Data were obtained on individuals 17 years or younger who developed acute pesticide-related illnesses while working. We excluded cases involving nonoccupational exposures, attempted suicides, intentional malicious use (e.g., attempted homicide), or exposure for a psychotropic effect. In addition, cases caused by disinfectants were excluded, because such cases are not tracked in many states.
Information on cases was provided by the Toxic Exposure Surveillance System (TESS), the California Department of Pesticide Regulation, the California Department of Health Services, the Texas Department of Health, the Washington State Department of Health, the Oregon Department of Human Services, the New York State Department of Health, the Florida Department of Health, the Louisiana Department of Health and Hospitals, and the Arizona Department of Health Services. TESS, maintained by the American Association of Poison Control Centers, collects poisoning reports submitted by approximately 85% of US poison control centers.6
Each of the state agencies that contributed data on cases maintains its own surveillance system for acute pesticide-related illness and injury. It should be noted that 4 states neither have poison control centers that participate in TESS nor have in place a state-based surveillance system (Maine, Mississippi, South Carolina, and Vermont).
The periods for which acute pesticiderelated illness and injury surveillance data were available varied by agency. TESS data were available for 1993 through 1998. Surveillance data from Texas are considered complete as of 1987; Oregon, as of 1988; New York and Washington State, as of 1991; Arizona and Louisiana, as of 1992; Florida, as of 1998; and California, as of 1989. Data from state agencies were collected through 1999.
The information collected by TESS and the state agencies includes date of illness, information on the ill individual (sex, age, signs, and symptoms), whether the illness occurred as a result of workplace exposures, and the pesticide or pesticides that produced the illness. Additional information collected by the state agencies but not by TESS includes race/ethnicity, occupation, industry, activity of the individual during the exposure, type of exposure (e.g., drift, direct spray, or exposure to a spill or leaking container), and whether personal protective equipment was used. For the present analysis, we defined use of personal protective equipment as use of goggles, face shields, gloves, or respirators.
The Environmental Protection Agency (EPA) acute toxicity category was sought for all pesticides responsible for illness. EPA classifies pesticide products into 1 of 4 acute toxicity categories based on established criteria. Pesticides having the highest toxicity are placed in category I, and those having the lowest are included in category IV. In the case of the present analyses, the acute toxicity category of the pesticide product responsible for causing an illness was often provided by the contributing state agency. When not provided, information on acute toxicity category was retrieved from a data set made available by EPA.
Information on illness severity was sought for all eligible cases. Except for Washington State and Louisiana, state agencies did not determine severity levels for the cases they identified. TESS criteria were used to assign severity levels to the cases provided by TESS and the other state agencies.6 Minor effects consisted of minimally bothersome health effects that generally resolved rapidly. Moderate effects consisted of non–life-threatening health effects that were more pronounced or prolonged than minor effects or of a systemic nature. Major effects consisted of life-threatening health effects or those resulting in “significant residual disability or disfigurement.”
To avoid repeated inclusion of the same case, we compared cases provided by each state agency with cases included in TESS. Cases that matched each other in terms of year and state of exposure, age, sex, and pesticide active ingredient were assumed to involve the same individual. Such individuals were included in the state agency totals only.
Case Definition
Cases were included only if health effects developed subsequent to pesticide contact and these effects were evaluated by poison control or state surveillance professionals as consistent with the known toxicology of the pesticide product. TESS relies on the experience and judgment of poison control center specialists managing specific cases to determine whether the affected individuals have symptoms and signs consistent with the pesticide exposure. No standardized criteria are used to make this determination. A full description of the standardized case definition used by each state agency is beyond the scope of the present article, but this information is available elsewhere.
7Data Analysis
SAS software (SAS Institute Inc, Cary, NC) was used for data management and in conducting χ
2 analyses to examine categorical data. Incidence rates among subjects aged 15 to 17 years were calculated for the period 1993 through 1998. The numerator was the total number of illness cases; the denominator was obtained from estimates of hours worked derived from the 1993 through 1998 administrations of the Current Population Survey.
8,9 The Current Population Survey does not provide data on workers younger than 15 years. In calculating incidence rates for young workers, it is preferable to use hours worked rather than employment counts.
9 The reason is that youths work fewer hours per week, and fewer weeks per year, than adults. Using employment counts would underestimate the risk of acute pesticide-related illnesses among young people.
Average annual incidence rates were calculated for young people employed in agricultural (Bureau of the Census industry codes 010–030) and nonagricultural (all other Census Bureau industry codes) industries. Because information on industry was not available from TESS, the assumption was made that the proportion of TESS cases involving individuals employed in agriculture was equal to the proportion found among the cases reported by state agencies. Male and female incidence rates and rates for each of 4 US regional areas were also calculated.
We calculated risks of acute pesticiderelated illness among individuals aged 15 to 17 years by comparing rates among these youths with those among adults aged 25 to 44 years.10 The data on adults were obtained from the same agencies that provided the data on youths, with the same exclusions applied. The age range of the adult comparison group was chosen a priori and was based on methodology used previously in examinations of occupational fatalities.11 We calculated the incidence rate ratio as the youth–adult ratio of number of acute pesticide-related illnesses per hour worked. A ratio greater than 1 would suggest that youths have a higher risk of acute pesticide-related illnesses than adults. Confidence intervals (CIs) were calculated according to methods described by Rothman.10