EPHT defines hazard as a factor that may adversely affect health. Many sources of pesticide hazard data exist. For example, national databases exist that describe the names and classes of pesticides, their federal and state registration status, and their toxicologic properties, although there is no single database that consolidates information on both the acute and chronic health effects of pesticides. The U.S. Food and Drug Administration (FDA) samples domestic and foreign food products for pesticide residues and funds states for local food surveillance (FDA 2001). However, no national-scale surveillance system exists that makes data available on pesticide production, import/export, sale, application, or use (Donaldson et al. 2002). Absent such data, the U.S. Environmental Protection Agency (EPA), Department of Agriculture, and Geological Survey estimate annual pesticide use by linking manufacturer, industry, grower, and crop survey data. The industry and trade association production and sales data used for these estimates are available for purchase (Kegley et al., unpublished data). Several states estimate pesticide use through similar combination of sales, use, and crop surveys, but the utility of this approach is limited to characterizing agricultural use and may be incomplete and inadequate to characterize geographic areas smaller than states or even regions (Thier 1997).

Five states mandate some form of comprehensive pesticide use reporting (PUR) and sales. California’s regulations require that agricultural and commercial applicators and government institutions file pesticide use reports with the state. Agricultural reports must contain information on the identity, quantity, location, method, date, and other volume and acreage data of restricted-use applications. Reports for nonagricultural applications are less detailed because they are aggregated by month and county. In addition, all pesticide sales must be reported at the first point of sale (California Department of Pesticide Regulation 2000). California’s use and sales systems permit public access to line-item data. Massachusetts, Oregon, New Hampshire, and New York require PUR that includes agricultural, nonagricultural, building, and institutional applications, with varying degrees of experience and public access (Kegley et al., unpublished data).

Oregon is the only state that currently requires tracking of household pesticide use through point-of-sale reporting, although the state’s fiscal crisis has prevented Oregon from collecting use and sales reports (PURS-Oregon 2004). New York’s system is the best equipped among state PURs to characterize urban pesticide use because address, type, and quantity must be provided for all structural and rodent applications. However, the New York legislature imposed the most restrictive of the states’ public access requirements, permitting release of raw data only for human health research and only if approved by a stakeholder health science board (New York State Environmental Conservation Law 1997).

Data from many of these state PUR systems have been used to produce research papers, reports, and white papers explaining the purpose, distribution, and quantities of largely agricultural pesticides. Investigators have used California’s PUR data for ecologic studies examining Parkinson disease (Ritz and Yu 2000) and cancer incidence (Mills 1998). Less specific pesticide use data from state and federal agricultural agencies have also been used to identify a pattern of birth defects associated with certain pesticide use (Garry et al. 1996; Schreinemachers 2003).

Exposure, for tracking purposes, is defined as the proximity and/or contact with a source of a disease agent in such a manner that effective transmission of the agent or harmful effects of the agent may occur (CDC 2003a). Pesticide exposure surveillance in the United States is largely limited to particular occupational cohorts—medical monitoring of applicators, for example—and to biomonitoring efforts to characterize exposures among representative samples of regional and national populations. In many states, occupational exposures resulting in depressed cholinesterase levels are reportable conditions and are useful for monitoring regulatory compliance, enforcing work rules, managing disease cases, and identifying risk factors (Calvert et al. 2004). But these data have limited generalizability to larger and more varied populations. The U.S. EPA National Human Exposure Assessment Survey, completed in the 1990s, evaluated pesticide exposures among representative populations in Arizona, the Midwest, and Maryland (Berry et al. 2000). This effort has not been replicated. As part of the Third National Health and Nutrition Examination Survey (NHANES III), CDC carried out biomonitoring for metabolites of several classes of pesticides. These data provide, for the first time, baseline exposure estimates for a representative U.S. population to a variety of pesticides (Barr et al. 2004). NHANES III has produced a wellspring of reports based on these results and illustrates how providing exposure data linked to personal descriptors can fill in critical knowledge gaps.

Health effects, for tracking purposes, are chronic or acute health conditions that affect the well-being of an individual or community and are measured in terms of illness and death (CDC 2003a). Although the health effects from pesticides may include acute and chronic conditions and reproductive effects, surveillance of their health impacts is effectively limited to nearly immediate toxic effects. The Toxic Exposure Surveillance System (TESS) is a national surveillance program that collects poison control data from all state and regional poison control centers. TESS records basic hazard, exposure, and individual information on pesticide-related inquiries, of which there were more than 96,000 in 2002 (Watson et al. 2003). Poison Control Center data are useful for identifying educational and outreach needs, identifying risk factors for poisonings, and investigating and identifying clusters and outbreaks. The Sentinel Event Notification System for Occupational Risk (SENSOR) at the National Institute for Occupational Safety and Health (NIOSH) supports pesticide-related illness and injury surveillance in 12 states and is used to identify outbreaks and emergency pesticide health effects (NIOSH 2004).

More than 40 states collect and report hospital discharge data, and pesticide-related hospitalizations are rare. Although more patients report to emergency departments for pesticide exposures than are admitted to hospitals, few states systematically collect and report these data.

The data described above that are systematically collected about pesticide hazards, exposures, and health effects describe the risks experienced by agricultural communities better than those experienced by other groups. There are many reasons, however, why large cities may be interested in developing pesticide tracking systems. A 1999 analysis of New York’s PUR data found that even though NYC accounts for < 1% of the total land area of the state, > 7% by volume of all pesticides applied in the state, and 13% by weight, were applied in NYC. Also, all five counties of NYC were included in the top 10 counties statewide for use of pesticides (Thier 2000).

Several events have elevated the city’s level of awareness about pesticides. Spraying of adulticides for controlling mosquitoes that carry West Nile virus (CDC 2003c), the rise in asthma hospitalizations in the late 1980s through the mid 1990s, the growing awareness of the links between pest infestations and health symptoms, high profile experiments in least-toxic pest control in low-income housing, and public hearings on methods of controlling rats (Kass and Outwater 2002) have all contributed to public concern regarding pesticide health effects. NYC residents have been the subject of several recent studies that have associated negative reproductive health outcomes among low-income women with residential exposure levels to chlorpyrifos (Berkowitz et al. 2003, 2004; Perera et al. 2003; Whyatt et al. 2004). As a result of these events, pesticides have taken on greater importance for public health and housing agencies.

Populations residing in large urban areas face special health risks from a variety of environmental concerns. In NYC and other older, densely populated, largely immigrant cities, environmental hazards tend to concentrate spatially, ethnically, and socioeconomically. Awareness of these hazards may sometimes be great, prompting important and appropriate advocacy and action by communities to ameliorate conditions that contribute to acute and chronic illness. Other times, communities or governmental officials have so little information that speculation, hyperbole, or inaction may result. Under these circumstances, public health agencies play a largely reactive role to public concerns. Failing to unite disparate information on hazards leaves agencies with an incomplete story, and inappropriate policy decisions may result. By linking data sources on pesticide use, housing quality and finance, demographics and socioeconomic status, exposures, and health, much more can be learned about where and why pesticides are used. This deeper understanding may promote the improved targeting of resources, education, and toxic use reduction efforts, as well as inform scientifically sound policy and legislation.

With feedback from a stakeholder advisory panel created to guide the development of the public health tracking program, DOHMH identified seven principles that would guide decision making on data acquisition, data architecture, analytic priorities, and public engagement: The pesticide tracking system should a) build upon existing and ongoing data collection systems; b) link hazard, human exposure, and human health effects data in scientifically valid and defensible ways; c) automate, to the extent possible, the importing, cleaning, and linking of data sources; d) build on, rather than duplicate, data and technical systems already under development by data providers; e) enable the development and tracking over time of public health environmental indicators; f ) satisfy the needs of a wide community of data users, analysts, advocates, and residents; and g) inform the development of public health and environmental interventions whose goals are to reduce health risks and improve environmental quality. In this section, we describe preliminary progress toward the creation of NYC’s pesticide tracking system.

We have completed the initial steps in the identification, acquisition, and assessment of data that can be used to characterize pesticide use, exposure, and health problems in NYC. The system we describe will be built largely on data sources that are pesticide related. Stakeholders are interested not only in the characterization of pesticide hazards, exposures, and poisonings but also in learning more about whether pesticide exposures are associated with Parkinson disease, neurologic disorders, development disabilities, and respiratory health. The potential of pesticide tracking to explore these concerns begins with building a base hazard and exposure system.

The final form, breadth, and analytic strength of this system will depend on many factors—data quality and completeness, the degree of sustained institutional and public support, sufficient funding, and staff resources among them. Despite logistical, resource, and methodologic limitations associated with the development of an urban pesticide tracking system, this system offers the potential for significant benefits for researchers, policy makers, residents, industry, and advocates. A hazard, exposure, and health outcome system has the potential to reveal relationships impossible to assess without linking data sets and to close significant gaps in our knowledge about how, where, when, why, and with what consequences pesticides are used in an urban environment.