a distance-weighted measure of crop area for specified crops (fruit trees, fruit bushes and vegetables). The second indicator used a distance-weighted measure of pesticide use. We used information at three scale levels: (a) information at individual's level, such as distance to crop fields; (b) information at the level of the municipality, such as time-series of crop area; and (c) regional information, such as pesticide use. Pesticide use data were available per group of pesticides (fungicides, herbicides, insecticides, growth regulators and group of other pesticides). Indicators were calculated for each individual in the case-control study. The indicators were calculated per year for a period of 20 years, taking into account address history. Variation of pesticide use and toxicity with time was addressed by a relative risk factor. A very strong correlation was found between the area of fruit trees and bushes and the use of fungicides as well as the use of "other pesticides", indicating that these groups of pesticides are predominantly used in fruit production. The indicator for fruit trees is highly skewed to the right, indicating a high number of subjects with low potential exposure to fruit trees. Pesticide pressure indicators are less skewed as they combine application to multiple crops; the highest skewness is found for fungicides corresponding with the distribution for fruit trees. Statistical analysis revealed no association between the indicators and the incidence of bladder cancer. The results show that, using GIS, it is possible to reconstruct potential environmental pesticide exposure accounting for changes in pesticide use, crop area and residence history. Validation of the method with measured exposure is considered essential in view of its future application in other studies.

A GIS-based method for household recruitment in a prospective pesticide exposure study

BACKGROUND

Recent advances in GIS technology and remote sensing have provided new opportunities to collect ecologic data on agricultural pesticide exposure. Many pesticide studies have used historical or records-based data on crops and their associated pesticide applications to estimate exposure by measuring residential proximity to agricultural fields. Very few of these studies collected environmental and biological samples from study participants. One of the reasons for this is the cost of identifying participants who reside near study fields and analyzing samples obtained from them. In this paper, we present a cost-effective, GIS-based method for crop field selection and household recruitment in a prospective pesticide exposure study in a remote location. For the most part, our multi-phased approach was carried out in a research facility, but involved two brief episodes of fieldwork for ground truthing purposes. This method was developed for a larger study designed to examine the validity of indirect pesticide exposure estimates by comparing measured exposures in household dust, water and urine with records-based estimates that use crop location, residential proximity and pesticide application data. The study focused on the pesticide atrazine, a broadleaf herbicide used in corn production and one of the most widely-used pesticides in the U.S.

RESULTS

We successfully used a combination of remotely-sensed data, GIS-based methods and fieldwork to select study fields and recruit participants in Illinois, a state with high corn production and heavy atrazine use. Our several-step process consisted of the identification of potential study fields and residential areas using aerial photography; verification of crop patterns and land use via site visits; development of a GIS-based algorithm to define recruitment areas around crop fields; acquisition of geocoded household-level data within each recruitment area from a commercial vendor; and confirmation of final participant household locations via ground truthing. The use of these procedures resulted in a sufficient sample of participants from 14 recruitment areas in seven Illinois counties.

CONCLUSION

One of the challenges in pesticide research is the identification and recruitment of study participants, which is time consuming and costly, especially when the study site is in a remote location. We have demonstrated how GIS-based processes can be used to recruit participants, increase efficiency and enhance accuracy. The method that we used ultimately made it possible to collect biological samples from a specific demographic group within strictly defined exposure areas, with little advance knowledge of the location or population.

Carcinogenicity of agricultural pesticides in adults and children

The role of specific agricultural pesticides in relation to adult and childhood cancers has not been firmly established due to the lack of precise exposure data in previous studies. Improvements in exposure assessment, disease classification, and application of molecular techniques in recent epidemiological evaluations is rapidly improving our ability to evaluate the human carcinogenicity of agricultural pesticides. The role of pesticide exposures in the etiology of human cancer is outlined by anatomical site and recent development in exposure assessment and molecular epidemiology are summarized and evaluated.

Environmental pollutants and breast cancer: epidemiologic studies

Laboratory research has shown that numerous environmental pollutants cause mammary gland tumors in animals; are hormonally active, specifically mimicking estrogen, which is a breast cancer risk factor; or affect susceptibility of the mammary gland to carcinogenesis. An assessment of epidemiologic research on these pollutants identified in toxicologic studies can guide future research and exposure reduction aimed at prevention. The PubMed database was searched for relevant literature and systematic critical reviews were entered in a database available at URL: www.silentspring.org/sciencereview and URL: www.komen.org/environment (accessed April 10, 2007). Based on a relatively small number of studies, the evidence to date generally supports an association between breast cancer and polycyclic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs) in conjunction with certain genetic polymorphisms involved in carcinogen activation and steroid hormone metabolism. Evidence regarding dioxins and organic solvents is sparse and methodologically limited but suggestive of an association. Methodologic problems include inadequate exposure assessment, a lack of access to highly exposed and unexposed populations, and a lack of preclinical markers to identify associations that may be obscured by disease latency. Among chemicals identified in toxicologic research as relevant to breast cancer, many have not been investigated in humans. The development of better exposure assessment methods is needed to fill this gap. In the interim, weaknesses in the epidemiologic literature argue for greater reliance on toxicologic studies to develop national policies to reduce chemical exposures that may be associated with breast cancer. Substantial research progress in the last 5 years suggests that the investigation of environmental pollutants will lead to strategies to reduce breast cancer risk.

Agricultural pesticide use and hypospadias in eastern Arkansas

INTRODUCTION

We assessed the relationship between hypospadias and proximity to agricultural pesticide applications using a GIS-based exposure method.

METHODS

We obtained information for 354 cases of hypospadias born between 1998 and 2002 in eastern Arkansas; 727 controls were selected from birth certificates. We classified exposure on pounds of pesticides (estimated by crop type) applied or persisting within 500 m of each subject's home during gestational weeks 6 to 16. We restricted our analyses to 38 pesticides with some evidence of reproductive, developmental, estrogenic, and/or antiandrogenic effects. We estimated timing of pesticide applications using crop phenology and published records.

RESULTS

Gestational age at birth [odds ratio (OR) = 0.91; 95% confidence interval (CI), 0.83-0.99], parity (OR = 0.79; 95% CI, 0.65-0.95), and delaying prenatal care until the third trimester (OR = 4.04; 95% CI, 1.46-11.23) were significantly associated with hypospadias. Risk of hypospadias increased by 8% for every 0.05-pound increase in estimated exposure to diclofop-methyl use (OR = 1.08; 95% CI, 1.01-1.15). Pesticide applications in aggregate (OR = 0.82; 95% CI, 0.70-0.96) and applications of alachlor (OR = 0.56; 95% CI, 0.35-0.89) and permethrin (OR = 0.37; 95% CI, 0.16-0.86) were negatively associated with hypospadias.

CONCLUSIONS

Except for diclofop-methyl, we did not find evidence that estimated exposure to pesticides known to have reproductive, developmental, or endocrine-disrupting effects increases risk of hypospadias. Further research on the potential effects of exposure to diclofop-methyl is recommended.

Individual lifetime exposure to inorganic arsenic using a space-time information system

OBJECTIVES

A space-time information system (STIS) based method is introduced for calculating individual-level estimates of inorganic arsenic exposure over the adult life-course. STIS enables visualization and analysis of space-time data, overcoming some of the constraints inherent to spatial-only Geographic Information System software. The power of this new methodology is demonstrated using data from southeastern Michigan where 8% of the population is exposed to arsenic >10 microg/l (the World Health Organization guideline) in home drinking water.

METHODS

Participants (N=440) are members of a control group in a population-based bladder cancer case-control study in southeastern Michigan and were recruited by phone using random digit dialing. Water samples were collected and analyzed for arsenic at current residence and participants were required to answer questions concerning lifetime mobility history and dietary habits. Inorganic arsenic concentrations were estimated at past residences and workplaces, and in select foods. Fluid and food consumption data were integrated with mobility histories and arsenic concentrations to calculate continuous estimates of inorganic arsenic intake over the adult life-course.

RESULTS

Estimates of continuous arsenic exposure are displayed, making use of both participant age and calendar year as measures of time. Results illustrate considerable temporal variability in individual-level exposure, with 26% of the participants experiencing a change in drinking water arsenic concentration of at least +/-10 microg/l over their adult lives. The average cumulative intake over the adult life-course ranges from 2.53 x 10(4)-1.30 x 10(5) microg, depending on the selected exposure metric.

CONCLUSIONS

The STIS-based exposure assessment method allows for flexible inclusion of different parameters or alternative formulations of those parameters, thus enabling the calculation of different exposure metrics. This flexibility is particularly useful when additional exposure routes are considered, input datasets are updated, or when a scientific consensus does not exist regarding the proper formulation of the exposure metric. These results demonstrate the potential of STIS as a useful tool for calculating continuous estimates of adult lifetime exposure to arsenic or other environmental contaminants for application in exposure and risk assessment.

Comparison of spatial scan statistic and spatial filtering in estimating low birth weight clusters

Background

The purpose of this study is to examine the spatial and population (e.g., socio-economic) characteristics of low birthweight using two different cluster estimation techniques. We compared the results of Kulldorff's Spatial Scan Statistic with the results of Rushton's Spatial filtering technique across increasing sizes of spatial filters (circle). We were able to demonstrate that varying approaches exist to explore spatial variation in patterns of low birth weight.

Results

Spatial filtering results did not show any particular area that was not statistically significant based on SaTScan. The high rates, which remain as the filter size increases to 0.4, 0.5 to 0.6 miles, respectively, indicate that these differences are less likely due to chance. The maternal characteristics of births within clusters differed considerably between the two methods. Progressively larger Spatial filters removed local spatial variability, which eventually produced an approximate uniform pattern of low birth weight.

Conclusion

SaTScan and Spatial filtering cluster estimation methods produced noticeably different results from the same individual level birth data. SaTScan clusters are likely to differ from Spatial filtering clusters in terms of population characteristics and geographic area within clusters. Using the two methods in conjunction could provide more detail about the population and spatial features contained with each type of cluster.

Residential proximity to agricultural pesticide use and incidence of breast cancer in California, 1988-1997

California is the largest agricultural state in the United States and home to some of the world's highest breast cancer rates. The objective of our study was to evaluate whether California breast cancer rates were elevated in areas with recent high agricultural pesticide use. We identified population-based invasive breast cancer cases from the California Cancer Registry for 1988-1997. We used California's pesticide use reporting data to select pesticides for analysis based on use volume, carcinogenic potential, and exposure potential. Using 1990 and 2000 U.S. Census data, we derived age- and race-specific population counts for the time period of interest. We used a geographic information system to aggregate cases, population counts, and pesticide use data for all block groups in the state. To evaluate whether breast cancer rates were related to recent agricultural pesticide use, we computed rate ratios and 95% confidence intervals using Poisson regression models, adjusting for age, race/ethnicity, and neighborhood socioeconomic status and urbanization. This ecologic (aggregative) analysis included 176,302 invasive breast cancer cases and 70,968,598 person-years of observation. The rate ratios did not significantly differ from 1 for any of the selected pesticide categories or individual agents. The results from this study provide no evidence that California women living in areas of recent, high agricultural pesticide use experience higher rates of breast cancer.

Community-initiated breast cancer and environment studies and the precautionary principle

The precautionary principle implies the need for research paradigms that contribute to "strength of the evidence" assessments of the plausibility of health effects when scientific uncertainty is likely to persist and prevention is the underlying goal. Previous discussions of science that inform precautionary decision making are augmented by examining three activist-initiated breast cancer and environment studies--the Long Island, New York, and Cape Cod, Massachusetts, studies and the National Institute of Environmental Health Sciences breast cancer and environment centers. These studies show how the choice of research questions affects the potential of results to inform action. They illustrate a spectrum of public involvement, population- and individual-level epidemiologic study designs, and the crucial importance of developing and applying new exposure assessment methods. The exposure studies are key because they are critical in assessing plausibility (without exposure to a causal agent, there is no health effect), are prerequisite to health studies, and identify preventable exposures that could be reduced by precautionary policies, even in the absence of strong evidence of harm. The breast cancer studies have contributed to environmental and biological sampling programs for endocrine-disrupting compounds in drinking water and household air and dust and the application of geographic information systems for surveillance and historical exposure assessment. They leave unanswered questions about when to invest in large epidemiologic studies, when negative results are sufficient, and how to pursue ambiguous positive results in further research and policy.

Positional Accuracy of Two Methods of Geocoding

BACKGROUND

Geocoding is often used in epidemiologic studies to map residences with geographic information systems (GIS). The accuracy of the method is usually not determined.

METHODS

We collected global positioning system (GPS) measurements at homes in a case-control study of non-Hodgkin lymphoma in Iowa. We geocoded the addresses by 2 methods: (1) in-house, using ArcView 3.2 software and the U.S. Census Bureau TIGER 2000 street database; and (2) automated geocoding by a commercial firm. We calculated the distance between the geocoded and GPS location (positional error) overall and separately for homes within towns and outside (rural). We evaluated the error in classifying homes with respect to their proximity to crop fields.

RESULTS

Overall, the majority of homes were geocoded with positional errors of less than 100 m by both methods (ArcView/TIGER 2000, median = 62 m [interquartile range = 39-103]; commercial firm, median = 61 m [interquartile range = 35-137]). For town residences, the percent geocoded with errors of </=100 m was 81% for ArcView/TIGER 2000 and 84% for the commercial firm. For rural residences, a smaller percent of addresses were geocoded with this level of accuracy, especially by the commercial firm (ArcView/TIGER 2000, 56%; commercial firm, 28%). Geocoding errors affected our classification of homes according to their proximity to agricultural fields at 100 m, but not at greater distances (250-500 m).

CONCLUSIONS

Our results indicate greater positional errors for rural addresses compared with town addresses. Using a commercial firm did not improve accuracy compared with our in-house method. The effect of geocoding errors on exposure classification will depend on the spatial variation of the exposure being studied.

Public involvement in breast cancer research: an analysis and model for future research

Public involvement in health program planning has been taking place for many years, and has provided a precedent for the emergence of public involvement in research conducted since the early 1990s. Such involvement is now widely seen in breast cancer research, due to the large public concern and major social movement activity. This article reviews current practices and general models of public involvement in research and constructs a prototype. The authors interviewed researchers, program officers, and laypeople in order to understand the obstacles, processes, and benefits. They conclude that public involvement has major ramifications for the democratization of science and the construction of knowledge by teaching lay people about science and sensitizing researchers to concerns of the public. There is growing support on the part of scientists and government agents for public involvement.

Current practices in spatial analysis of cancer data: data characteristics and data sources for geographic studies of cancer

The use of spatially referenced data in cancer studies is gaining in prominence, fueled by the development and availability of spatial analytic tools and the broadening recognition of the linkages between geography and health. We provide an overview of some of the unique characteristics of spatial data, followed by an account of the major types and sources of data used in the spatial analysis of cancer, including data from cancer registries, population data, health surveys, environmental data, and remote sensing data. We cite numerous examples of recent studies that have used these data, with a focus on etiological research.

Residential proximity to agricultural pesticide use and incidence of breast cancer in the California Teachers Study cohort

We examined the association between residential proximity to agricultural pesticide use and breast cancer incidence among members of the California Teachers Study cohort, a large study of professional school employees with extensive information on breast cancer risk factors, followed for cancer incidence since 1995. We identified 1552 invasive breast cancer cases, diagnosed between 1996 and 1999, among 114,835 cohort members. We used California Pesticide Use Reporting data to select pesticides for analysis based on use volume, carcinogenic potential, and exposure potential; a Geographic Information System was used to estimate pesticide applications within a half-mile radius of subjects' residences. We applied Cox proportional hazard models to estimate hazard rate ratios (HR) for selected pesticides, adjusting for age, race, and socioeconomic status. We saw no association between residential proximity to recent agricultural pesticide use and invasive breast cancer incidence. HR estimates for the highest compared to the lowest exposure categories for groups of agents were as follows: probable or likely carcinogens (1.07, 95% confidence interval (CI): 0.86-1.32), possible or suggestive carcinogens (1.06, 95% CI: 0.87-1.29), mammary carcinogens (1.15, 95% CI: 0.90-1.48), and endocrine disruptors (1.03, 95% CI: 0.86-1.25). HR estimates for other groups and individual pesticides did not differ from unity, nor was there a trend for any groupings of or individual pesticides examined. Stratifying by menopausal status or family history of breast cancer did not substantially affect our results. Our analyses suggest that breast cancer incidence is not elevated in areas of recent, high agricultural pesticide use in California.

The geographic distribution of breast cancer incidence in Massachusetts 1988 to 1997, adjusted for covariates

BACKGROUND: The aims of this study were to determine whether observed geographic variations in breast cancer incidence are random or statistically significant, whether statistically significant excesses are temporary or time-persistent, and whether they can be explained by covariates such as socioeconomic status (SES) or urban/rural status? RESULTS: A purely spatial analysis found fourteen geographic areas that deviated significantly from randomness: ten with higher incidence rates than expected, four lower than expected. After covariate adjustment, three of the ten high areas remained statistically significant and one new high area emerged. The space-time analysis identified eleven geographic areas as statistically significant, seven high and four low. After covariate adjustment, four of the seven high areas remained statistically significant and a fifth high area also identified in the purely spatial analysis emerged. CONCLUSIONS: These analyses identify geographic areas with invasive breast cancer incidence higher or lower than expected, the times of their excess, and whether or not their status is affected when the model is adjusted for risk factors. These surveillance findings can be a sound starting point for the epidemiologist and has the potential of monitoring time trends for cancer control activities.

Using geographic information systems for exposure assessment in environmental epidemiology studies

Geographic information systems (GIS) are being used with increasing frequency in environmental epidemiology studies. Reported applications include locating the study population by geocoding addresses (assigning mapping coordinates), using proximity analysis of contaminant source as a surrogate for exposure, and integrating environmental monitoring data into the analysis of the health outcomes. Although most of these studies have been ecologic in design, some have used GIS in estimating environmental levels of a contaminant at the individual level and to design exposure metrics for use in epidemiologic studies. In this article we discuss fundamentals of three scientific disciplines instrumental to using GIS in exposure assessment for epidemiologic studies: geospatial science, environmental science, and epidemiology. We also explore how a GIS can be used to accomplish several steps in the exposure assessment process. These steps include defining the study population, identifying source and potential routes of exposure, estimating environmental levels of target contaminants, and estimating personal exposures. We present and discuss examples for the first three steps. We discuss potential use of GIS and global positioning systems (GPS) in the last step. On the basis of our findings, we conclude that the use of GIS in exposure assessment for environmental epidemiology studies is not only feasible but can enhance the understanding of the association between contaminants in our environment and disease.

Historical reconstruction of wastewater and land use impacts to groundwater used for public drinking water: exposure assessment using chemical data and GIS

Land use in geographic areas that replenish groundwater and surface water resources is increasingly recognized as an important factor affecting drinking water quality. Efforts to understand the implications for health, particularly outcomes with long latency or critical exposure windows, have been hampered by lack of historical exposure data for unregulated pollutants. This limitation has hindered studies of the possible links between breast cancer risk and drinking water impacted by endocrine disrupting compounds and mammary carcinogens, for example. This paper describes a methodology to assess potential historical exposure to a broad range of chemicals associated with wastewater and land use impacts to 132 groundwater wells and one surface water body supplying drinking water to 18 public distribution systems on Cape Cod, MA. We calculated annual measures of impact to each distribution system and used the measures as exposure estimates for the residential addresses of control women in the Cape Cod Breast Cancer and Environment Study (Cape Cod Study). Impact was assessed using (1) historical chemical measurements of nitrate at the water supply sources (performed as required by the Safe Water Drinking Act) and (2) a geographic information system analysis of land use within the zones of contribution (ZOCs) delineated for each well in a state-mandated wellhead protection program. The period for which these impact estimates were developed (1972-1995) was constrained by the availability of chemical measurements and land use data and consideration of time required for groundwater transport of contaminants to the water supply wells. Trends in these estimates for Cape Cod suggest increasing impact to drinking water quality for land use over the study period. Sensitivity analyses were conducted to assess the effect on the distribution of controls' cumulative exposure estimates from (1) reducing the area of the ZOCs to reflect typical well operating conditions rather than extreme pumping conditions used for the regulatory ZOCs, (2) assuming residences received their drinking water entirely from the closest well or cluster of wells rather than a volume-weighted annual district-wide average, and (3) changing the travel time considered for contaminants to reach wells from land use sources. We found that the rank and distribution of controls' cumulative exposure estimates were affected most by the assumption concerning district mixing; in particular, assignment of exposure estimates based on impact values for the closest well(s) consistently produced a larger number of unexposed controls than when a district-wide average impact value was used. As expected, the results suggest that adequate characterization of water quality heterogeneity within water supplies is an important component of exposure assessment methodologies in health studies investigating impacted drinking water.

GIS and disease

Geographic information systems (GIS) and related technologies like remote sensing are increasingly used to analyze the geography of disease, specifically the relationships between pathological factors (causative agents, vectors and hosts, people) and their geographical environments. GIS applications in the United States have described the sources and geographical distributions of disease agents, identified regions in time and space where people may be exposed to environmental and biological agents, and mapped and analyzed spatial and temporal patterns in health outcomes. Although GIS show great promise in the study of disease, their full potential will not be realized until environmental and disease surveillance systems are developed that distribute data on the geography of environmental conditions, disease agents, and health outcomes over time based on user-defined queries for user-selected geographical areas.

Public health, GIS, and spatial analytic tools

We review literature that uses spatial analytic tools in contexts where Geographic Information Systems (GIS) is the organizing system for health data or where the methods discussed will likely be incorporated in GIS-based analyses in the future. We conclude the review with the point of view that this literature is moving toward the development and use of systems of analysis that integrate the information geo-coding and data base functions of GISystems with the geo-information processing functions of GIScience. The rapidity of this projected development will depend on the perceived needs of the public health community for spatial analysis methods to provide decision support. Recent advances in the analysis of disease maps have been influenced by and benefited from the adoption of new practices for georeferencing health data and new ways of linking such data geographically to potential sources of environmental exposures, the locations of health resources and the geodemographic characteristics of populations. This review focuses on these advances.